论文部分内容阅读
针对多目标跟踪中的未知杂波强度,提出了基于熵分布的杂波强度在线估计算法.利用有限混合模型对未知杂波强度建模,将仅依赖于混合权重的熵分布作为混合参数的先验;利用拉格朗日乘子法推导了混合权重在极大后验意义下的在线估计公式;以随机近似过程为在线估计策略,推导了基于缺失数据的分量均值和方差的在线估计公式.仿真结果表明,基于熵分布的杂波强度在线估计算法改进了概率假设密度滤波器在多目标跟踪中的性能.
Aiming at the unknown clutter strength in multi-target tracking, an on-line estimation algorithm of clutter strength based on entropy distribution is proposed. Using the finite mixture model to model the unknown clutter intensity, the entropy distribution which only depends on the mixture weight is taken as the first The Lagrange multiplier method is used to derive the online estimation formula of mixed weights under the great a posteriori. On the basis of the stochastic approximation method, the online estimation formula of the component mean and variance based on the missing data is deduced. The simulation results show that the on-line estimation algorithm based on entropy distribution improves the performance of probability hypothesis density filter in multi-target tracking.