Written evidence submitted by Dr Julian Estevez (SDV0017)
Dr. Julian Estevez. Computational Intelligence Research Group. University of the Basque Country (UPV/EHU)
This is the report based on my knowledge and opinions about the technology of self-driving vehicles. As suggested in the Call, I will organize my ideas around the different list of aspects that the Transport Select Committee refers:
Likely uses, including private cars, public transport and commercial vehicles / and potential effects on patterns of car ownership, vehicle taxation and decarbonisation in the car market.
I think that self-driving vehicles should not be thought that they will replace one by one all the private cars, and keep the same private property schema as today [1]. On the one hand, an autonomous vehicle will be far more expensive than a normal car [2], and on the other hand, the persistence on almost one-car-per-person would inevitably still provoke traffic jams, accidents, and pollution.
Most of the utopian visions of self-driving cars assume that they will be shared, rather than owned privately. This would be a more sustainable option, and governments should take advantage of this technology to increase public transport around big cities and create more flexible and adaptive routes [3].
However, it is difficult to make hard statements on the reduction of pollution with the combined usage of autonomous and not autonomous cars [4], as it depends on the new usage habits of the users. States should lead their efforts to take advantage of the autonomous driving technology to reduce the total rides in the whole country and enhance the public transport.
From the point of commercial vehicles, the most promising application is goods delivery. At this moment, some self-driving trucks are performing different tests, both in America [5] and in Europe [6]. In 2011, Tesco supermarkets implemented virtual stores in South Korea, which made that people could buy online their products in the underground and reduced traffic jam considerably. According to research studies, 40% of the car rides in Britain are used to going to the supermarket [7].
Progress of research and trials in the UK and abroad
The most advanced country in self-driving vehicle technology is United States. While the US NHTSA has offered guidance to the autonomous vehicle industry, its recommendations have so far been voluntary, and the Federal Motor Vehicle Safety Standards have not been updated yet in order to reflect new technology regulations. Several US states have passed laws specifying conditions for the testing of self-driving vehicles.
In the case of UK, self-driving cars faster advance is restricted due to the difficulties of regulation. As [8] states, “lawmakers haven’t created a good enough set of rules for accountability when it all goes wrong”.
However, all over the Globe, researchers and industry seem to have reduced the big hype and enthusiasm for the rapid achievement of self-driving technology. In the beginnings of the last decade, deep-learning revealed as a great tool to extract working pattern out of data. There was a great enthusiasm due to the big success of Deep Mind, Alpha Go, Alpha Zero and artificial intelligence based on the collection of huge datasets that elaborated some restricted and standard rules to learn how to play StarCraft, Go or Chess. However, self-driving technology revealed to be one of the biggest problem in artificial intelligence, and a highly chaotic scenario where it is almost impossible to test all the possible traffic situations.
According to [9], $16 billion dollars have been spent on this technology, and the article wonders what is the real step ahead that the industry made. Most of car manufacturers recognized that the deadline for the upcoming of this type of vehicles will delay much more than they initially expected, or that even they could never achieve it [9].
Toyota and Honda manufacturers have made the bid for a totally different approach towards the development of autonomous cars, based on ADAS (Advanced Driver Assistance Systems). They call it Augmented Driving.
Cars have been equipped for some years with safety systems called ADAS, which assist the driving process and give extra information to the driver. For instance, they say whether there is a cyclist in the dead angle of the car, or they read the road traffic signals, suggest the most suitable gear, or make the steering wheel vibrate when there is a not indicated lane change [11].
This method only activates when the driver makes a mistake or creates a dangerous traffic situation. It requires the attention of the driver on the road, while preventing accidents. Moreover, drivers might lose some motor-skills in driving after using for long periods totally autonomous cars [12], while in augmented driving probably might not.
Potential implications for infrastructure, both physical and digital
The upcoming of self-driving vehicles would have a huge impact on many aspects of infrastructure. Roads should be better sensorized in order to communicate with the vehicles and notify them about traffic jams, near accidents, road conditions, or any other situation. Moreover, all the accidents liability should be clearly specified, and thus, vast amounts of data should be collected (in the car or in the road) and transmitted to insurances, entertainment and navigation companies.
However, this fact presents an intense legal debate about the ownership of the data, anonymization and aggregation, and drivers should give specific consent for the flow of that data [10].
Our cities might need to re-arrange. Depending on the level of autonomy of self-driving cars, the most appropriate solution might be to separate some city areas where only autonomous cars can circulate, but not cyclists and pedestrians, as they might ruin the traffic flow. That would not be an easy mission for city governors, as people would not like to easily give up their cherished walkability.
The regulatory framework, including legal status and approval and authorisation processes / The role of Government and other responsible bodies, such as National Highways and local authorities
There is a wide heterogeneity about autonomous driving regulations along the Globe. Undoubtedly, the country that goes beyond the rest is the United States. The U.S. Department of Transportation released an update to the agency's policy regarding autonomous vehicles on January 11, 2021, the Automated Vehicles Comprehensive Plan [13], which attempts to prepare the transportation system for the upcoming of self-driving transport. However, regulation in this country heavily depends on the states, and 10 years before the Plan, in 2011 Nevada state introduced a law focused on this technology for first time in history. However, there is no vehicle currently available for sale that is fully automated.
Every vehicle currently for sale in the United States requires the full attention of the driver during all the driving process [14]. There is an increasing number of vehicles now that offer some automated features designed to assist the driver under specific conditions, but these vehicles are not fully automated. Actually, NHTSA (National Highway Traffic Safety Administration) have drawn attention various times to Tesla car manufacturer about the tricky marketing of their famous autopilot. Now, NHTSA just considers Level 2 autonomy cars for consumer sale in the country. Currently, states permit a limited number of “self-driving” vehicles to conduct testing, research, and pilot programs.
These tests take place in limited and restricted conditions and roads, and NHTSA is responsible for the safe testing and deployment of these systems under their Standing General Order. Moreover, there are big differences between the 34 different states that enact any kind of legislation, depending on the requirements to have an operator on the vehicle, or require liability insurance [15].
Another big issue is the regulation of car liability in case of accidents. Who is the main responsible? The death of Elaine Herzberg (August 2, 1968 – March 18, 2018) was the first recorded case of a pedestrian fatality involving a self-driving car, in this case, developed by Uber. The trial is being extremely complex, and the Uber car driver, Rafaela Vasquez, was found guilty of the accident, as she was distracted streaming a TV show. However, everything is not said on this liability regulation, and in 2022 a Tesla driver who was behind the wheel with autopilot engaged when his vehicle crashed and killed two people. Now, the man is charged with manslaughter [16].
On the other hand, legislation also affects closely to autonomous delivery and self-driving trucks. These rules also depend on individual states, but they tend to be laxer than consumer driving. For this reason, some commercial brands already got permission to carry out totally autonomous 7-mile delivery with autonomous trucks in Arkansas [17].
Europe is also advancing on autonomous car legislation, and Germany is considered to be the first country to develop a legal framework on autonomous driving. In February 2021, this country discussed a bill on autonomous driving, specifically addressing level four of autonomous driving. The act finally passed in July 2021, titled Act Amending the Road Traffic Act and the Compulsory Insurance Act, and it caught the attention of all the agents in self-driving world due to its focus on ethics, such as human dignity, the right to life and the right to non-discrimination. For this purpose, the act addresses the very definition of autonomous driving, general division of responsibilities and assignment of tasks as mandated by legislators and precautions taken to prevent accidents. Thus, Germany welcomes the usage of full autonomous cars on their roads, but let also restricted conditions for that, which has been widely criticized by associations (i.e., the obligation to keep the car wirelessly connected at any moment).
Parallel to this regulation, European Commission is planning to make a step forward, and in July 2022 they announced the enhancement of General Safety Regulation that existed on Europe, which permitted the tests in very restricted conditions of autonomous cars. However, by the end of September 2022, they plan to allow each country to approve the registration and sale of up to 1,500 autonomous vehicles each year [18]. In this process, European Commission passed on April 2019 the “Guidelines on the exemption procedure for the EU approval of automated vehicles”, which aims to “harmonize the practice of Member States for the national ad-hoc assessment of automated vehicles and to streamline the mutual recognition of such assessment, as well as to ensure fair competition and transparency”. These guidelines already include some interesting points about the European vision about full autonomous driving vehicles, based on ethics, sustainability, transparency and safety.
The plans of European Commission on this field is to count on 2024-26 with L1 or L2 level of autonomy in all new sold cars on the continent [19].
Thus, most developed European countries are already preparing for enacting their own regulations on this type of technology, as already described for Germany. Another particularly interesting example is France.
In July 2021, France became the first European country that adapted its road and traffic regulations (Highway Code and Transport Code) to allow the operation of fully automated vehicles on public roads through the decree published by the French government [20], which amends road regulations to authorize testing of automated driving vehicles on public French roads. This decree will allow from September 2022, on limited and previously planned roads, vehicles equipped with delegated driving systems and automated road transport systems.
The decree introduces important changes in transport legislation, such as liability. Their view is that an accident will not be the driver’s fault as soon as the automated driving system operates in accordance with its conditions of use. Another newly defined feature by the decree is the terms of interaction between the driver and the automated driving system, as well as the emergency manoeuvres that the system may need to perform automatically. Finally, the last relevant point in the decree is that it regulates and identifies level of attention expected from the driver when an autonomous driving system is activated.
Similarly, the UK government announced [21] in April 2021 that by the end of that year “hands-free” automated vehicles offering automated lane-keeping systems will be legal to operate on UK roads. Previously described Germany legislation, ruled that will allow companies to deploy robotaxis and driverless delivery services on public roads by 2022.
Finally, another technology leader in the world of autonomous driving is China. On 24th March 2021, the Ministry of Public Security of China issued the Draft Proposed Amendments of the Road Traffic Safety Law [22] (the “MPS Proposed Amendments”). The MPS Proposed Amendments clarify the requirements related to road testing of vehicles equipped with automated driving functions, as well as regulating how liability for traffic violations and accidents will be ruled.
China let quite open the liability for car accidents, as the regulation states that when vehicles equipped with automated driving and manual operation modes are involved in road traffic accidents, the responsibility of the driver or the automated driving system developer shall be determined in accordance with laws, as well as the liability for damage. But for vehicles that are equipped with automated driving functions but lack manual operation modes, this liability issue be separately dealt with by relevant departments of the State Council.
In August 2021, the country updated the regulation, and permitted qualified firms to conduct trials of autonomous vehicles used for transporting passengers and goods on highways and city roads. Driverless cars that are equipped with real-time remote monitoring systems and can record and store driving data for at least 90 seconds before any crash or system failure are eligible to apply for the trial. This law came into force in March 2022. At this moment, also Chinese autonomous driving standards will be effective, which are different from SAE’s.
SAE launched its classification in 2014 and has updated it several times since [23]. It also has a six-tier rating based on human driver engagement and automation. Analysts state that Chinese standard put more emphasis on technology input. For instance, China's L0, L1 and L2 levels require the driver and the autonomous driving system to cooperate in detecting and responding to objects and events while the SAE version only requires the driver to perform tasks for these levels [24].
As a measure to enhance and boost the race of autonomous driving with the rest of the world, Shenzen Government announced on July 2022, that they will allow L3 autonomy vehicles from August of that same year [25].
Safety and perceptions of safety, including the relationship with other road users such as pedestrians, cyclists and conventionally driven vehicles
As previously mentioned in this document, humans might be one of the biggest causes of self-driving vehicles accidents [26] and lead to struggle the chaotic conditions of traffic. As a consequence of that, a rearrangement of city areas for pedestrians, cars and cyclist might be needed.
However, there are still many cognitive and psychological aspects to solve. First, to some people, it is insanity to think that they will just let their hands off their steering and let the car do the driving for them.
A recent research by the American Automobile Association (AAA) says that just 22% of Americans actually feel at ease riding in a self-driving car. But most of their fears are hinged on the fact that the technology is still under-explored [27]. Obviously, in Europe the same reaction was perceived [28].
Second, a human behaviour dilemma might take place: autonomous cars are supposed to save human lives in the long term, and self-driving cars need to be allowed on the road as soon as they are safer than the average driver. That presents an immediate problem: These autonomous vehicles would have many accidents and media coverage of these accidents could lead the public to doubt their safety. But this begs another question: Who would want to buy one of these cars? [29]
Third, despite its many benefits, cognitive offloading when using self-driving vehicles also introduces a new set of problems. When we offload activities, we also offload learning and judgment. In one study [30], researchers showed that self-driving cars drivers may see their driving skills degrade over time, which leads to forgetfulness, overconfidence and lose of ability.
Fourth, science showed that pedestrians behaviour when interacting with autonomous cars changes. Pedestrians tend to be much more dubious and hesitant when crossing the roads, for instance. Thus, more research in this field is required [31].
August 2022
REFERENCES
[1] de Almeida Correia, Gonçalo Homem, and Bart van Arem. "Solving the User Optimum Privately Owned Automated Vehicles Assignment Problem (UO-POAVAP): A model to explore the impacts of self-driving vehicles on urban mobility." Transportation Research Part B: Methodological 87 (2016): 64-88.
[2] https://www.bbc.com/future/article/20211126-how-driverless-cars-will-change-our-world
[3] Acheampong, Ransford A., and Federico Cugurullo. "Capturing the behavioural determinants behind the adoption of autonomous vehicles: Conceptual frameworks and measurement models to predict public transport, sharing and ownership trends of self-driving cars." Transportation Research part F: Traffic Psychology and Behaviour 62 (2019): 349-375.
[4] Silva, Óscar, et al. "Environmental impacts of autonomous vehicles: A review of the scientific literature." Science of The Total Environment (2022): 154615.
[5] https://www.cnbc.com/2021/11/08/walmart-is-using-fully-driverless-trucks-to-ramp-up-its-online-grocery-business.html
[6] https://www.eenewseurope.com/en/first-self-driving-trucks-on-european-motorway/
[7] Cairns, Sally. "Delivering supermarket shopping: more or less traffic?" Transport Reviews 25.1 (2005): 51-84.
[8] https://www.carmagazine.co.uk/autonomous/are-self-driving-cars-legal-in-the-uk/
[9] https://www.thedrive.com/tech/31816/key-volkswagen-exec-admits-level-5-autonomous-cars-may-never-happen
[10] https://themarkup.org/the-breakdown/2022/07/27/who-is-collecting-data-from-your-car
[11] Lindemann, Patrick, Niklas Müller, and Gerhard Rigolll. "Exploring the use of augmented reality interfaces for driver assistance in short-notice takeovers." 2019 IEEE Intelligent Vehicles Symposium (IV). IEEE, 2019.
[12] Trösterer, Sandra, et al. "What we can learn from pilots for handovers and (de) skilling in semi-autonomous driving: An interview study." Proceedings of the 9th international conference on automotive user interfaces and interactive vehicular applications. 2017.
[13] https://www.transportation.gov/sites/dot.gov/files/2021-01/USDOT_AVCP.pdf
[14] https://www.nhtsa.gov/technology-innovation/automated-vehicles-safety
[15] https://www.iihs.org/topics/advanced-driver-assistance/autonomous-vehicle-laws
[16] https://www.businessinsider.com/driver-who-had-tesla-on-autopilot-in-crash-manslaughter-trial-2022-5
[17] https://www.simmons-simmons.com/en/publications/cklcdtylu2wtt0970pwjocnti/new-german-draft-law-on-autonomous-driving
[18] https://www.politico.eu/article/eu-plans-to-approve-sales-of-fully-self-driving-cars/
[19] https://www.connectedautomateddriving.eu/wp-content/uploads/2021/04/EUCAD2021_D3_BO8_Maria-Cristina-Galassi.pdf
[20] French Decree on automated vehicles’ conditions of use and automated road transport systems’ commissioning
[21] https://www.bbc.com/news/technology-56906145
[22] https://www.chinalawinsight.com/2021/04/articles/corporate-ma/chinas-legislation-on-autonomous-cars-rolls-out/
[23] https://www.sae.org/blog/sae-j3016-update
[24] https://voi.id/en/technology/86943/china-sets-national-standard-for-autonomous-cars-l5-still-away
[25] https://www.arenaev.com/shenzhen_is_the_first_city_to_allow_l3_autonomous
_driving_from_august_1-news-455.php
[26] https://fortune.com/2018/08/29/self-driving-car-accidents/
[28] https://ec.europa.eu/research-and-innovation/en/horizon-magazine/trusting-self-driving-cars-going-be-big-step-people
[29] https://thereader.mitpress.mit.edu/whos-afraid-of-driverless-cars/
[30] Javadi, Amir-Homayoun, et al. "Hippocampal and prefrontal processing of network topology to simulate the future." Nature communications 8.1 (2017): 1-11.
[31] Rasouli, Amir, and John K. Tsotsos. "Autonomous vehicles that interact with pedestrians: A survey of theory and practice." IEEE Transactions On Intelligent Transportation Systems 21.3 (2019): 900-918.