带反馈多传感器分层融合算法性能研究

Chang K.C.于近期研究了两传感器非实时分层融合算法的稳定性能,其研究结果表明全反馈条件下的算法融合性能要劣于部分反馈算法,特别是当过程噪声较大时提高传感器与融合中心的通信频率并不能改善全反馈算法的性能。与上述研究不同,重点考察了融合系统中传感器数目对非实时有反馈分层融合算法性能的影响。研究结果表明:当过程噪声较大时,增加传感器的数目不仅不能改善全反馈算法的性能反而会使性能降低;所得结论进一步推广和发展了Chang K.C.的研究成果,并从另一侧面验证了融合信息的质量比数量更为重要的论断。

Study of Performance of Multisensor Hierarchical Fusion Algorithm with Feedback

Of late years,steady performance of two sensor hierarchical fusion algorithm with nonfullrate communication was evaluated by Chang K.C.,the result indivates that performance of algorithm with complete feedback is not good as that of algorithm with partial feedback and moreover,in the case of complete feedback,increment of data rate does not improve the performance.Different from their research,we pay more attention to the effect of number of sensors on the steady performance of the algorithm with different feedback mechanism.It shows that for algorithm with complete feedback,in the case of process noise being relatively large,increase on the number of sensors does not improve but worsens the performance.This conclusion is extension and development of work of Chang K.C.and,from another aspect,verifies that quality of information is much more important than quantity.