摘要
针对多目标跟踪中的传感器控制问题,提出一种基于有限集统计(FISST)理论的传感器控制策略.首先,基于随机有限集(RFS),建模给出基于信息论的传感器控制的一般方法;然后,利用多目标概率密度的近似统计特性推导柯西施瓦兹(CS)距离的表达形式,并最终利用粒子概率假设密度滤波器对评价函数进行求解,在信息增益最大化的准则下得到传感器的最优控制方案.传感器的轨迹控制实验仿真验证了所提出算法的有效性.
In consideration of the sensor control problem for multi-target tracking, a sensor control strategy is proposed based on the finite set statistics（FISST） theory. Firstly, the general method of sensor control based on the information theory is given based on the random finite set（RFS） modeling. Then, the expression of Cauchy-Schwarz（CS） distance is deduced by the approximate statistical properties of multi-target probability density. The evaluation function is solved through the particle probability hypothesis density filter. Thus, the optimal control scheme of the sensor is obtained under the criterion of maximizing the information gain. Simulation results of sensor trajectory control show the effectiveness of the proposed algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2018年第2期337-344,共8页
Control and Decision
基金
国家自然科学基金项目(61370037
61763029
61461026)
甘肃省自然科学基金项目(1506RJZA090)
关键词
传感器控制
随机有限集
多目标概率密度
柯西施瓦兹
sensor control
random finite set
multi-target probability density
Cauchy-Schwarz
