Approaches to identification and analysis of control patterns in motorcycle emergency braking response

Abstract

A new research area explores the relationship between motorcycle rider muscle activation patterns and vehicle kinematics and kinetics to understand the interaction between human control inputs and vehicle kinematic outcome. A common factor in serious motorcycle accidents is the absence or inadequacy of braking by the rider to avoid collision. This is a complex perceptual-motor task, and it is unknown whether the key aspect in expert performance is related more to visual perceptual cueing, mechanical control, or an ideal combination of both. In the early stages of this project, the first goal is to equate neural control of movement, - evaluated from muscle activation and kinematic patterns - with objective outcome measures of front and rear braking force and ratio, stopping distance, and vehicle stability. Riders were equipped with wireless integrated EMG/IMU sensors placed bilaterally at back and forearm muscles. The test vehicle was a scooter equipped with sensors to measure vehicle kinematics, wheel speeds and brake force. Our protocol involves a controlled environment in which the rider approaches an intersection, while a car (driven by an experimenter) approaches from the opposite direction. The car driver may go straight or initiate a turn in front of the rider (never passing the midline) while the rider must brake in the shortest distance possible as soon as the car’s turning is perceived visually. We have confirmed that the experimental paradigm is safe and feasible, and that measures of human and vehicle performance during the braking event are readily identified and amenable to quantitative analysis and comparisons. Braking events can be divided into identifiable phases to analyze the sequence of the rider’s control response. Performance outcomes such as stopping distance are compared to muscle onsets and synergy patterns to characterize different categories of rider response and provide objective measures of skill and performance. Initial results show characteristic differences between less-skilled and expert riders related to functional versus non-functional variability in muscle activation patterns. For example experts show more systematic, repeatable modulation profiles of right forearm muscle activity (front brake control) to produce different effects in the phases of the braking event, while greater variability in left arm muscle patterns (rear brake) likely reflects voluntary adjustment in response to vehicle motion to regulate stability and traction control. Greater control modulation by experts is interpreted as a greater capacity to quickly develop an internal model of vehicle motion resulting from rider control outputs. Outcomes will be applicable to improving rider training methods, and to furthering understanding of rider-vehicle interactions, which is necessary for the development and design of automated rider assistive technology such as automated braking.