Determination of Hands-Off Detection Timespan Based on Asymmetric Nash Bargaining

Hands-off detection (HOD) is used in autonomous driving vehicle (ADV). However, how to choose an appropriate HOD timespan (HODT) either practically or legislatively remains unsolved. In this paper, an asymmetric Nash bargaining-based HODT determination is introduced. Utility functions of autonomous driving system (ADS) and driver are developed. HODT is determined based on the asymmetric Nash bargaining solution. Experimental result shows existing adopted HODTs can be obtained using the proposed method. Appropriate HODT can be obtained considering safety requirement and driver's need. Under high safety requirement, HODT is short regardless of driver's need. When safety requirement is not strict, HODT can be prolonged to satisfy driver's need. By constructing situation awareness (SA) evolution process considering HODT, ceiling HODT can be obtained, which can be used in legislation as the longest HODT that can be chosen. Experiment also shows superiority of dynamical HODT over the constant HODT. The proposed method offers a possible theoretical solution to HODT determination, showing potential in practice and legislation.