基于箱式粒子滤波的群目标跟踪算法

在现有群目标跟踪方法中,粒子滤波(Particle filter, PF)算法常被用来解决点量测的非线性滤波问题.而当量测数据受到测量偏差或未知分布边界误差的影响时,传感器获得的点量测需要转换成区间量测,此时原有PF算法不能直接适用.因此,本文提出基于广义似然(Generalized likelihood, GL)函数加权的PF算法.该算法在原有PF算法的基础上,利用广义似然函数的积分解来计算区间量测下的粒子权重.为了降低算法的运算量问题,又提出基于箱式粒子滤波(Box particle filter, Box-PF)的群跟踪算法.首先,在目标状态空间内抽样矩形区域的箱式粒子.然后采用区间分析和约束传播方法,利用区间量测压缩后的粒子与预测粒子的容积比来计算粒子权重.最后,在群目标状态估计结果和群演化网络模型的基础上估计群结构.仿真实验结果表明,与GL-PF算法相比, Box-PF算法具有更高的运算效率,并能降低估计结果中的峰值误差.

Group Targets Tracking Algorithm Based on Box Particle Filter

Particle filter(PF) algorithm is often used to solve the nonlinear filtering problem for point measurements in the existing group targets tracking algorithms. However, the traditional PF algorithm cannot be directly applied to the case where the point measurements should be converted to interval measurements when the measurements are affected by biases or bounds errors of unknown distributions. Therefore, this work presents an improved PF algorithm based on the generalized likelihood(GL) weighting method. The GL-PF algorithm uses the definite integral solution of generalized likelihood function to calculate the weighting of particles under interval measurements. For the sake of reducing computational burden, this work presents another group tracking algorithm based on box particle filter(Box-PF). Firstly, the rectangular box particles are sampled in the target state space. Then, the ratio between the contracted and the predicted box particle volumes is used to calculate the weighting of particles based on the interval analysis and constraints propagation method. Lastly, the group structure is estimated based on the estimation results of group target state and the evolving network model. Computer simulations show that compared with the GL-PF algorithm, the Box-PF algorithm can achieve a greater computational efficiency and reduce the peak error of the estimation results.