Response to the draft implementing act on certain automated driving systems (ADS)

The European Commission published a draft of its upcoming implementing act setting out the technical requirements and test procedures for the type approval of certain vehicles with an automated driving system (ADS). The rules would apply to the following use cases:

• ‘Self-driving shuttles’, which are designed to carry passengers or goods on a predefined area in an urban or suburban environment;
• ‘Hub-to-Hub’ transport, which covers fully automated vehicles or dual mode vehicles that carry passengers or goods on a predefined route with fixed start and end points, and which may include urban or suburban or motorway environments; and
• ‘Automated valet parking’, which are vehicles that have a fully automated driving mode for parking applications.

ETSC has expressed concerns about several of the draft act’s provisions, which can be read in more detail by downloading the document on the ETSC comments here. In short, the concerns include:

• The interaction with different types of vulnerable road users in urban environments is not sufficiently covered by the draft act. Most of the tests that should be conducted to ensure compliance with the safety provisions feature primarily vehicle-to-vehicle scenarios. While pedestrians and cyclists are occasionally explicitly considered as well, they are not taken into account sufficiently. Moreover, the act fails to systematically include other road users that are commonly encountered in urban areas, such as wheel chair users and e-scooters, amongst many others. The lack of validation of the automated vehicle’s performance in situations when such other types of road users are encountered is from ETSC’s point of view insufficient to guarantee the operational safety of the vehicles in urban environments.
• It remains unclear how type approval authorities or their technical services would validate compliance with the performance requirements in the context of vehicles designed to also operate in other countries, and in particularly their compliance with the other country’s/countries’ traffic rules.
• While anticipatory behaviour by the vehicle/system is required, which is warmly welcomed by ETSC, the testing provisions do not guarantee that compliance with this important requirement is verified. The lack of provisions/requirements for the real world testing aspect, as well as the lack of a dedicated test on this topic, means that it is not assured that the vehicle will be assessed based on its anticipatory behaviour during the physical testing process, as there is no guarantee that the vehicle will encounter a scenario that would allow for such assessment.
• Contrary to the current provisions, ETSC considers that as a principle, acceptable safety levels should be set by regulators, and it should be up to manufacturers to demonstrate their compliance with it. 

ETSC followed-up on the public consultation with a letter co-signed by the European Cyclists’ Federation (ECF) and the International Federation of Pedestrians (IFP), which can be downloaded here. In the letter, the three organisations set out more details and provide supporting data for their concern that the velocities used in the draft implementing act regarding the maximum speeds at which pedestrians and cyclists can travel in the context of the tests on assessing compliance with the collision avoidance requirements, are set too low and are unrealistic. In short:

• The values are based on often quoted average figures. However, using average figures as maximum values represents a logical error, as this means that automated vehicles only have to avoid collisions with that half of the population walking or riding at or below the average speed. It therefore leaves the other half of the population unprotected, even though they are walking and riding at their normal pace.
• The often quoted average figures of 5 km/h for pedestrians and 15 km/h for cyclists are not representative for the average speeds found in Europe. For example, a study from the Netherlands showed that the average speed of cyclists in urban areas was 17.3 km/h instead of 15 km/h. And to reiterate, being an average, it is therefore known that half of the Dutch population travels at greater speeds while riding their bicycles normally.
• The average values moreover do not take into account the speeds of pedelecs and speed pedelecs (respectively 20.1 km/h and 22.2 km/h on average in Dutch urban areas), nor their maximum speeds (45 km/h for the speed pedelecs) which they can reasonably be expected to travel at. The values also do not take into account people jogging or (recreational) running – which are very common activities in urban areas, and should therefore be accounted for as well.

The organisations therefore called on the European Commission to increase the values for the speeds of pedestrians and cyclists, to better take into account the speeds at which they can reasonably be expected to be encountered in urban areas.

Image attribution: Autonomous Shuttle Pilot Project by Paul Kimo McGregor, available here on flickr.com. Used unaltered in accordance with the requirements of the CC BY-NC-SA 2.0 license.