Weight-adaptive parameter estimation assisted event-triggered model predictive guidance for reentry

This paper proposes an event-triggered model predictive guidance (ET-NM PG) method assisted by weight-adaptive parameter estimation (WAPE) and applies it to reentry guidance. Guidance methods based on online trajectory optimization (TO) often face a trade-off between guidance accuracy and the efficiency of guidance command computation when dealing with complex problems. By using state deviation exceeding a threshold as an event-triggering condition, it is possible to effectively reduce computational resource consumption while ensuring a certain level of guidance accuracy. However, the actual model parameter values often deviate from the reference values, leading to an excessively high event trigger frequency and potentially rendering the trajectory optimization unfeasible. To address this issue, we propose updating the model parameters online using WAPE within the general ET-NMPG framework, thereby enhancing guidance accuracy and reducing guidance frequency. Importantly, for reentry processes, this approach can further ensure that the flight state remains within acceptable path constraints. Additionally, we designed a change point detection step to avoid data contamination in the case of in-flight faults (parameter mutations). The numerical simulation results confirm the effectiveness of the proposed method.