Airborne multi-sensor target recognition method based on weighted fuzzy reasoning network and improved DS evidence theory

Abstract

Difficult obtainment of basic probability assignment (BPA) and inaccurate recognition due to conflicting data are the problems in applying Dempster-Shafer (DS) evidence theory to airborne multi-sensor target recognition. In order to effectively deal with them, an airborne multi-sensor target recognition method based on weighted fuzzy reasoning network (WFRN) and improved Dempster-Shafer (IDS) evidence theory is proposed in this paper. First, the feature vector consisting of 5 feature components is constructed. And then a 4-layer WFRN consisting of 3 kinds of basic units is established to obtain BPA of the feature vector. Finally, conflict data is processed through IDS evidence theory to obtain the final fusion recognition result. The simulation results indicate that the proposed airborne multi-sensor target recognition method is able to obtain BPA reasonably and deal with conflict information.