基于可观测性的序列图像自主导航观测时序规划方法

Observation Timing Planning Method for Sequential Images of Autonomous Navigation Based on Observability

针对行星着陆自主导航中图像处理计算负担重的问题,分析了序列图像自主导航的可观测性,并提出了一种观测时序规划方法。通过可观测性分析得到了在未知环境中使状态可观测的最少观测次数,这是切换观测陆标的边界条件。在此基础上,通过优化所构建的深度估计误差模型获得最佳观测间隔时间,从而自适应地规划观测时序,减少图像处理次数。仿真结果验证了可观测性分析的正确性,以及提出的观测时序规划方法的有效性,相比每个采样时刻均观测陆标,在不明显影响导航精度的条件下减少陆标观测次数45.9%,有效降低序列图像在线处理的计算负担,大幅提升未知环境中行星着陆器基于序列图像的自主导航能力。

To reduce the computational burden of image processing during the planetary landing autonomous navigation,the observability of sequential image-based autonomous navigation is analyzed,and an observation timing planning method is proposed.The minimum number of the observations to make the state observable in unknown environments is obtained from the observability analysis,which is the boundary condition for switching landmarks.On this basis,the optimal observation time interval is obtained by optimizing the constructed depth estimation error model,so that the observation timing can be planned adaptively and the number of image processing can be reduced.The simulation results verify the correctness of the observability analysis results and the effectiveness of the proposed observation timing planning method,which effectively reduces the number of landmark observations by 45.9%without significantly affecting the navigation accuracy compared with the observing landmarks at each sampling moment.It is shown that the proposed method can effectively reduce the computational burden of the onboard image processing algorithm and greatly improve the autonomous navigation capability of the planetary lander in unknown environments based on sequential images.