多传感器多目标数据互联中的拉格朗日松弛算法研究

Study on Lagrangian Relaxation Algorithm of Multisensor-Multitarget Data Association

拉格朗日松弛算法在现代优化启发式算法中占有重要地位，本文对多维分配问题的拉格朗日松弛算法进行研究，重点研究松弛次序对数据互联结果的影响。研究方法是针对不同测量误差情况下进行计算机仿真实验。结果表明，多传感器多目标数据互联的拉格朗日松弛算法中，松弛探测效果较好的传感器的观测数据，有助于提高目标关联的准确率，且算法的计算量相应减少。该结论使多维分配问题的拉格朗日松弛算法能满足工程上对数据进行实时处理，具有实际应用价值。

Lagrangian relaxation algorithm of S-D assignment problem is studied in this paper. It plays a dominant role in modern optimal heuristic algorithms. This paper centered on the influence of relaxing order. The study shows that not only the accuracy of data association is improved but also the calculation burden can be reduced when the perfect detection is relaxed in algorithm of multisensor-multitarget data association. It avoids the disadvantage of slowly processing existed in engineering. The principle analysis is described and the simulation results are given to verify the validity of the conclusion.