Research on Heterogeneous Redundancy Emergency Decision Strategy based on Hierarchical State Machine

In order to meet the functional safety requirements of unmanned driving technology at critical moments under decision failure, a heterogeneous redundant emergency decision-making strategy based on hierarchical state machine is proposed. According to the centralized architecture of the whole vehicle domain, the overall redundancy hierarchical state machine model is constructed from the data end, the decision-making end to the execution end, the two-heartbeat communication detection mechanism based on the time wheel is used as the boundary condition for the main redundant decision state transfer, and the AEBS control model of redundant decision is established using RSS theory, and finally the overall hierarchical decision state machine architecture is processed recursively. Taking a pure electric bus as the test object, from the main decision start operation to the redundant decision intervention after failure, the hierarchical decomposition verification is carried out according to different trigger conditions. The results show that the overall time efficiency of redundant decision-making from trigger initiation to policy control execution is only 0.79% compared with the main decision, and the double-heartbeat timeout threshold can effectively trigger the precise intervention of redundant decision-making at 160ms, so that the decision security is further improved and computing resources are rationally utilized. At the same time, the hierarchical state machine theory can realize the circulation of various working modes of the vehicle after the failure of the main decision, so that the sub-states can be reasonably transferred in the redundancy decision, so as to meet the design feasibility and effectiveness requirements of the redundancy function.