Optimizing Motion Comfort in Automated Driving: Motion Sickness Reduction and Augmented Trajectory Planning

As research in automated driving progresses, vehicles take over driving tasks and enable occupants to focus on non-driving-related activities. However, a certain level of comfort is required for this new technology to be accepted, as noticeable discomfort can hinder the ability to engage in these activities. Motion comfort should therefore be considered an important aspect when developing automated driving functions. This thesis focuses on improving motion comfort through trajectory planning, utilizing not only classical actuators – such as motor, brake and front steering – but also additional actuation systems to extend the range of feasible motion. The formulation as an optimal control problem (OCP) is the preferred planning method, as it allows for finding the best feasible trajectory that is optimal w.r.t. an objective cost function. A widespread and severe trigger of discomfort is motion sickness (MS), which refers to motion-induced symptoms such as dizziness, nausea, or even vomiting. In addition to discomfort factors that relate to physical disturbances – e.g. acceleration and jerk – this work specifically investigates MS and how it can be considered both effectively and efficiently in optimal planning. It is generally accepted that MS is linked to a sensory conflict between actual and perceived motion, and that head movements are strongly associated with MS. Since head motion is crucial for predicting MS, this work focuses on modeling the vehicle-to-head motion transmission and presents a subject study with ten participants to identify such a model. An essential finding is that the transient dynamics of motion transmission can be neglected when assuming comfortable trajectories, and that only the steady-state head tilt significantly influences MS. As a result, minimizing vehicle jerk strongly correlates with the reduction of MS-causing stimuli, which can be used to significantly simplify the planning problem. Another focus of this thesis is exploring the potential of additional actuation – such as active suspension systems or rear-axle steering. This is referred to as augmented motion, as it extends the range of feasible motion perceived by the passenger. A key innovation is the comfort drift function, which exploits coordinated front and rear-axle steering, simulating both oversteer and understeer behavior to effectively mitigate perceived jerk. On a test track, it was found that the comfort drift function applied to a series production Audi A8 with a rear steering range of 4.5° can achieve a 10.6% reduction in jerk, while increasing the steering range to 10° could achieve a 20.3% reduction for the same travel time. The proposed method for quantifying the benefits of motion comfort can be applied to any actuator configuration and range. Overall, this work provides a comprehensive overview of mitigating discomfort through trajectory planning in an optimization framework. It covers the modeling of comfort-relevant cost features, explains and discusses the optimization results, and demonstrates the real-world applicability of the theoretical findings through various measurements in the Audi A8 demonstration vehicle.