Target Tracking Oriented Self-Organizing Sensor Networks

Sensor networks are the enablers of a technology which can best be described as omnipresence. Small, inexpensive, low power sensors distributed throughout an environment can provide ubiquitous situational awareness. The technology lends itself well to surveillance and monitoring tasks—including target tracking—and it is in this application where our interests lie. This thesis analyses some key problems involved in applying sensor network in target tracking, which are Deployment, Coverage and Node Localization. And the thesis also brings forward a set of innovative solutions to settle down these associated problems. The contributions of this thesis are as follows:1) Target involved sensor deployment. Improving on the existing sensor detection model and virtual force method, we present a novel sensor deployment optimization strategy based on target involved virtual force. Additionally, based on the information that all sensors gathered, we propose an improved sensor-ranking algorithm as well as a sensor protection strategy with targets' importance sequence in consideration.2) Deployment analysis and optimizing based on coverage. We analyze Area Coverage Problem, propose a novel triangle-arranged circle division method to obtain the formula of the relationship of probability of coverage and connectivity, sensing radius, transmission radius, failure probability and the numbers of the nodes using probability method. And we also considered Barrier Coverage Problem. Based on the analysis of "Maximum Support Path", we get some heuristic rules of incremental sensor deployment, using which we could patch up some weakness of a type of established deployment.3) A novel sensor localization method based on optimized path of a mobile beacon. We advanced two methods to get appropriate position of mobile beacon in a rectangular ROI and an arbitrary ROI. And, we apply TSP algorithm to the position sequence in order to get the optimal path touring it. When mobile beacon moves according to the optimal path, as well as sending RF signal at every position, the sensors in ROI could work out their position with Trilateration.At last, combing all the algorithms above, we carry through the target tracking simulation experiments. Furmore, we design and implement a sensor...