一种招投标机制的多星自主协同任务规划方法

针对多星自主协同遥感背景下非预期任务的快速响应问题,考虑到星上计算资源有限、计算能力较弱等特点,为寻找一种满足星上自主任务规划能力需求的优化算法,提升遥感卫星星群在非预期情况下的快速响应能力,通过多星自主协同规划问题建模、算法设计和仿真分析等模型及算法研究,提出了一种基于招投标机制的自主任务规划方法.该方法首先针对多星自主协同任务规划问题,构建了星上自主任务规划的数学模型,进而在问题求解过程中将一次完整的任务规划合理分解为招标、投标和评标3个过程,并详细设计了求解流程及相应的约束检验规则,由此得到基于招投标机制的多星自主协同任务规划求解算法.该方法与常用智能优化方法相比,能够显著降低计算量,更加适应星上紧张的计算资源约束.通过仿真算例结果表明,针对典型的非预期任务,算法平均仿真运行时间约为1 s,能够在40 s内完成对非预期任务的响应,并且充分保证了原规划任务的完成率,基于此验证了该方法的有效性与正确性.

A multi-satellite autonomous coordination task planning method based on bidding mechanism

Based on the bidding mechanism, an autonomous mission planning method was proposed to solve the rapid response problem of unexpected tasks in the background of multi-satellite cooperative remote sensing, which considers the limited computing resources and the weak computing power on the satellite. In order to find an optimization algorithm that satisfies the requirements of on-board autonomous mission planning and enhances the rapid response capability of the constellation of the remote sensing satellites in unexpected situations, the multi-satellite collaborative planning modeling, algorithm design, and simulation analysis were studied. First, the mathematical model of autonomous mission planning on the satellite for multi-satellite autonomous collaborative task planning was constructed. Then, in the problem solving process, a complete task planning was reasonably decomposed into three processes including invitation for bid, bidding, and evaluation. With the solving process and the corresponding constraint inspection rules designed, the multi-satellite autonomous collaborative mission planning algorithm based on bidding mechanism was obtained. Compared with commonly used intelligent optimization methods, this method can significantly reduce the computational complexity and adapt to the tight computational resource constraints on the satellite. The simulation example shows that for a typical unexpected task, the average simulation running time of the algorithm was about 1 s, the response to unexpected tasks could be completed within 40 s, and the completion rate of the original planning task was fully guaranteed, which verifies the validity and the correctness of the method.