一种动态分组的多节点协同定位编队构型优化方法

A dynamic grouping formation configuration optimization method for multi-node cooperative localization

针对并行式多节点协同定位时协同节点选用的问题,提出了一种基于改进遗传算法的多节点动态分组协同定位编队构型优化方法。以节点间测距信息为量测信息,建立基于容积卡尔曼滤波的多节点协同导航模型;针对实际系统中编队构型约束和测距信息限制,改进了标准遗传算法的编码、交叉和变异操作,以Cramer-Rao边界不等式求解的编队误差界作为适应度函数,完成了多节点系统最优分组问题的求解。分别针对10、20、50节点协同系统进行编队优化仿真试验,验证了改进遗传算法的可行性和有效性。仿真试验结果表明,编队构型优化方法可以在固定编队协同定位基础上进一步提升节点定位精度,动态编队的定位精度较固定编队平均提高了36%。

Aiming at the problem of cooperative nodes selection in parallel multi-node cooperative localization, a multi-node dynamic grouping cooperative localization formation configuration optimization method based on improved genetic algorithm is proposed. A multi-node collaborative navigation model based on Cubature Kalman filtering is established by taking the inter-node ranging information as the measurement information. Considering the constraints of formation configuration and ranging information in practical system, the coding, crossover and mutation operations of standard genetic algorithms are improved.And the optimal grouping problem of the multi-node system is solved by taking the formation error bound solved by the Cramer-Rao boundary inequality as the fitness function. Formation optimization simulation experiments are carried out for cooperative systems with 10, 20, and 50 nodes, which illuminate the feasibility and effectiveness of the improved genetic algorithm. The simulation experiment results show that the formation configuration optimization method can further improve the node positioning accuracy on the basis of the fixed formation cooperative localization, and the positioning accuracy of the dynamic formation is improved by about 36% on average compared with the fixed formation.