Zoeken op trefwoord in de kennisvragen:

 

De meest bekeken kennisvragen:

DEPLOYMENT-Business models

There are multiple models/players:

The Branded Integrated Life-Style Model

It’s a sleekly designed experience, riding in this self-driving car. As elegantly designed as the sleekest smart phone.You use an app on your phone to summon your car when you need it or to program a daily pick-up. It’s as simple as setting the alarm on your phone.Your windshield doubles as a screen, synching seamlessly with your other connected devices. As you ride along, you swipe through applications and web sites, checking your progress and the local weather on a digital dashboard, uploading photos to your favorite web site or watching a video. When you arrive at your destination, the screens you’ve opened are synched and waiting for you on whatever device you pick up next.

In this model, perhaps a company with no traditional presence in the auto industry that is already an integral part of the consumer’s life outside the vehicle could become a key participant in the ecosystem. Since self-driving vehicles will no longer need the same level of rigorous testing and validation, and manufacturing could potentially be outsourced, their emphasis would be on consumer research, product development, and sale of integrated lifestyle experiences.

The Branded Lifestyle Value Proposition: Design, Technology, Software, Consumer experience

The Open System Model

It’s all about the data and how to use these data to customize the consumer value proposition.The market for big data

is growing exponentially. Market intelligence provider IDC predicts that by 2015 the “Big Data” market will be $16.9 billion, up from $3.2 billion in 2010.35 A major player in the data market might not want to manufacture vehicles, but could

well design a vehicle operating system. With more than a billion cars serving up trillions of data points about consumer behavior, traffic patterns, and topography, an operating system (OS) developer could afford to give away the OS but accrue significant value from the data they could aggregate. Who would manufacture the vehicle? The OS provider could partner with any of the world’s vehicle manufacturers—and not just the traditional automotive manufacturers. Partnerships could be established with one or more new players who might compete in the branded technology arena.

The Open System Value Proposition: Utility, Technology, Customization

Mobility On Demand Model

Zipcar was the pioneer in the shared-vehicle field, but other players are breaking into the market. Whereas current mobility on demand providers must make vehicles easily accessible for customers in urban areas, their vehicle maintenance and parking fees are high. With self-driving vehicles, proximity to end-users would no longer be necessary. Vehicles could be dispatched by taxi and car service companies.

Giant retailers with a core competence in managing complex distribution channels or fleet providers with the capability

to manage the complexity of renting and allocation of fleets could enter the fray and accrue significant value in the new ecosystem. New entrants in the market might compete at either end of the spectrum—with generic, low-cost utilitarian transportation on demand at one end (the low-cost airline model) and super-luxury mobile executive suites and sleeping pods at the other (the first class or private jet experience). Success will be determined by efficiency, reliability, flexibility, vehicle maintenance, customer service, ease of human-vehicle interface, and integration with existing consumer devices—and all the other psychographic factors that determine consumer behaviors and brand preferences.

The Mobility on Demand Value Proposition: Flexibility, Reliability, Convenience, Cost

The OEM Model

Traditional automotive manufacturers have decades of experience in designing and manufacturing vehicles, and shaping an emotional connection with consumers. But will they move fast enough to maintain their brand dominance? Smart automotive manufacturers should be planning now, thinking about how to restructure their organizations and what potential strategic investments they should be making. History has not been kind to those who get stuck protecting the status quo in the face of disruptive change. In fact, collaboration is already taking place across the ecosystem as companies strive to stay relevant.The joint project between Intel and DENSO36 to develop in-vehicle communication and information systems exemplifies the new cross-industry synergistic relationships.

Vertical integration is an option for companies looking to bring a critical skill or technology in house. Some vehicle manufacturers have established venture capital subsidiaries to invest in promising new technologies as a means of bridging any skill or technology gaps. Doing so may provide a competitive advantage in this rapidly evolving ecosystem.

The OEM Value Proposition: Design, Technology, HMI, Supply Chain Management “

Gevonden in (p.32-33): Self-Driving Cars, The Next Revolution

Regarding making PT more flexible:

Such a transformation of the system could also breathe new life into ideas of financing basic public-transportation services—on the one hand in the form of pay-as-you-drive, but also on a flat-rate basis financed via taxes or levied on all citizens, as is often debated for cities. Also, a high service density in suburban and even rural areas would justify a flat-rate levy and could in the process help to reduce private car use.”

Regarding offering new service options for PT:

Concerning intermodality, possibilities include more public transport services, even in the suburban and rural areas mentioned above (for urban areas, see Chap. 11). The benefits resulting from the use of autonomous vehicles are equally true in spatial and temporal terms, that is both for districts on the outskirts and off-peak hours. An economic lower limit resulting from frequency of use also applies here, however, even in view of the saved labor costs. This also means that a spatially highly dispersed use can only be covered to a limited extent by providing larger fleets. In any case, operating these vehicles would have to pay for itself in terms of initial outlay and operating costs.”

Gevonden in (p. 186 & 187): New Mobility Concepts and Autonomous Driving: The Potential for Change

The automotive industry is a global industry in which value is generated predominantly by suppliers to automakers. The Dutch automotive industry is no exception to this rule. Within specific areas in the automotive industry, the Netherlands even plays a significant role with leading innovative companies that are involved in automotive activities worldwide. In these areas, the Dutch automotive sector is highly innovative and possesses a considerable knowledge base. To further strengthen its role, the Dutch automotive sector has developed a vision supported by a strong ambition of the Dutch automotive industry to increase its annual revenues from Eur 12 bn to Eur 20 bn.

The Dutch automotive sector has two responses to the opportunities and challenges of today’s automotive industry: innovation and cooperation. Innovation is vital in the continuous struggle for cost reductions alongside increasing levels of quality, individuality, and personalisation, and legal requirements (e.g., noise, safety and emission). Effective cooperation is becoming more and more crucial as competitive advantage will gravitate towards those that discern their strengths and move quickly to build or join appropriate new collaborative networks.”

 

Gevonden in (p.7): Vision for the Dutch automotive sector

“We estimate that autonomous vehicles can save the US economy $1.3 trillion per year. We believe the large potential savings can help accelerate the adoption of autonomous vehicles.

We see five drivers of the cost savings: Fuel cost savings ($158 bn), accident costs ($488 bn), productivity gain ($507 bn), fuel loss from congestion ($11 bn), productivity savings from congestion ($138 bn).

This is our base case estimate. Our bull case estimate of savings is $2.2 tn/year and a bear case is $0.7 tn/year

This is a rough estimate. It does not account for the cost of implementing autonomous vehicles (one-time), offsetting losses, and investment implications. It also assumes 100% penetration of autonomous vehicles to achieve the full run-rate of potential savings.”

Gevonden in (p. 48): https://www.dropbox.com/s/mhckyj2bha4id0u/Morgan%20Stanley%20%282013%29%20AUTONOMOUS-CARS:-SELF-DRIVING-THE-NEW-AUTO-INDUSTRY-PARADIGM.pdf?dl=0

 

  • Use of existing and public infrastructure:
  • The concept of truck platooning combines automation with the usage of existing public infrastructure. This increases the compatibility of the concept. New infrastructure is often expensive and it may not be clear who is responsible for the infrastructure.
  • Realization of fuel savings:
  • Truck platooning has the potential to significantly save fuel. This also results in less emissions. Though, the order of fuel savings in practice is still uncertain. It was found that truck platooning is probably not feasible when only fuel savings are taken into account.
  • Larger truck driver productivity:
  • Drivers in following trucks may be in a standby mode at highways or even disappear for high levels of automation. It can be concluded that coordinated platoon formation is needed to benefit from larger trip distances that may be possible due to platooning. Labor cost savings may largely contribute to the adoption of platooning as many flows become feasible for platooning when labor costs drop.

 

 

Gevonden in (p. 49): https://www.dropbox.com/s/wnl33x2zqwwe5jy/MScThesisBakermans2016.pdf?dl=0

 

DEPLOYMENT-Samenwerking

Key players:

 

  • Evolutionary: auto industry (OEMs)
  • Revolutionary: non-automotive technology companies (google, apple, etc)
  • Transformative: high-tech start-ups

 

Gevonden in (p. 206):

https://www.dropbox.com/s/91n2z7i19wfgzu6/Beiker%2C%20S.%20%282016%29%20Deployment%20Scenarios%20for%20Vehicles%20with%20Higher-Order%20Automation.pdf?dl=0
Note Joop: In het onderzoek van Surf STAD is voor Nederland door Bart Stoffels een mooi overzicht van stakeholders opgenomen.

https://www.dropbox.com/s/vvb3xj5m1d4ntkp/zelfrijdende%20stad_20%20maart_BS.pptx?dl=0

Vehicle-to-X connectivity (V2X): Connectivity is an important element of the automated vehicles especially secure V2X communication requiring low latency. V2X technologies encompass the use of wireless technologies to achieve real-time two-way communication among vehicles (V2V) and between vehicles and infrastructure (V2I). The convergence of sensor-based solutions (current advanced driver assistance – ADAS) and V2X connectivity will promote automated driving. “

 

“Digital infrastructure: Digital infrastructure (for road automation) includes static and dynamic digital representations of the physical world with which the automated vehicle will interact to operate. Issues to address include: sourcing, processing, quality control and information transmission. “

Gevonden in: https://www.dropbox.com/s/4hor6dyblxeeinb/15CPB_AutonomousDriving.pdf?dl=0

Innovatie bevorderende wetgeving ontwikkelen

  • Om (testen met) zelfrijdende auto’s op de openbare weg juridisch mogelijk te maken, wordt de bestaande AMvB (Besluit ontheffing verlening exceptionele transporten) voor ontheffingverlening door de RDW aangepast. Ik verwacht dit voorstel begin 2015 aan uw Kamer te sturen. Tot die tijd is testen op kleinschalig niveau mogelijk. De RDW verleent dan ontheffing en beoordeelt samen met de wegbeheerders de veiligheid. Hierbij kijken we uiteraard naar wat er in de rest van de wereld gebeurt op dit terrein, bijvoorbeeld naar de regelgeving die in Californië is ontwikkeld.
  • Ik streef daarnaast naar (inter)nationale regelgeving die marktintroductie van zelfrijdende voertuigtechnologie mogelijk maakt. Daarvoor nemen we het initiatief in internationale overleggen (EU en VN) en steunen we relevante voorstellen. Ter voorbereiding op het EU-voorzitterschap van Nederland inventariseer ik welke regelgeving/kaders ten behoeve van zelfrijdende auto’s op Europees niveau zouden moeten worden aangepast of waar een gezamenlijk kader wenselijk is. Uiteraard werk ik hierbij samen met andere landen.
  • Grootschalige testen in de praktijk faciliteren en kennisontwikkeling:
  • Ik geef na de zomer uitsluitsel over de voorwaarden en de locatie waaronder eerder genoemde testaanvraag kan worden uitgevoerd. Hierbij betrek ik eventuele andere aanvragen.
  • Deze test gebruiken we om in de praktijk een basisprocedure en voorwaardenset voor het structureel testen van automatische voertuigtechnologie te ontwikkelen. Doel is veilig testen en de kennis structureel borgen voor volgende initiatieven en projecten. Hierbij werk ik samen met kennisinstellingen, bedrijfsleven, de RDW en wegbeheerders
  • We gaan actief deelnemen aan internationale initiatieven. Zo nemen we deel aan het World Economic Forum waarbij met de auto-industrie en andere relevante partijen barrières en mogelijke oplossingsrichtingen voor de zelfrijdende auto in kaart worden gebracht. Te denken valt aan vraagstukken rondom data (eigendom, beheer, uitwisseling en beveiliging) en aansprakelijkheid. Op nationaal niveau zal ik ook onderzoeken laten uitvoeren naar deze onderwerpen, daarbij neem ik ook privacy en rijvaardigheidseisen mee.

Gevonden in: https://www.dropbox.com/s/ze5qzm20upsqlye/grootschalige-testen-van-zelfrijdende-auto-s-4.pdf?dl=0

 

DEPLOYMENT-Toekomstverkenningen-en-Transitiepaden

“Automated driving, with its minimal space requirements and rather equal speed levels, could at least double the existing average road infrastructure capacity. “

Gevonden in (p.380): Autonomous Vehicles and Autonomous Driving in Freight Transport

 

There are multiple models/players:

The Branded Integrated Life-Style Model

It’s a sleekly designed experience, riding in this self-driving car. As elegantly designed as the sleekest smart phone.You use an app on your phone to summon your car when you need it or to program a daily pick-up. It’s as simple as setting the alarm on your phone.Your windshield doubles as a screen, synching seamlessly with your other connected devices. As you ride along, you swipe through applications and web sites, checking your progress and the local weather on a digital dashboard, uploading photos to your favorite web site or watching a video. When you arrive at your destination, the screens you’ve opened are synched and waiting for you on whatever device you pick up next.

In this model, perhaps a company with no traditional presence in the auto industry that is already an integral part of the consumer’s life outside the vehicle could become a key participant in the ecosystem. Since self-driving vehicles will no longer need the same level of rigorous testing and validation, and manufacturing could potentially be outsourced, their emphasis would be on consumer research, product development, and sale of integrated lifestyle experiences.

The Branded Lifestyle Value Proposition: Design, Technology, Software, Consumer experience

The Open System Model

It’s all about the data and how to use these data to customize the consumer value proposition.The market for big data

is growing exponentially. Market intelligence provider IDC predicts that by 2015 the “Big Data” market will be $16.9 billion, up from $3.2 billion in 2010.35 A major player in the data market might not want to manufacture vehicles, but could

well design a vehicle operating system. With more than a billion cars serving up trillions of data points about consumer behavior, traffic patterns, and topography, an operating system (OS) developer could afford to give away the OS but accrue significant value from the data they could aggregate. Who would manufacture the vehicle? The OS provider could partner with any of the world’s vehicle manufacturers—and not just the traditional automotive manufacturers. Partnerships could be established with one or more new players who might compete in the branded technology arena.

The Open System Value Proposition: Utility, Technology, Customization

Mobility On Demand Model

Zipcar was the pioneer in the shared-vehicle field, but other players are breaking into the market. Whereas current mobility on demand providers must make vehicles easily accessible for customers in urban areas, their vehicle maintenance and parking fees are high. With self-driving vehicles, proximity to end-users would no longer be necessary. Vehicles could be dispatched by taxi and car service companies.

Giant retailers with a core competence in managing complex distribution channels or fleet providers with the capability

to manage the complexity of renting and allocation of fleets could enter the fray and accrue significant value in the new ecosystem. New entrants in the market might compete at either end of the spectrum—with generic, low-cost utilitarian transportation on demand at one end (the low-cost airline model) and super-luxury mobile executive suites and sleeping pods at the other (the first class or private jet experience). Success will be determined by efficiency, reliability, flexibility, vehicle maintenance, customer service, ease of human-vehicle interface, and integration with existing consumer devices—and all the other psychographic factors that determine consumer behaviors and brand preferences.

The Mobility on Demand Value Proposition: Flexibility, Reliability, Convenience, Cost

The OEM Model

Traditional automotive manufacturers have decades of experience in designing and manufacturing vehicles, and shaping an emotional connection with consumers. But will they move fast enough to maintain their brand dominance? Smart automotive manufacturers should be planning now, thinking about how to restructure their organizations and what potential strategic investments they should be making. History has not been kind to those who get stuck protecting the status quo in the face of disruptive change. In fact, collaboration is already taking place across the ecosystem as companies strive to stay relevant.The joint project between Intel and DENSO36 to develop in-vehicle communication and information systems exemplifies the new cross-industry synergistic relationships.

Vertical integration is an option for companies looking to bring a critical skill or technology in house. Some vehicle manufacturers have established venture capital subsidiaries to invest in promising new technologies as a means of bridging any skill or technology gaps. Doing so may provide a competitive advantage in this rapidly evolving ecosystem.

The OEM Value Proposition: Design, Technology, HMI, Supply Chain Management “

Gevonden in (p.32-33): Self-Driving Cars, The Next Revolution

“This chapter explored three scenarios for the deployment of vehicles with higher-order automation: the continuous evolution of driver assistance systems by the established auto industry, the revolution of personal mobility by non-automotive technology companies, and the transformative merging of private and personal mobility by start-ups and transportation service providers. “

Gevonden in (p. 207):

https://www.dropbox.com/s/91n2z7i19wfgzu6/Beiker%2C%20S.%20%282016%29%20Deployment%20Scenarios%20for%20Vehicles%20with%20Higher-Order%20Automation.pdf?dl=0

Note Joop: Zie ook recente verzameling transitiepaden door Tom Alkim samengesteld:  https://www.dropbox.com/s/bzo215xwhwdyvth/Overview%20Roadmaps%20Automated%20Driving_final_withoutaspect.pdf?dl=0

Brand equity has been found to be positively related to customer loyalty and willingness to pay. While strong brands are generally helpful for the marketing of products and services, the importance of brands has been found to vary across industry sectors, with a high relevance for the marketing of automobiles [22]. The relevance of branding strongly depends on the function of the brand as risk reducing factor, its function to enhance information efficiency, and its symbolic value. Since the purchase of a new car is an extensive decision involving comparably high expenditures and the collection of extensive information, strong brands can promote the purchasing process.

Besides the sparse empirical evidence for the risk-reducing effects of strong brands in the context of automated driving [13], the aforementioned brand functions should be positively related to consumer acceptance of automated driving systems. Knowledge and experience of consumers with automated driving technology is marginal. In combination with additional cost for automated driving abilities, consumers are likely to evaluate a purchase decision as risky. Strong brands can effectively help to reduce perceptions of risk. “
Gevonden in (p.691): Consumer Perceptions of Automated Driving Technologies: An Examination of Use Cases and Branding Strategies

“While consumers still have many questions about safety, liability and the operation of self-driving cars, their receptivity increased significantly when presented with the right value proposition, which can be summed up as follows: shorter commute times + reduced traffic-related variability + the ability to use the vehicle in either self-driving or human- operated mode (self-driving on/off) = a strong incentive for consumer adoption.

Companies that get the value proposition right – and deliver a mobility/driving experience that is esthetically and emotionally pleasing could dominate the market. Companies that miss the mark on either the technology or the mobility experience could find themselves left behind. “

Gevonden in (p.4): Self-Driving Cars: Are We Ready?

80% Driverless future A policy roadmap for city leaders Page 12 report AVs will reduce demand for parking, gas stations, and other auto-related land uses. Some uses, particularly those in highly desirable areas, may be reused and repurposed over time. AVs are highly likely to reduce parking demand by taking personally owned automobiles off the street. Past studies estimate that, depending on the success of merging AV into city infrastructure, parking demand may be reduced by up to 90%. Parking, roads and other auto-related uses occupy a significant amount of land. The U.S. contains as many as two billion parking spaces, occupying up to 16,000 square miles of land (the equivalent of Connecticut and Vermont combined). The quantity of parking spaces in the country amounts to as many as eight parking spaces for every car. Parking consumes a significant amount of land, especially in suburban areas where auto use is highest and surface lots are more common than multi-story garages. At a typical suburban mall, parking or driveways make up 80% of the land, while only 20% is used for the mall. Even in denser, more urban areas, parking requires significant land area. For example, streets and parking take up 45% of land in downtown Washington, D.C. and up to 65% in downtown Houston.

HUMAN BEHAVIOUR-Acceptatie

Brand equity has been found to be positively related to customer loyalty and willingness to pay. While strong brands are generally helpful for the marketing of products and services, the importance of brands has been found to vary across industry sectors, with a high relevance for the marketing of automobiles [22]. The relevance of branding strongly depends on the function of the brand as risk reducing factor, its function to enhance information efficiency, and its symbolic value. Since the purchase of a new car is an extensive decision involving comparably high expenditures and the collection of extensive information, strong brands can promote the purchasing process.

Besides the sparse empirical evidence for the risk-reducing effects of strong brands in the context of automated driving [13], the aforementioned brand functions should be positively related to consumer acceptance of automated driving systems. Knowledge and experience of consumers with automated driving technology is marginal. In combination with additional cost for automated driving abilities, consumers are likely to evaluate a purchase decision as risky. Strong brands can effectively help to reduce perceptions of risk. “
Gevonden in (p.691): Consumer Perceptions of Automated Driving Technologies: An Examination of Use Cases and Branding Strategies

“While consumers still have many questions about safety, liability and the operation of self-driving cars, their receptivity increased significantly when presented with the right value proposition, which can be summed up as follows: shorter commute times + reduced traffic-related variability + the ability to use the vehicle in either self-driving or human- operated mode (self-driving on/off) = a strong incentive for consumer adoption.

Companies that get the value proposition right – and deliver a mobility/driving experience that is esthetically and emotionally pleasing could dominate the market. Companies that miss the mark on either the technology or the mobility experience could find themselves left behind. “

Gevonden in (p.4): Self-Driving Cars: Are We Ready?

30% Ethics and selfdriving cars

Niet concreet een keuze, maar mogelijkheden voor de keuze van verantwoordelijkheid.

Road traffic also raises important issues of tort liability: who should pay for the costs of the accident. From a legal perspective the introduction of ADS presents at least three new issues (Pagallo 2013: 110): first, the law has so far seen robots and autonomous systems merely as tools and not as agents and doesn’t seem equipped to cope with the presence of non-human intelligent systems (see also Calo 2016); secondly and relatedly, when systems equipped with complex artificial intelligence are used the driver/owner may not always be responsible for the behaviour of the system: sometimes others should, other times nobody may, for instance in the event of a malfunctioning that no reasonable person could have predicted; in this respect things can be even more complex in the case of a shared vehicle, where owner and user do not coincide (see section below on ownership); third, unlike what happens for instance with robo-traders, liability for road accidents concern also “extra-contractual” third parties, that is parties not bound by any contractual relationship with the owner/driver (a typical example here would be an unknown pedestrian).

30% Societal Risk Constellations for Autonomous Driving. Analysis, Historical Context and Assessment

Geeft niet daadwerkelijk antwoord op de vraag. Wel wordt genoemd op welke gebieden actie ondernomen moet worden

Significant problems with data protection and privacy are to be expected if autonomous driving takes place within networked systems (although autonomous driving does not pose specific issues in this regard compared to other fields). Technical and legal measures (Chap. 24) should be introduced here to take account of the wide-ranging public debate on these issues (e.g. NSA, indiscriminate data collection).

 

30% Ethics and selfdriving cars

Niet zozeer maatregelen, maar mogelijke gebeurtenissen die kunnen plaatsvinden op gebied van sociale veiligheid door de ZRA.

From a technical point of view, many of the valuable goals listed and discussed so far: safety, accountability, accessibility may be achieved by a huge acquisition, storage and elaboration of data. Both partial automation and supervised automation systems may need to acquire a huge amount of data about the behaviour of the vehicles and their drivers or passengers via sensing and communication technologies. Moreover, by being equipped with sensors and cameras, the vehicles are also likely to incidentally acquire many data on other road users interacting with them.

Two ethical and societal risks highlighted in the ethical literature on privacy and data protection (e.g. van Den Hoven 2008) are clearly present also in the case of ADS. Firstly, ADS may be the target of cyber attacks or hacking. Secondly, the massive acquisition and storage of personal data about road users may threaten their moral autonomy in two ways: a) by creating an information asymmetry: a huge quantity of information about individual persons may become available to those who own or control transport infrastructures. This information may be used to benefit citizens, but there is also a risk that it will be collected and used against the interests of minorities or even the majority of people and in violation of their rights; b) by creating an imbalance of power: similarly, a dramatic increase in the information capabilities of governments or other agencies may enhance their capacity to promote the citizen’s safety and well-being, but this capacity may also be used to control, coerce, exploit, discriminate and even oppress people.

HUMAN BEHAVIOUR-Gebruikersgedrag

50%

A Proposed Psychological Model of Driving Automation

Niet perse voor het behalen voor bepaalde doelen, maar die invloed hebben op het gebruik van de zelfrijdende auto

From the review of the literature, we believe that the interdependency between
the psychological concepts underrepresented. The mental model that a user develops
about a system is critical to performance and operations with that system. Given the
previous review of psychological factors associated with the use of automation, there are
some obvious (and some not-so-obvious) interrelations between the variables. From the
literature, it would seem that mental workload plays a central role in the relationship.
For instance, it is apparent that high workload in the form of traffic congestion can
increase stress (Wilson and Rajan, 1995), but there is some evidence that this
relationship is bi-directional. Matthews and Desmond (1995a; 1997) provide evidence
for the mechanism behind this relationship, and from this there are two novel yet logical
conclusions relating workload to stress. The first is that stress can affect performance in
low as well as high workload conditions. The second is that the effort involved in coping
with stress actually adds to the task demands.

30% Intelligent Cruise Control Field Operational Test

 

As implied above, it is assumed that high levels of ACC penetration into the vehicle population will cause extended strings of ACC-equipped vehicles to form spontaneously simply due to the probabilities of traffic mixing— even in the absence of any peculiar natural tendencies toward aggregation of vehicles under ACC control. Thus, the dynamic stability of ACC strings and their impact on the natural inter lane weaving movements of other traffic will constitute real issues if ACC becomes a successful product. Observations from these tests have indicated that significant traffic impacts could arise from ACC strings. Firstly, considering simply the ACC system that was fielded here, (with its low deceleration authority and relatively sluggish re-acceleration response) a string of more than four of these vehicles will exhibit marginal stability levels, yieldingexaggerated responses when longitudinally disturbed from the forward end of the string. With strings of eight vehicles equipped with this ACC controller, significant disruptions in the smooth movement of a traffic stream would ensue following modest disturbances. Further to the string-stability issue, the authors of this report are not aware that this characteristic is being considered in the current design of automotive ACC products. Infact, an opposite approach has been apparent by which ACC control algorithms are “detuned” in some emerging products to render the controller unresponsive to brief misdetections by the range sensor. While string-stability problems would not manifest themselves as long as ACC-equipped vehicles are a rarity on the road, the issue will become highly important whenever the population density begins to precipitate long string formation on a regular basis. On the matter of cross-lane movements of other traffic, an important issue arises when an ACC string constitutes a sort of “moving wall” that impedes the natural weaving movement of other traffic. That is, due to ACC’s regularization of headway spacing,randomly extended gaps do not occur in the same manner as seen in manually-controlled traffic. Further, the ACC controller does not, by itself, respond to the “body language” of other drivers who maneuver alongside, in an adjacent lane, with the clear intention of weaving across into another destination lane. When headway time is in the vicinity of 1.0second, at highway speed, it was seen that other motorists were basically thwarted in their attempts to change lanes through an eight-car string that occupied the next-to-right-mostlane— occasionally exhibiting a fairly dramatic rate of penetrating the string in their apparent frustration to find a fully suitable gap in line with their exit/entrance transition plans. (Note that, upon entering a freeway, some more aggressive drivers seek to occupy the “fast,” left-most lane as soon as possible— thus experiencing some frustration when they remain “stuck” in the rightmost lane while searching for a suitable gap.) When ACC headway times were uniformly set to 1.5 seconds, other drivers appeared to penetrate the string with minimal difficulty.

“One central aspect of human-machine interaction is the perceived autonomy of the consumer [4, 29]. While the role of consumer autonomy has been addressed directly or indirectly by some studies, its criticality for consumer acceptance of automated technologies might not be fully captured in the contexts studied. Restricting or removing the autonomy of individuals could cause reactance, i.e., negative psychological and contrary behavioral responses of consumers as reactions to a perceived restriction of their personal freedoms [6, 44]. Automated driving systems could be perceived as a threat to drivers’ autonomy, and reactance could arise in terms of consumer boycott intentions or low adoption rates. Presently, it it is unclear if consumers are willing to accept a loss in control [56].”

Gevonden in (p.690): Consumer Perceptions of Automated Driving Technologies: An Examination of Use Cases and Branding Strategies

 

50% Behavioural adaptation, risk compensation, risk homeostasis and moral hazard in traffic safety

The researchers concluded that the observed behaviours are associated with increased accident likelihood and that one should be cautious about the potential safety of the ACC systems. The results showed that drivers using ACC react too late and collide more often with a stationary queue than unsupported drivers. The researcher suggests that these collisions are due to the driver’s expectations that the ACC system would cope with the situation. Although informed about ACC limitations, drivers seem to have problems to identify situations requiring them to take over control. Similar conclusions were drawn in a study of Stanton et al. (1997, quoted in Rudin-Brown and Parker, 2004).

In an emergency stop scenario, the average maximum braking was larger and the average minimum time headway was smaller when driving with ACC. The type of driving style group made little
difference to these behavioural adaptations. The researchers concluded that the observed behaviours are associated with increased accident likelihood and that one should be cautious about the potential safety of the ACC systems.

The results show that the drivers reacted more slowly and less often within

a safe time period (approximately 33% less often) on this task when using ACC. This effect was particularly pronounced in ‘high sensation seekers’. Furthermore, ACC was also associated with impaired lane-keeping performance and high sensation seekers were more deviating within the lane then the other test persons.

50% Misconceptions and self-reported behavioral adaptations associated with advanced in-vehicle systems: lessons learned from early adopters.

Gevaren van het gebruik van ACC

While nearly all drivers reported reading all or some part of the owner’s manual relating to the ACC system, many held misconceptions about the functional capabilities of the system. For example, most owners mistakenly believed the ACC system would react to a stopped in-path vehicle, and many were not aware that the system provided an approach warning feature that alerts the driver when manual intervention is required in situations where the system’s braking authority is exceeded. Of greater concern are drivers who think that the warning feature operates all the time, when in fact it does not. Over 6% of drivers in our sample were under the mistaken impression that the approach warning feature is active in their vehicles even when the ACCsystem is disengaged; these individuals are assuming a greater level of protection than the system actually provides. Again, this misinterpretation of the system’s capability was not moderated by experience with the system.

Of some concern is the finding that a substantial percentage of system owners (27%) were unsure of the underlying basis for how the system triggered the warning (distance or a combination of speed and distance). Inexperienced users, those with less exposure to the park aid system, were more likely to incorrectly assume that the system adjusts the timing of the warning based on both speed and distance to the obstacle, rather than using a fixed warning timing based on distance alone. Nearly twice as many “low experience” users (22%) were operating under this false assumption compared to their more experienced counterparts (11%). The vast majority of drivers were also unaware of the system’s functional speed limitations; 67% believed that the park aid system operates under any speed when backing (most systems only operated at speeds under 6 mph). Experience with the system also did not appear to improve understanding of the system’s functional speed range.

 

Nevertheless,misconceptions are common, suggesting that drivers need to be better educated about the system’s capability and limitations. Some form or degree of driver behavioral adaptation was reported to have occurred for each of the systems examined (Llaneras, 2006). Despite access to a wide array of information about their in-vehicle system, responses to knowledge-based questions about the systems themselves suggest that key information was not necessarily acquired or understood by a large number of drivers. Many drivers held misconceptions about the performance capabilities of their advanced systems. For example, 99% of ACC system owners did not know that the system ignores stopped vehicles. The fact that the system ignores stopped or slow-moving vehicle is available in the owner’s manuals for all of the ACC systems reviewed, yet drivers were not aware of this important operational characteristic. Similarly, 41% of park aid system owners did not know that the system warning is tied solely to the distance to objects and does not take into account their closing speed. This suggests that drivers’ mental models of how these systems function and perform do not always match reality, and additional efforts are needed to increase driver understanding of how these systems operate. This is particularly important for safety-related misconceptions.

HUMAN BEHAVIOUR-Human Machine Interaction

50% Transition of control in highly automated vehicles A literature review R-2015-22

According to Flemisch et al. (2012), there are four factors that define the relationship between drivers and highly automated vehicles where the automated systems primarily perform the driving task and the driver performs the driving task occasionally. These factors are: ability, authority, control and responsibility.

Factoren die relevant kunnen zijn voor de interactie.

  • Interactie have en have-not’s? Vaste set van indicatoren om rijgedrag te monitoren?

 

100% Effects of adaptive cruise control and highly automated driving on workload and situation awareness: A review of the empirical evidence.

 

Many Human Factors researchers would probably agree that workload and situation awareness are two of the most important Human Factors constructs that are predictive of performance and safety (McCauley & Miller, 1997;Parasuraman, Sheridan, & Wickens, 2008;Sarter & Woods, 1991; Stanton & Young, 2000). Accordingly, the aim of this study is to quantify the effects of ACC and

HAD on workload and situation awareness.

 

Een aantal indicatoren om rijgedrag te monitoren, terwijl er met een zelfrijdende auto gereden wordt.

 

100% Rijtaakindicatoren voor C-ITS-projecten

 

Indicatoren ten behoeve van de doorstroming

Snelheid (puntsnelheid, gemiddelde snelheid, continue snelheid)

Afstand tot de voorligger

Rijstrookkeuze en aantal wisselingen

Acceleratie (met name bij oplossen file)

Aantal voertuigen op de weg (situationele variabele)

Longitudinale positie

Laterale positie

Gebruik signalering van auto, bv richtingaanwijzer

Remgedrag

 

Rijtaakindicatoren verkeersveiligheid

Snelheid (puntsnelheid en gemiddelde snelheid)

Snelheid naderen kruispunt en voorliggers

Aantal overschrijdingen maximum snelheid

Acceleratie/Deceleratie

Afstand tot voorligger (TTC)

Tijdsduur binnen bepaalde TTC

Remkracht

Aantal keren dat bovengemiddeld geremd wordt

Aantal rijstrookwisselingen

80% AUTOMATED DRIVING FUNCTIONS GIVING CONTROL BACK TO THE DRIVER

Heeft hier onderzoek gedaan, maar met klein aantal participanten (16) and in een simulator. Verder onderzoek is nodig met meer mensen en met echte voertuigen.

Objective Results

The effect of the additional task is evaluated through the reaction time of the drivers on the confirmation request,and the steering behaviour after regaining control and taking the exit. This is shown in Figure 6.

Figure 6.  zijn de grafieken met tijden, kan deze hier niet weergegeven, zie document.

 

100% Transition of control in highly automated vehicles A literature review R-2015-22

Measures for vehicle control were the standard deviation of the lateral position (SDLP) and the frequency of steering adjustments. There were two conditions: (1) moments when a switch to manual driving was required while drivers were attentively scanning the forward roadway while the vehicle was in fully automated mode, and (2) at moments the eyetracking equipment indicated that drivers were not attentively scanning the forward roadway while the vehicle was in the fully automated mode. When drivers were attentive, switching to manual and regaining proper control over the vehicle took on average 10 s. When drivers were less attentive when driving in the fully automated mode, switching to manual and regaining full control over the vehicle took circa 35-40 s. These results imply that especially when drivers are not attentive, messages about a switch tomanual must be provided properly and timely. These results also indicate that planned switches to manual driving have to occur in traffic situationswhere crash risk is low.

 

Ook dit komt uit enkel onderzoek, geeft wel een precieze tijd, dus antwoord op de vraag. Geeft andere resultaten dan het artikel hierboven. Wel is dit onderzoek accurater.

5/10% AUVSI-TRB-Symposium2015-Presentaties-Human-Factors

 

S.Hill presentatie geeft resultaten van een onderzoek naar tijd van out-of-loop naar in-to-loop

  1. Green presentatie geeft een formula om de tijd van controle overname te berekenen

K,Lee presentatie geeft ook resultaten van een onderzoek naar de tijd van controle overname

De overige presentaties geven geen antwoord op een van de vragen.

 

100% Effects of adaptive cruise control and highly automated driving on workload and situation awareness: A review of the empirical evidence.

Verschillende experimenten uitgevoerd, waarin verschillende tijden naar voren zijn gekomen. De antwoorden staan verdeeld over grote stukken tekst in het artikel, het antwoord is meerdere pagina’s lang.

100% AUTOMATED DRIVING FUNCTIONS GIVING CONTROL BACK TO THE DRIVER

At a certain distance upstream from the exit, the participant was warned and requested to provide a confirmation by pushing a button on the touch screen of the interface (Figure 2). If the driver did not confirm within a certain time the warning and confirmation request was repeated. Closer to the exit, irrespective of the driver reacting to the confirmation request, a warning was displayed that provided the amount of meters till the exit (Figure 1 without the confirmation request and button). The timings of the warnings and feedback requests were different between the attentive and inattentive driver states (see Table 1). The unadapted transition strategy was to warn the participant and ask for confirmation the first time at 1000 m before the exit. From 500 m before the exit the participant was continously informed on the distance (‘count down’) till the VTB system would switch off. In the adapted strategy, Willemsen 5 the participant was warned and asked for confirmation earlier, at 2000 m before the exit and the ‘count down’ was shown from 1000 m before the exit. In both strategies, if the participant did not react to the first confirmation request, a second one was issued at 750 m before the exit.

100%  The experimental setup of a large field operational test for cooperative driving vehicles at the A270.

To be able to choose a suitable way of communicating with the driver through a HMI in the A270 experiments, several HMI alternatives are tested by means of a driving simulator.

The chosen HMI design consists of a triangle which fills up with red when (more) deceleration is needed or a circle which fills up with green when (more) acceleration is needed, see left and right plot of Figure 3, respectively. The color signs are only shown when needed. As soon as no acceleration or deceleration is requested from the driver, i.e. a constant speed must be kept, the display is either a gray triangle or gray circle. When acceleration or deceleration gets more urgent an acoustic signal is added to the visual display saying “speed up” or “slow down”, respectively. The reason behind the sound is that more attention is attracted to the needed action from the driver and it gives the driver an extra motivation to follow up the advice.

Ook hier is een geëxperimenteerd met mogelijke manieren om de bestuurder weer in de loop te krijgen.

 

100% Effects of adaptive cruise control and highly automated driving on workload and situation awareness: A review of the empirical evidence.

In a study by Brook-Carter et al. (2002), a red rectangle appeared on the simulator screen and the participant had to respond as quickly as possible by pressing the horn.

In a driving simulator study by Ma (2006), participants were requested to press a button on the steering wheel when the navigation aid was activated, which occurred after about 9 min of driving.

De Winter et al. (2014) found that drivers responded faster to arrow-shaped stimuli projected on the simulator screen during HAD as compared to manual driving

 

Hier worden verschillende manieren toegepast.

HUMAN BEHAVIOUR-Rijvaardigheidseisen

0% t.a.v. rijvaardigheidseisen zijn weinig tot geen onderzoeken en presentaties beschikbaar en zijn de vragen nog open.

Een algemeen beeld is dat het rijden eerst ingewikkelder wordt (level 3), waarbij de bestuurder (fall back ready user) de beperkingen van de automatische piloot goed moet kennen. Vanaf SAE level 4 naar 5, nemen de eisen af en is het bv denkbaar dat de automatische pilot het rijbewijs moeten behalen als veiligheidseis. [Joop Veenis]

Yes. Until 1 January 2016 such (level 1) systems should be switched off by the candidate while taking the test to ensure a similar test situation for every candidate, focused on the driving task itself (the use of ADAS changes this driving task). However, to create a realistic and future-proof test situation, it is necessary to implement the use of such systems in driver training and testing. Therefore, CBR allows candidates to use ADAS during the driving test on a voluntary basis. If so, they will be judged on using these systems in relation to a proper performance of the driving task.

IMPACT-Digitale Infrastructuur

50% 

Zoeken naar strepen op het asfalt

Page 4 of document : Jene van der Heide, Senior adviseur Strategie en Beleid bij het Kadaster: ‘Voor kaartenmakers is het te duur om ook de afgelegen wegen in het buitengebied in kaart te brengen. En de overheid verzamelt uitsluitend informatie die nodig is voor het onderhoud van zulke wegen. Rijd je zo’n gebied binnen, dan vraagt een zelfrijdende auto waarschijnlijk aan zijn bestuurder om het even van hem over te nemen. Het kan zomaar 10 jaar duren voordat dit verschil tussen buitengebied en snelweg is weggewerkt.’

50%

ROAD SAFETY WITH SELF-DRIVING VEHICLES: GENERAL LIMITATIONS AND ROAD SHARING WITH CONVENTIONAL VEHICLES

Chapter 3.2.1

Gomes(2014) argued that, “all 4 million miles of U.S. public roads will need to be mapped, plus driveways, off-road trails, and everywhere else you’d ever want to take the car” and this information would need to include “locations of street lights, stop signs, crosswalks, lane markings, and every other crucial aspect of a roadway.”

70%

Zoeken naar strepen op het asfalt

Page 5 report: Welke rol is weggelegd voor de overheid?

‘Past de definitie die de overheid hanteert voor een wegvak op de definitie die de automobielindustrie heeft van een wegvak? Dat loopt waarschijnlijk scheef. Want geautomatiseerde auto’s onderscheiden wegvakken in verband met verschillen in rijgedrag, en de overheid onderscheidt wegvakken in verband met de planning van beheer- en onderhoudswerkzaamheden.

Het Kadaster kan kaartenmakers helpen bij het opbouwen van een statisch wereldbeeld. Maar kaartenmakers hebben behoefte aan meer detail. Die extra nauwkeurigheid is niet alleen handig voor geautomatiseerd rijden, maar ook voor het beheer van de openbare ruimte. De afweging waar de overheid voor staat, is of ze zelf gaat investeren in de extra nauwkeurigheid, of dat ze het overlaat aan de markt.’ ‘Zeker als het gaat om statische informatie die ‘in advance’ beschikbaar is voor zelfrijdende auto’s, kan de overheid als leverancier een grote rol spelen. Andersom kan het voor wegbeheerders handig zijn om van autofabrikanten informatie af te nemen die met sensoren en camera’s ‘on the fly’ is verzameld, bijvoorbeeld over kuilen in de weg.

‘Ik zie een verschuiving in de rol van de overheid, van producent van kaartinformatie naar platform voor kaartdiensten, ook de betaalde diensten van kaartenmakers. De vraag is welke investeringen we moeten doen om die rol goed in te vullen? En we moeten nu alvast nadenken over de consequenties van de nieuwe verhoudingen. Gaat de overheid betalen voor de kaarten die ze voor haar eigen doelen nodig heeft?’ Stephen

T’Siobbel, Sr. Project Manager Advanced Driving bij TomTom Maps: ‘Ik denk niet dat TomTom ooit  kadasterkaarten gaat maken. Ik ga er van uit dat de overheid voldoende eigen use cases heeft om zelf topografische en kadastrale kaarten te beheren en te onderhouden. Net zomin als overheden kaarten zullen samenstellen die direct geschikt zijn voor private kaartenmakers, zullen kaartenmakers producten leverden die direct geschikt zijn voor overheden. Daarvoor zijn de verschillen te groot. Wel zullen onze kaarten voor Automated Driving over specifieke attributen beschikken die ook voor overheden relevant kunnen zijn, bijvoorbeeld informatie over het type van vangrails, of de exacte breedte van een rijstrook.’

70%

State of Art on Infrastructure for Automated Vehicles

Chapter 6

This  section  summarizes,  based  on  insights  from  the current scientific literature, projects, test sites, and  initiatives,  the  implications  of  vehicle  automation  on  the  infrastructure  for  each  SAE  level  of automation (in each case assuming 100% penetration level). According to Shladover (31) level 5 will not be here until 2075, while level 3 is problematic because of the difficulty to attain drivers’ attention after  being  out  of  the  loop  and  because  some  automakers  simply  will  not  attempt  level  3.  However, level 4 automation will probably be realized within the coming decade. In Table 6 a first attempt was made to summarize the requirements from the physical infrastructure to facilitate vehicle automation, followed by Table 7 which summarizes the requirements from the digital infrastructure. These results should  be  considered  with  caution,  as  many  of  the findings  from  the  scientific  literature  were  not explicitly based on empirical data and results, but on experts’ opinions.

 

IMPACT-Infrastructuur

90%

Rapport Zelfrijdende auto’s, verkenning van implicaties op het ontwerp van wegen

Chapter 3.3
– Als  begrenzing  van  rijstroken  is fysieke  markering belangrijk, naast de eventueel al toe te passen   digitale  markering.   Dat   geldt  voor   alle   wegonderdelen.   De   markering  moet   goed waarneembaar  zijn,  door  in-car  sensoren  (camera’s)  en door  de  menselijke  bestuurder,  bij verschillende weers- en lichtcondities.

-Naarmate  de  automatiseringsgraad  toeneemt,  zijn  er steeds  meer  voertuigen  die  op smallere rijstroken kunnen rijden. Echter bestuurders van voertuigen die nog niet geautomatiseerd koers houden, zijn gebaat bij de huidige rijstrookbreedte (vrees marge en vetergang). Rijstroken kunnen dus   nog   niet   overal   smaller   gemaakt   worden.   Als   tussenoplossing   kunnen smallere doelgroepstroken voor de hogere SAE level voertuigen worden geïntroduceerd.

-Aanpassingen van het dwarsprofiel en de berminrichting (redresseerstrook, obstakelvrije zone, vluchtstroken) kunnen nog niet plaats vinden.

-Boogstralen kunnen ook nog niet aangepast worden. Wel zou overwogen kunnen worden om in bogen  met meer  stroken  de manueel  bestuurde voertuig en  alleen  in  de  buitenste  strook/stroken toe  te  staan  en  de  binnenste  strook/stroken  te  reserveren voor  automatische voertuigen  die  hun snelheid   kunnen   optimaliseren   op   de   infrastructuurkenmerken   en   de   voorkeuren   van   de inzittenden.  Een  risico  is  dat  manueel  bestuurde  voertuigen  het  gedrag  van  automatische voertuigen gaan imiteren, wat ertoe zou kunnen leiden dat ze met een te hoge snelheid de bocht in gaan. Ook de korte volgtijden van automatische voertuigen zouden overigens door menselijke bestuurders geïmiteerd kunnen worden. Dat geldt ook op andere wegonderdelen.

-Over  het  algemeen  geldt  dat  de  mix  van  verschillende voertuigtypen  een  aanvankelijk  wat onvoorspelbaar  verkeersbeeld  kan  geven.  Dat  heeft  met name invloed op de dimensionering van  uitwisselpunten. (in- en uitvoeger, weefvak, kruispunt, rotonde). De mix van voertuigen van verschillende  SAE  levels  kan  zorgen  voor  interactie  tussen  de  verschillende  voertuigtypen  die tegen  de  intuïtie  van  menselijke  bestuurders  ingaat. ZRA’s  gedragen  zich  anders  dan  de bestuurders van niet-ZRA’s, of bestuurders van voertuigen met een lager SAE level, op basis van hun intuïtie verwachten. De uitwisseling op sommige plaatsen is mogelijk te complex voor ZRA’s, die   nog   niet   met   elkaar   communiceren   en   meer   tijd   nodig   hebben,   door   de   grotere veiligheidsmarges  dan  die  die menselijke  bestuurders aanhouden  en vroegtijdig  remmen.  Dit  zal ertoe   kunnen  leiden   dat   uitwisselpunten   (in-   en   uitvoegstroken,   weefvakken)   eerst   ruimer gedimensioneerd  moeten  worden.  Dat  sluit  aan  bij  de praktijkobservatie  dat  ACC in  zijn huidige vorm leidt tot grotere volgafstanden.

-Voor onderliggende wegen geldt dat met name de interactie  met  langzaam  verkeer  (fietsers  en voetgangers) veel dilemma’s oplevert. Daardoor wordt de situatie veel complexer en daarvoor zijn op  dit  moment  nog  geen  (veilige)  oplossingen  beschikbaar.  Op  gebiedsontsluitingswegen  met gescheiden verkeersstromen gelden dit ook (op kruispunten en rotondes).

-Bij gemengd verkeer dient nog vastgehouden te worden aan het originele kruispuntontwerp en voorrangsregels.  De  ZRA  moet  zich  zoveel  aan  de  menselijke  bestuurders  aanpassen,  zodat verwarring voorkomen wordt bij bestuurders van niet-automatische voertuigen. Mogelijk heeft het automatische voertuig meer tijd  nodig  om  de  situatie op  een  kruispunt  in te  schatten,  als  er  niet met alle voertuigen in de buurt gecommuniceerd kan worden.

-Doorzicht op een rotonde is voor een ZRA, die communiceert met het overige verkeer, geen probleem, voor menselijke bestuurders wel.

– (truck) platooning brengt nog de nodige vragen met zich mee, als het streven is om vrijwel continu te  kunnen  platoonen  (dus  niet  moeten  opsplitsen  bij ieder  knooppunt)  om  de  voordelen  vanplatooning te kunnen behalen

Uit  het  bovenstaande  ontstaat  het  beeld  dat  de  mogelijke  consequenties  voor  het  wegontwerp  in  de situatie  met  gemengd  verkeer  waarschijnlijk  nogal  beperkt  zullen  zijn  (ofwel  dat  je  niet  veel  kunt veranderen  aan  het  wegontwerp  zolang  er  gemengd  verkeer  is).  Bij  gemengd  verkeer  kan  er  in  eerste instantie niets veranderd worden aan het ontwerp, dat gebaseerd is op wat menselijke bestuurders nodig hebben  om  veilig,  vlot  en  comfortabel  te  rijden.  Alleen  op  wegen  met  veel  capaciteit/rijstroken  kan overwogen  worden  een  deel  hiervan  voor  ZRA’s  te  reserveren en dit deel ook een nieuw wegontwerp te geven (scheiding in het dwarsprofiel van ZRA en niet-ZRA).

Smart Infra, Eerste schetsonderzoek naar level 4 snelwegen en kruispunten voor zelfrijdende auto’s

Chapter 5

Voor de transitiefase geldt dat zowel de zelfrijdende voertuigen als de conventionele voertuigen gebruik maken van dezelfde rijbaan. Zolang de conventionele voertuigen gebruik maken van de rijbaan zullen deze, ten behoeve van de verkeersveiligheid, maatgevend z

ijn voor de ontwerpcriteria aan de rijbaan. In de transitiefase is het daarom naar verwachting niet wenselijk versoberingen aan het ontwerp van autosnelwegen door te voeren.

Bij een gescheiden transitie zal er een doelgroepenstrook aan de linkerzijde van de rijbaan worden aangewezen voor de zelfrijdende voertuigen. In het geval dat de doelgroepenstrook als extra strook wordt toegepast, zullen de overige stroken versmald moeten worden en wellicht dat de vluchtstrook moet worden opgeofferd. Beide maatregelen leiden tot een

afwijking van de vigerende richtlijn (ROA 2014) en hebben mogelijk een negatief effect op de

verkeersveiligheid van met name de conventionele voertuigen. Met de extra strook wordt de totale capaciteit van de weg wel vergroot.

 

“Automated driving, with its minimal space requirements and rather equal speed levels, could at least double the existing average road infrastructure capacity. “

Gevonden in (p.380): Autonomous Vehicles and Autonomous Driving in Freight Transport

 

70% State of Art on Infrastructure for Automated Vehicles

 (See figure in Chapter 6))

This  section  summarizes,  based  on  insights  from  the current scientific literature, projects, test sites, and  initiatives,  the  implications  of  vehicle  automation  on  the  infrastructure  for  each  SAE  level  of automation (in each case assuming 100% penetration level). According to Shladover (31) level 5 will not be here until 2075, while level 3 is problematic because of the difficulty to attain drivers’ attention after  being  out  of  the  loop  and  because  some  automakers  simply  will  not  attempt  level  3.  However, level 4 automation will probably be realized within the coming decade. In Table 6 a first attempt was made to summarize the requirements from the physical infrastructure to facilitate vehicle automation, followed by Table 7 which summarizes the requirements from the digital infrastructure. These results should  be  considered  with  caution,  as  many  of  the  findings  from  the  scientific  literature  were  not explicitly based on empirical data and results, but on experts’ opinions.

80%

Automated Vehicles. The Coming of the Next Disruptive Technology

Benefits page 17 report

With AVs, the demand for parking will decrease substantially because an AV can relocate itself to an area of free parking. Or, as an automated taxi, it can pick up its next ride. In some cases, a commuter can send the car home for his/her spouse to use.

 

Besides changes in parking spots Urban land use is expected to change as well with the implementation of AV’s. This Autonomous Driving and Urban Land Use report discusses the possible effects.

80% Driverless future A policy roadmap for city leaders Page 12 report AVs will reduce demand for parking, gas stations, and other auto-related land uses. Some uses, particularly those in highly desirable areas, may be reused and repurposed over time. AVs are highly likely to reduce parking demand by taking personally owned automobiles off the street. Past studies estimate that, depending on the success of merging AV into city infrastructure, parking demand may be reduced by up to 90%. Parking, roads and other auto-related uses occupy a significant amount of land. The U.S. contains as many as two billion parking spaces, occupying up to 16,000 square miles of land (the equivalent of Connecticut and Vermont combined). The quantity of parking spaces in the country amounts to as many as eight parking spaces for every car. Parking consumes a significant amount of land, especially in suburban areas where auto use is highest and surface lots are more common than multi-story garages. At a typical suburban mall, parking or driveways make up 80% of the land, while only 20% is used for the mall. Even in denser, more urban areas, parking requires significant land area. For example, streets and parking take up 45% of land in downtown Washington, D.C. and up to 65% in downtown Houston.

IMPACT-Veiligheid

60% 

Zoeken naar strepen op het asfalt

Introduction page 4 of document

Van de 33.000 verkeersslachtoffers die de VS jaarlijks betreurt, zijn er volgens deskundigen 22.000 te voorkomen als we de mens achter het stuur vandaan halen.

Tomorrow’s Road Infrastructure for Automated Driving

Slide 11

Point made by an online respondent of a survey:

“I am extremely concerned that proponents have little regard to or understanding of the level of reliability required to class any of these systems as safe . For example in regard to Google cars : ‘Ultimately, Google aims to provide a solution for the millions of car accidents that occur worldwide – 93 percent due to human error .’Statement is misleading/ wrong . Human factors contribute to 93 percent of crashes but many other factors also contribute. And the most responsible drivers cause a crash where someone is injured around once in 2,000,000 Miles. And public would expect autonomous cars to have a much lower rate-say once in 20,000,000miles.That requires a system that will not fail/malfunction more than once in ~ 80 vehicle lives or once in 1250 years of average driving.”

The Release of Autonomous Vehicles

Chapter 21.1

60%

ROAD SAFETY WITH SELF-DRIVING VEHICLES: GENERAL LIMITATIONS AND ROAD SHARING WITH CONVENTIONAL VEHICLES

Chapter 3/ Conclusion

Self-driving vehicles could compensate for some but not all crashes caused by other traffic participants (Pedestrian error could be compensated by AV). Lighting failures might turn out to be irrelevant to safety from the perspective of being able to control one’s vehicle at night, because self-driving vehicles might not rely on visual input. / (1) The expectation of zero fatalities with self-driving vehicles is not realistic. (2) It is not a foregone conclusion that a self-driving vehicle would ever perform more safely than an experienced, middle-aged driver. (3)During the transition period when conventional and self-driving vehicles would share the road, safety might actually worsen, at least for the conventional vehicles.

Safety Benefits of Automated Vehicles: Extended Findings from Accident Research for Development, Validation and Testing

17.4  Significance of Possible Predictions based on Accident Data

 

70%

Safety Benefits of Automated Vehicles: Extended Findings from Accident Research for Development, Validation and Testing
17.1 introduction

→ report goes into more detail
For testing methods in order to develop and validate safe automated vehicles with reasonable expenditure, the author recommends combining area-wide traffic, accident, weather, and vehicle operation data as well as traffic simulations. Based on these findings, a realistic evaluation of internationally and statistically relevant real world traffic scenarios as well as error processes and stochastic models can be analyzed (in combination with virtual tests in laboratories and driving simulators)to control critical driving situations in the future.

Self-Driving Regulation, Pro-Market Policies Key to Automated Vehicle Innovation
50%
One important challenge, which is expected to be met by late 2014 or early 2015 , is providing sufficient evidence that road – tested autonomous vehicles are in fact safer than manually driven vehicles. As Bryant Walker Smith of Stanford Law School has noted, a high degree of statistical confidence must be reached in order for automakers and component developers to begin scaling up technology deployment beyond testing.

Google’s self- driving cars have logged over 500,000 miles on U.S. public roads to date. To demonstrate their safety over manually driven vehicles with 99 percent confidence, Google will need to log approximately an additional 200,000 miles of crash-free automated driving (see Table 2).

70%

Automated Vehicles, Are we ready ?

Chapter 3.3

AVs are capable of providing large amounts of data that could assist investigation in case of a crash. By recording the actions and forces involved in the minutes before and after a crash, they may help determine the cause of the crash and assist in resolving any liability dispute.

Motoring of the future

Point 33, page 15 in report

Telematics also known as ‘black boxes’ monitor the location of a driver and driving performance.

Safety Benefits of Automated Vehicles: Extended Findings from Accident Research for Development, Validation and Testing

However, a safety prognosis of highly or fully automated vehicles depends on assumptions, as so far no series applications of such features exist. For testing methods in order to develop and validate safe automated vehicles with reasonable expenditure, the author recommends combining area-wide traffic, accident, weather, and vehicle operation data as well as traffic simulations.

Two questions discussed in paper:

–What significance do analyzes and findings from road-accident research hold for the introduction of automated vehicles?

–How can the potential safety benefits of automated vehicles be established?

The validity of accident data regarding potential safety benefits varies considerably depending on the collection method.

60%

The Release of Autonomous Vehicles

21.3  Requirements for a Test Concept

In order to discuss in the following section why full automation poses a particular challenge for safety validation, we will first describe the requirements for test concepts to assess safety. These are divided into effectiveness and efficiency criteria.

21.5.1 Validity of the current test concept for autonomous driving

At present, real driving is the most important method for the approval; the reason for this, in particular, is the validity combined with the justifiable economic overhead. However, along with the economic overhead, autonomous driving also presents a systematic challenge for the known methods. At present, real driving stands for driving in public road traffic with test drivers. The task of the test driver is to drive or supervise the vehicle in every situation in accordance with the task of the vehicle user. Transferred to autonomous driving, the use of a test driver in the driver’s seat would be non-real behavior of a user, as the user does not have to supervise the vehicle and the environment anymore and intervene.

Motoring of the future

Point 64, page 27 in report

50%

Witnesses discussed the research evidence for the effectiveness of different systems. Professor Sampson said that it was very difficult to research which technologies were most effective in terms of reducing accidents, because of the difficulties in running controlled trials of different features, with sufficient numbers of vehicles. Professor Carsten explained that the key struggle was with the continual monitoring and evaluation of technology, and developing an understanding of how casualty rates were affected over time by different technologies.

He explained that while it was statistically possible to show the safety benefits arising from car impact regulations, it was “really hard” to do this in relation to other safety approaches.

IMPACT-Verkeersafwikkeling

70%

The Effect of Autonomous Vehicles on Traffic

Chapter 16.4.2

The models developed for traffic flow and capacity, assuming a given share of autonomous vehicles, show that capacity increases disproportionately highly as the share of autonomous vehicles increases. It should be noted that the shortening of the time gaps comes into effect as early as the first autonomous vehicle; the speed increase at high densities, however, will only be possible for purely autonomous traffic. The introduction of autonomous vehicles will succeed, in the opinion of the author, only in their ability to move safely in mixed traffic, as reserved transit areas would not be socially or economically acceptable, particularly with a low share of autonomous traffic. However, once a sufficient number of vehicles with autonomous capabilities are participating in traffic, it will be very beneficial to the transport efficiency to create reserved lanes for autonomous driving. The benefits of autonomous vehicles can be maximized by separation due to the nonlinear course of the capacity once non autonomous vehicles are added to autonomous traffic. In conjunction with specially dedicated lanes, the column speed could also be increased even when traffic demand is higher, which would lead to further significant capacity gains. This is not possible in mixed traffic, since even in trafficwith only a few human-driven vehicles, these would dictate the speed.
50%

Autonomous Driving: Disruptive innovation that promises to change the automotive industry as we know it

Page 7 report

Looking forward, we project “Level 3: limited self-driving automation” to be available by 2018-2020 with features such as highway chauffeur (automated driving on highways). Furthermore, we expect “Level 4: full self-driving automation” to be first offered for low speed situations by 2020-25 (e.g., in parking lots or low-speed areas) and eventually, including more complex operations to be offered by 2025-30 (e.g., city driving). Even with the introduction of new technologies, we do not expect global adoption of full self-driving automation with “door-to-door” capabilities across all vehicle segments before 2030-40.

 

50%

Methodische Verkenning Zelfrijdende Auto’s  en  Bereikbaarheid
Chapter 2.2.1 (Capacity with and without bottlenecks mixed AV non AV)

Arnaout en Bowling (2011) vonden voor een weg met 4 rijstroken in een scenario met en zonder oprit dat CACC een positief effect op de capaciteit heeft (tot +60% bij een penetratiegraad van 100%) als de penetratiegraad groter is dan 40% en de instroom hoog genoeg is. Bij lagere penetratiegraden was het positieve effect klein. Als de instroom laag is (vrije doorstroming), vonden ze geen effect op de capaciteit. Ze veronderstelden dat CACC voertuigen een volgtijd van 0,5 seconde aanhouden als ze achter een ander CACC-voertuig rijden en 0,8 tot 1,0 seconde (uniform verdeeld) als ze achter een ander voertuig rijden. Of men in praktijk deze korte volgtijden durft aan te houden is een grote uitdaging volgens hen (Shladover, Su, & Lu, 2012).

De CACC-voertuigen kunnen hun voorliggers volgen zonder dat de bestuurder gas hoeft te geven of hoeft te remmen; de bestuurder moet wel het voertuig in de strook houden. Er werd een overbelaste snelweg in gemodelleerd, met een lengte van 6,5 km en een snelheidslimiet van 105 km/uur zonder bottleneck . In de simulaties waren vier voertuigtypen aanwezig:

50%

Methodische Verkenning Zelfrijdende Auto’s en Bereikbaarheid

Hoofdstuk 2.6

ACC kan zowel een klein negatief als een klein positief effect hebben op de capaciteit (~ -5% -+10%). Voor CACC rapporteren de meeste studies een kwadratische toename van de capaciteit als penetratiegraad toeneemt met een maximale theoretische toename van 100% (verdubbeling). ACC en CACC hebben een positief effect op de stabiliteit. Bij hogere penetratiegraden ontstaan minder schokgolven en isde duur van de schokgolven aanmerkelijk korter.

 

The Effect of Autonomous Vehicles on Traffic

16.3 Gives theory for why capacity increases of purely AV traffic

70%

Traffic Control and Traffic Management in a Transportation System with Autonomous Vehicles

Chapter 15.8 Conclusion

It was demonstrated in Sect.15.4, for example, that the capacity of a traffic signal can certainly be doubled. If the demand is low at the corresponding signal,this doubling is scarcely noticeable. But if the signal is working at the limits of its capacity, by contrast, even a minor increase in its capacity can lead to a dramatic

Improvement.

This can be observed quite clearly in the scenario in Sect.15.5: here the demand runs the values from very low to (temporary) over-saturation. Although the introduction of autonomous vehicles has little impact on green times and delays when demand is low, it yields major improvements when the system is operating beyond capacity. Nevertheless, the magnitude of these improvements does depend on the details of the scenario being examined. If the peak value for demand were just a bit lower, the benefit would also be significantly diminished. That notwithstanding, it may be asserted with confidence that at least in the urban context, the introduction of autonomous vehicles has the potential to generate substantial time gains at traffic signals which would then be available for other road users—if the introduction of these vehicles does not lead to an increase in demand for automotive transportation

“Autonomous driving could have a dramatic, albeit gradual, effect on each of the traffic concepts discussed in the previous section. Absent other phenomena, the total cost of motor vehicle travel is likely to decrease, and demand for that travel is likely to increase faster than corresponding capacity.”

Gevonden in (p.1409): Managing Autonomous Transportation Demand

 

LEGAL-Aansprakelijkheid

0%

Verschillende papers benoemen de relevantie van v2i-technologie in de context, maar nergens wordt gesuggereerd dat de wegbeheerder aansprakelijk kan zijn.

50%

In (Fundamental and Special Legal Questions for Autonomous Vehicles) wordt gesteld dat hier elastisch mee om moet worden gegaan. Dat wil zeggen dat handhaving alleen plaats moet vinden als dat bijdraagt aan de veiligheid.

25%

(Als de auto autonoom wordt.), (How Autonomous Vehicle Policy in California and Nevada Addresses Technological and Non-Technological Liabilities)

Er worden verschillende cyberrisico’s benoemd. Het kan bijvoorbeeld gaan om een gebrekkig product of een cyberaanval. Wie er aansprakelijk is, wordt niet besproken.

25%

(Autonomous Vehicle Liability—Application of Common Carrier Liability) geeft enkele voorbeelden, zoals de automatische piloot op vliegtuigen en schepen, en de eerste liften.

50%

In (Aansprakelijkheidsaspecten van zelfrijdende auto’s Een verkennende analyse) staat een citaat van Coelingh & Solyom:

“To join from the rear, a driver would send a request to the lead vehicle, get confirmation, approach the platoon from behind, and then put the car into semiautomatic mode, in which braking and accelerating is automatic and the steering is still manual. This ensures that the driver will pay full attention to traffic in case anything unusual happens. Only when the car is locked into the determined following distance does lateral control pass to the automatic system. An indication of the change appears on the car’s display, accompanied by a voice message, letting the driver know that he can release the steering wheel, lean back, and just enjoy the ride…..”

Dit impliceert dat de volgauto’s het voorste voertuig volgen, en dat voertuig het peloton dus bestuurt.

LEGAL-Juridische Aspecten

0%

Het antwoord hierop moet m.i. niet in het juridische, maar in het technische en het human factors domein gezocht worden.

25%

Uit (Automated Vehicles are probably legal in the United States): “The rules of the road as codified assume human judgment, and the rules of the road as observed reflect that judgment. These dependencies  may  complicate  the  lawful  operation  of  automated  vehicles.”

“These references to reasonableness, prudence, practicability,  and  due  care  demonstrate  that  the  law  accepts  risk  at  a  certain level, they neither specify this level nor prescribe a basis for determining  it. ”

“Application  of  these  laws  to  automated  vehicles  may  present both  design  challenges  and  liability  concerns.”

 

  • Welke strategieën voor regelgeving zijn er en welke hanteren we nu?

75%

Verschillende dilemma’s worden gepresenteerd in (Automated and Autonomous Driving: Regulation under uncertainty):

  1. Treat automated vehicles specifically or generally?
  2. Let policy lead or lag technology?
  3. Privilege uniformity or flexibility?
  4. Emphasise ex-ante or ex-post regulation?
  • Hoe verbreden we het perspectief van de regelgeving zonder het te groot en te complex te maken? (bv door een of meerdere frameworks adhv use cases, in specifieke deployment situaties).

0%

Dit is niet in de literatuur genoemd.

 

  • Welke relaties in de bestaande regelgeving zijn relevant voor Automatisch rijden en in welke regelsets, bv: Het voertuig (toelating), Het verkeer (veiligheid, doorstroming), De infrastructuur en het beheer daarvan in de openbare ruimte, de mobiliteit, transport van personen en of goederen, DAta en privicy, datainfrastructuren, (geo)informatiesystemen en informatiemanagement, eurpese regelgeving/harmonisatie

10%

(Regulation and the Risk of Inaction) benoemt de juridische complexiteit. Gesteld kan worden dat elk van de relaties relevant is.

25%

In (How Governments Can Promote Automated Driving) is een hoofdstuk gewijd aan het vergroten van de maatschappelijke acceptatie. In Amerikaanse staten waar geëxperimenteerd wordt met automatische voertuigen doet de overheid dit actief. In de literatuur wordt het risico dat de burger ervaart, niet genoemd.

0%

Het antwoord hierop moet m.i. niet in het juridische, maar in het technische en het human factors domein gezocht worden.

75%

(Zelfrijdende auto’s en het Verdrag van Wenen inzake het wegverkeer Een verkennende analyse) en (Automated Vehicles are probably legal in the United States)

Nationale wetgeving moet internationale verdragen volgen. De interpretatie daarvan heeft invloed op de implicaties.

LEGAL-Privacy

50%

Uit (Opportunities and Risks Associated with Collecting and Making Usable Additional Data):

“Limiting access rights and encryption are typical instruments of information security. Limiting access rights is also mentioned under the “Information Security” principle of ISO/IEC 29100. It follows the concept of “need-to-know”, limiting access to PII to those individuals who require such access to perform their duties, and limiting the access of those individuals to only that PII which they require to perform their duties. Access rights can be defined by defining exactly which entity can access which PII. This asks for a fine-grained specification of the system, and can best be achieved if privacy is already considered during the design phase, e.g. when designing which data are collected by the vehicle and for which application they are needed.”

Category: LEGAL-Privacy

75%

Dit is niet in wetgeving vastgelegd, maar afhankelijk van de overeenkomsten die een consument met bijvoorbeeld de autofabrikant en de verzekeraar sluit. Voorbeelden hiervan staan in (Opportunities and Risks Associated with Collecting and Making Usable Additional Data)

Category: LEGAL-Privacy

0%

Dit is niet in de literatuur genoemd.

Category: LEGAL-Privacy

90%

Volgens (The connected car. Who is in the driverseat?) is er een groot privacyrisico verbonden aan event data recorders. De wetgever kan beperkingen opleggen aan fabrikanten en verzekeraars. In de VS is er al wetgeving die consumenten beschermt.

Category: LEGAL-Privacy

25%

Camerabeelden worden niet separaat besproken. Algemene inzichten over data zouden ook op camerabeelden toegepast kunnen worden.

Category: LEGAL-Privacy

LEGAL-Toelatingseisen

100%

Uit (Societal Risk Constellations for Autonomous Driving. Analysis, Historical Context and Assessment): “Risk management must be adapted to the respective risk constellations and be conducted on the appropriate levels (public debates, legal regulations, politically legitimated regulation, business decisions, etc.). It is based on the description of the risk constellation, in-depth risk analyses in the respective fields and a societal risk assessment.”

0%

Dit aspect is niet genoemd in de literatuur.

50%

(Motoring of the future) De eisen aan voertuigen moeten de stand van de techniek volgen. Dit maakt het moeilijk om vandaag een framework op te zetten voor het voertuig van morgen.

75%

Uit: (The pathway to driverless cars, a code of practise for testing) “It is expected that automated driving systems will rely on the interaction and correct operation of several computers and electronic control modules. It will be important that:

  • Software levels and revisions running on each vehicle to be tested are clearly documented and recorded.
  • All software and revisions have been subjected to extensive and well documented testing. This should typically start with bench testing and simulation, before moving to testing on a closed test track or private road. Only then should tests be conducted on public roads or other public places.”

Uit: (Products Liability and Driverless Cars): ” Common sense would hold that, if an original manufacturer in no way participates in or promotes the post-sale installation of autonomous vehicle technology manufactured by a third party, the original manufacturer should not be liable for alleged defects in that technology. Unfortunately, some of the case law relating to liability for third-party conversions in other contexts doesn’t necessarily support this common sense conclusion.”

0%

De rechtspositie van de overheid bij defectieve automatische voertuigen is niet benoemd.

TECHNICAL-(Cyber)Security

90% Adviesrapport Cybersecurity Autonoom rijdende voertuigen , Fox IT (2014)

Dit rapport geeft inzicht in kwetsbaarheden van moderne voertuigen, risico’s bij introductie van autonoom rijdende voertuigen en mogelijke maatregelen om deze risico’s tegen te gaan.
Ondanks de dreiging die klein is, is het toch ontzettend belangrijk om goede maatregelen te nemen.Voor de korte termijn is een zestal generieke maatregelen voorgesteld waarmee het op korte termijn mogelijk is een beeld te krijgen van de beveiliging van systemen voor autonoom rijdende auto’s, daarmee van de veiligheid van het rijden erin en de beveiliging te verbeteren. Deze 6 maatregelen richten zich op het onderzoeken van autonoom rijdende voertuigen op kwetsbaarheden, deze op te lossen en op het klaar zijn voor een aanval
middels een gevonden kwetsbaarheid. Gezien de verwachting dat cybercrime voor autonoom rijdende voertuigen sterk zal groeien en dat niet alle problemen van tevoren te voorspellen zijn, wordt voor de lange termijn geadviseerd te streven naar een framework waarin maatregelen zich kunnen ontwikkelen.

60% Comprehensive Experimental Analyses of Automotive Attack Surfaces

Geeft inzicht in manieren van digitale inbraak van autos.
‘’We discover that remote exploitation is feasible via a broad range of attack vectors (including mechanics tools, CD players, Bluetooth and cellular radio), and further, that wireless communications channels allow long distance vehicle control, location tracking, in-cabin audio exfiltration and theft.’’

50% Security Challenges for Cooperative andInterconnected Mobility Systems

Identificeert en kwantificeert risico’s van Interconnected Mobility Systems.

‘’The biggest security risk factors foreseen are application data integrity validation, the usage of insecure position information and systems that are currently not secure by design. These risk factors will have to be addressed in the coming years, to pave the road for successful introduction of cooperative and interconnected mobility systems’’

TECHNICAL-Architectuur

60%

Model Based Vehicle Detection and Tracking for Autonomous Urban Driving

This paper describes the moving vehicle detection and tracking module that we developed for our autonomous driving robot Junior. The module provides reliable detection and tracking of moving vehicles from a high-speed moving platform using laser range finders. The paper presents the notion of motion evidence, which allows us to overcome the low signal-to-noise ratio that arises during rapid detection of moving vehicles in noisy urban environments.

100% Ja dat kan; 

Learning to Drive: Perception for Autonomous Cars Chapter 3 smooth road detection

‘’Accurate perception is a principal challenge of autonomous off-road driving. Perceptive technologies generally focus on obstacle avoidance. However, at high speed, terrain roughness is also important to control shock the vehicle experiences. The accuracy required to detect rough terrain is significantly greater than that necessary for obstacle avoidance. self-supervised learning approach for estimating the roughness of outdoor terrain. Our main application is the detection of small discontinuities that are likely to create significant shock for a high-speed robotic vehicle. By slowing, the vehicle can reduce the shock it experiences. Estimating roughness demands the detection of very small surface discontinuities – often a few centimeters. Thus the problem is significantly more challenging than finding obstacles.’’ Experimental results from this paper show that speed control – reduction  used offers significant improvement in shock  detection.

80% Self-supervised Road Detection in Desert Terrain

The paper presents  a  method  for  identifying  drivable  surfaces   in   difficult   unpaved   and   offroad   terrain   conditions   as encountered in the DARPA Grand Challenge robot race. Instead of relying on a static, pre-computed road appearance model, this method  adjusts  its  model  to  changing   environments.  It  achieves robustness  by  combining  sensor  information  from  a  laser  range finder,  a  pose  estimation  system  and  a  color  camera.  Using  the first  two  modalities,  the  system  first  identifies  a  nearby  patch of  drivable  surface.  Computer  Vision  then  takes  this  patch  and uses  it  to  construct  appearance  models  to  find  drivable  surface outward  into  the  far  range.  This  information  is  put  into  a  drivability  map  for  the  vehicle  path  planner.

50%

Do Autonomous Vehicles Learn?

Self learning capabilities of autonomous vehicles. The paper discusses discuss why, whether, and with which challenges and approaches machine learning is possible in its current form in autonomous driving. The paper portrays the view of vehicle technology in particular on this question, and is based on experience from the literature for the area of machine learning.