| The acquisition of indoor target information is an important research content of the location based services.Now,the Ultra-Wideband(UWB)radar technology attracts much more attention because it can truly acquire information of the targets without contact and work with low power consumption,high resolution,high penetration and strong anti-interference ability.However,there are still many challenges to put the UWB radar in practice:a single radar is difficult to obtain comprehensive information because of its limited line-of-sight;the indoor environment is closed and the objects such as walls or ornaments interfere with the signal;the targets block each other and seriously affect the received signal.So,this thesis will research and implement the indoor multi-target tracking based on UWB radar network.Firstly,the thesis analyzes the characteristics of the indoor environment,the principle of the UWB radar,the type of the radar signal and the propagation process of the signal.Then,this thesis discusses clutter suppression process and target detection process and proposes a two-step CLEAN algorithm.The proposed algorithm filters the detection results based on density and optimizes the detection effect of the targets far from the radar.In addition,a multi-target tracking algorithm combining with clustering is proposed.The algorithm detects the number of the targets through Density-Based Spatial Clustering of Applications with Noise(DBSCAN)clustering and KMEANS clustering and associates the measured values with targets using JPDA.Experimental results show that the proposed algorithm effectively responds to the changing of the number of the targets and maintain sufficient tracking accuracy.Finally,this thesis builds a multi-target tracking system based on the theoretical analysis.The system uses three UWB radars for networking and it works for data collection,data processing and results demonstration.The results of this thesis have certain reference to the multi-target tracking. |
PDF Full Text Request