Micromobility: an opportunity for future transport and a challenge for intelligent vehicles (Keynote)

Other conference contribution, 2023

Today's vehicles are increasingly connected, automated, and smart. The challenge for vehicle intelligence is to satisfy all road users’ mobility needs while keeping them safe and comfortable. Ensuring the safety of micromobility users (e.g., cyclists and e-scooterists) is particularly tricky, not only because of their intrinsic vulnerability but also because their behavior is hard to predict. Maneuvers involving high speed, close proximity, and multiple threats—such as overtaking when there is oncoming traffic—are particularly difficult to handle because a tiny human error may result in a fatality. In order for vehicles to provide support during overtaking maneuvers (or perform the maneuver themselves), they must be able to predict human behavior accurately and intervene promptly when necessary, before a human shortcoming can cause harm to any road user. We can combine simulator, test-track, and naturalistic data to devise models (i.e., algorithms) that can explain to machines how a reference driver may safely interact with micromobility users. These models are intended for advanced driving assistance systems (to support drivers maneuvering among micromobility users) and for automated vehicles (to efficiently interact with micromobility vehicles while maximizing the safety and comfort of all road users).