Motorcycle rider posture measurement for on-road experiments on rider intention detection

Abstract

Motorcycle riders represent a highly vulnerable group of road users with high risk of heavy injuries and fatalities per distance traveled. Hence there is an ongoing demand for the development of assistance systems to improve riding safety. Collecting information about a rider's intention – the desired maneuver to carry out or trajectory to travel through – is considered as an enabler for new systems that can warn or intervene before or assist in dangerous driving situations. The observation of the rider's posture is necessary for a holistic understanding of the human-machine interface as riders typically move their body during riding for various reasons. The authors develop and test an on-road capable measurement system of high accuracy and robustness for the detection of rider upper body posture in riding experiments as off-the-shelf systems are not existent. AprilTag optical markers applied to the back of the rider that are filmed by a camera from behind prove to be superior to other concepts tested. Two new methods named subarea and dynamic frame rate evaluation are introduced to reduce computational effort from raw video data to rider posture information. First measurement results from on-road riding are presented and reveal positional errors below 1 cm or 3 deg rider lean angle. Based on the data that is collected in an ongoing riding study, the meaning of posture information for the identification of rider behavior and intention will be further investigated.