Automated vehicles (AVs) are developing fast, and we need to understand how to integrate these vehicles into our roads, because they are likely to be involved in interactions with other road users, such as pedestrians and cyclists (Vulnerable Road Users or VRUs). When considering a two-way interaction between the AV and VRUs, the key challenge is ensuring that they are safe and acceptable to human road users, and do not affect the normal flow of traffic.
One part of addressing this challenge is providing AVs with the know-how to recognise and anticipate subtle behaviours and intentions of VRUs, while the other part is to understand how the AV should communicate its own intentions and behaviours. To this end, our researchers from the Human Factors & Safety group at ITS are conducting a series of studies, supported by the InterACT project, to understand pedestrians’ road-crossing behaviour.
In the first study, researchers used a Wireless Head Mounted Display (WHMD; see picture left) to investigate pedestrians’ gap acceptance, when crossing a virtual road, which included a number of vehicles approaching at different speeds. Initiation time for the crossing and the time it took pedestrians to cross the road was compared for different approaching speeds and deceleration profiles of the vehicles.
The second study is investigating the extent to which pedestrians note driver presence and distraction when crossing the road in a virtual environment, and whether this affects their crossing behaviour.
These studies will help build a picture of how drivers and other road users interact and communicate in a conventional setting, and this information will then be used as a basis to develop effective AV-VRU interaction and communication strategies.
Another important consideration in this context is how an automated vehicle might communicate its own behaviour and intentions with other road users, to convey certain messages or behaviours, and whether this is useful to a pedestrian’s crossing behaviour.
Therefore, a third study, currently in development, will employ rapid virtual prototyping to understand whether some vehicle-based signals from an AV are better understood, or are more acceptable, than others, for communicating information about the AV’s intentions and behaviours.
For example, using messages such as those shown below, a WHMD is used to ask participants to compare two such signals, and determine which message is best linked to which signal. These signals are then used to create new external Human Machine Interfaces (eHMI) to evaluate pedestrian’s comprehension of vehicle intention and investigate their crossing behaviour in response to these messages.
Future studies are currently being developed to investigate pedestrians’ responses to such eHMI signals in our new Highly Immersive Kinematic Experimental Research (HIKER) pedestrian laboratory (see picture below). Wearing a pair of glass-less frames, our researchers will be able to track participants’ position in the virtual road, assessing their response to different eHMIs, presented on a series of approaching vehicles. The most advanced of its kind in the world, this CAVE-based pedestrian simulator, funded by the UK EPSRC, will help researchers understand how road users respond to differently designed AVs, a task which is not easily achieved in the real world. Such controllable and affordable experiments are a good example of how virtual prototyping can be used to inform better design for future forms of transport, while also focussing heavily on users’ preferences and behaviour.