Per­cep­tion of elec­tric vehic­les with AVAS

The number of electric vehicles is increasing. They are quieter than vehicles with internal combustion engines and may be harder for other road users to perceive.

Since 2021, all newly registered electric vehicles in the EU have to be equipped with an Acoustic Vehicle Alerting System (AVAS) which emits an acoustic warning. The legal provisions are laid down in the EU Regulation No 540/2014 and its amendments as well as the UN ECE Regulation 138. The AVAS needs to be e.g.: 

• in operation up to a speed of 20 kilometres per hour,
• acoustically similar to an internal combustion engine of the same vehicle category and 
• indicative for vehicle behaviour, e.g. through variations in sound levels or characteristics that are synchronized with the vehicle speed.

This is why the German Insurers’ Accident Research (UDV) commissioned a research study to assess how well electric vehicles with or without an AVAS are perceived by pedestrians compared to conventional vehicles with an internal combustion engine. In a virtual reality simulation of a street crossing situation participants were presented with different spatial vehicle sounds. 

• 1st experiment: In a street-crossing scenario, the participants estimated the time until the vehicles would reach their own position (contact time).
• 2nd experiment: In a street-crossing scenario, the participants decided whether or not they would cross the street before the arrival of the approaching vehicle (crossing decisions).
• 3rd experiment: The participants were asked to indicate whether the vehicle was accelerating or traveling at a constant speed (acceleration detection).

When the vehicles were driving at constant speeds, there were hardly any differences in any of the three experiments between electric vehicles with and without AVAS and the vehicle with an internal combustion engine.
When the vehicles were accelerating the contact time of the electric vehicles was estimated to be longer than it actually was, especially at high accelerations. The estimates were better for the vehicle with AVAS but still did not reach the level of the internal combustion engine vehicle, whose contact time was estimated best.
With regard to crossing decisions in the acceleration condition, higher collision probabilities were found for the electric vehicles (with and without AVAS) than for the vehicle with an internal combustion engine, especially at high accelerations. AVAS again mitigated this effect slightly, but decisions were still riskier than for the internal combustion engine vehicle. 

The acceleration of electric vehicles without AVAS was not well detected. With AVAS, acceleration was detected significantly better, but not as well as for the internal combustion engine vehicle.
Accelerating electric vehicles do not give pedestrians sufficient auditory information to permit an accurate estimation of vehicle travel. An AVAS improves the perception of electric vehicles without reaching the level of a vehicle with an internal combustion engine. It is therefore necessary to further develop AVAS solutions as well as the regulation itself within the current AVAS minimum and maximum overall sound pressure levels (in dB(A)).

• The AVAS regulation should include acceleration as an indicator in addition to frequency shifts as indicators of speed changes. It appears that such frequency shifts provide inadequate information about actual acceleration. Pedestrians use this information for their decision-making and are at risk if they fail to do so correctly. 
• Even above speeds of 20 kilometres per hour, the travel behaviour of electric vehicles with and without AVAS is estimated worse than that of an internal combustion engine. Therefore, the speed range in the AVAS legislation should be extended.
• The design of AVAS solutions should be optimized for acceleration estimations, especially in road-crossing situations and not only for acceleration detection. 
• The AVAS test protocols should be further developed to include acceleration.