DFA based autonomous decision-making for UGV in unstructured terrain
Ning Li, Xijun Zhao, Jianfeng Gao, Xing Ke Cui · 2017 IEEE International Conference on Unmanned Systems (ICUS) · 2017
In this paper, a novel approach of autonomous decision-making system based on the deterministic finite automation (DFA) is proposed for unmanned ground vehicle (UGV) in unstructured terrain. According to vehicle states and environmental perception by sensors, driving states are decomposed into eight different modes, therefore a DFA based decision-making algorithm is designed through probabilistic analysis of perception results. Subsequently, the planning method and control strategy for each mode are developed in detail to optimize the autonomous navigation capability. Meanwhile self-recovery mechanism is adopted to improve the autonomous ability when the UGV is blocked by dense obstacles. Numerous experiments are conducted and the experiment results show that the proposed method is helpful to improve the flexibility, security and intelligence of autonomous navigation, even in the complex unstructured terrain and significantly reduce the times of manual intervention.