基于轮询相对测距的多无人机协同优化方法

Multi-UAV collaborative optimization method based on polling relative ranging

针对多无人机协同定位时,轮询通讯架构下传统滤波方法无法抑制惯性导航系统误差累积发散的问题,提出了一种基于相对距离信息辅助的多无人机协同优化方法。以惯性导航系统为核心构建惯性预积分因子节点和相对距离信息协同因子节点组成因子图优化模型,引入时间滑动窗口,构建量测残差代价函数,利用Gauss-Newton最小二乘迭代对无人机的最优状态进行求解。仿真实验表明,在轮询通讯架构下,所提出的方法能够实现多无人机编队的高精度协同定位,达到广播式通讯架构定位精度水平,相较传统扩展卡尔曼滤波(EKF)方法,定位误差降低了73.8%。

Aiming at the problem that traditional filtering methods cannot suppress the accumulation and divergence of inertial navigation system errors under the polling communication architecture,a multi-UAV collaborative optimization method based on relative distance information assistance is proposed.With the inertial navigation system as the core,a factor graph optimization model composed of inertial pre-integration factor nodes and relative distance information collaborative factor nodes is constructed.A time sliding window is introduced.A measurement residual cost function is constructed,and Gauss-Newton least squares iteration is used to solve the optimal state of the UAV.Simulation experiments show that under the polling communication architecture,the proposed method can achieve high-precision collaborative positioning of multiple UAV formations,reaching the positioning accuracy level of the broadcast communication architecture.Compared with the traditional extended Kalman filter(EKF)method,the positioning error is reduced by 73.8%.