Velocity Control Strategies to Improve Automated Vehicle Driving Comfort

Automated vehicles are rapidly emerging as the new generation of transport tools. Relevant research mainly focuses on collision avoidance, lane keeping, and coordination control to ensure driving safety and improve efficiency, with little emphasis on passenger comfort. This paper proposes a set of comfort-based velocity control strategies on the basis of driving safety. Driving comfort was evaluated by multiple accelerometers fixed inside vehicles. Power spectral density analysis and one-third octave band filtering were applied to derive the weighted root mean square acceleration as an evaluation indicator for comfort. The annoy rate model is proposed to describe individual sensitivity to vibration. Field tests were conducted to investigate interactions between pavement roughness, speed, and comfort, and multiple linear relationships were identified. Three typical processes were considered: deceleration phase, uniform speed phase and abnormal occasions. The hyperbolic tangent function was utilized to provide moderate deceleration. Model parameters were set to guarantee maximal deceleration and jerk within acceptable limits. Nonlinear programming was used to balance between rapid deceleration and bumps caused by abnormal obstacles. The proposed strategy can potentially be applied for automated vehicles to improve the perceived quality of automated driving.