一种抑制DPA评价函数扩散的方法

Method for Restraining Merit Function Scattering Based on Dynamic Programming Algorithm

传统DPA算法在跟踪目标的过程中存在评价函数的扩散现象,即目标周围的评价函数会被"抬高",形成以目标所在位置为顶点的"目标锥"。若目标相距较近,各目标锥会相互融合,导致DPA算法难以有效地将全部目标检测出来。且经研究发现,目标的信噪比越高、或检测时间越长,扩散的程度就越大,故抑制各目标(特别是较高信噪比目标)的扩散很有必要。为此,提出了一种对评价函数扩散的抑制方法,目标信噪比越高,该方法对扩散的抑制效果越显著。仿真结果表明,采用新方法后目标周围评价函数的扩散程度相比传统DPA算法有了明显减弱,提高了DPA算法检测密集目标的能力。

The merit function （MF） of traditional dynamic programming algorithm （DPA） scatters during target tracking, and MFs around target arises forming a ＂MF group＂ Once in dense multi - target environment, the MF groups are merged, which makes the DPA hardly detect all targets successfully. Researches show that, the higher the signal-to-noise ratio （SNR） of a target or the longer a detecting period is, the more a scattering will be. Thus, restraining the MF scattering of targets, especially of higher SNR targets, is necessary. A novel method is presented for restraining MF scattering （MRMFS） , especially the scattering of higher SNR targets. Simulation results show that, the scattering of MRMFS is reduced, indicating that the performance of multi-target detection is improved with this method.