基于引导优化的拒止环境无人系统相对定位

Relative Positioning of Unmanned Systems in Denial Environments Based on Guided Optimization

在卫星拒止条件下,包含无人机、无人艇等多个节点的无人系统可以基于节点间距离信息,使用相对定位方法解算所有节点的绝对地理位置。由于无人系统节点的高机动特性,节点间的距离信息具有瞬时有效性,因此需要实现快速、精确的相对定位解算。旋转矩阵的计算是相对定位解算的核心步骤,使用传统的最小二乘法、牛顿法、遗传法等无法兼顾效率与精度因素,基于自由度松弛的引导优化算法通过放松旋转矩阵的自由度能够将寻优参数的初始值一步引导至真值邻域附近,再利用迭代法进行精细寻优,从而实现旋转矩阵的短时间、高精度计算。仿真实验结果表明,在无人系统测距误差达到5 m时,基于引导优化的相对定位方法解算时间为0.014 s,产生的定位误差为25.73 m,与牛顿法的解算精度相当,同时减少了55.56%的计算时间;与最小二乘法的计算时间相当,同时减少了39.47%的定位误差。

Under satellite denial conditions,unmanned systems composed of multiple nodes such as drones and unmanned surface vehicles can utilize relative positioning methods to determine the absolute geographical locations of all nodes based on inter-node distance information.Due to the high mobility of unmanned system nodes,the distance information between nodes has instantaneous validity,necessitating the implementation of rapid and precise relative positioning solutions.Calculation of rotation matrices is a core step in relative positioning.Traditional methods such as least squares,Newton s method,and genetic algorithms cannot balance both efficiency and accuracy.The guided optimization algorithm based on freedom degree relaxation can guide the initial value of parameter optimization to the neighborhood of the true value in only one step by relaxing the freedom degree of the rotation matrix.Subsequently,fine optimization is performed using iterative methods,enabling short-time and high-precision calculation of rotation matrices.Simulation experimental results show that the calculation time of the relative positioning method based on guided optimization is 0.014 s with a positioning error of 25.73 m when the ranging error of the unmanned system reaches 5 m.This is comparable to the accuracy of Newton s method but reduces the calculation time by 55.56%.The calculation time is equivalent to the least squares method,but the positioning error is reduced by 39.47%.