Fusion of Heterogeneous Friction Estimates for Traction Adaptive Motion Planning and Control
This work addresses the challenge of improving accident avoidance in critical situations for automated vehicles, though it is incremental as it builds on existing friction estimation techniques.
The paper tackled the problem of enabling traction adaptive motion planning and control for automated vehicles by addressing the need for accurate, real-time friction estimates of the road ahead, showing that fusing high-accuracy local friction estimates with rough forward-looking camera classifications yields near-optimal behavior.
Traction adaptive motion planning and control has potential to improve an an automated vehicle's ability to avoid accident in a critical situation. However, such functionality require an accurate friction estimate for the road ahead of the vehicle that is updated in real time. Current state of the art friction estimation techniques include high accuracy local friction estimation in the presence of tire slip, as well as rough classification of the road surface ahead of the vehicle, based on forward looking camera. In this paper we show that neither of these techniques in isolation yield satisfactory behavior when deployed with traction adaptive motion planning and control functionality. However, fusion of the two provides sufficient accuracy, availability and foresight to yield near optimal behavior. To this end, we propose a fusion method based on heteroscedastic gaussian process regression, and present initial simulation based results.