A Research On Algorithm For Target Localization And Tracking In Internet Of Things Based On Gegographic Information

Internet of things(IoT) is in a prevalent concept in recent years, through various sensing equipment and networking, it can connect people and things in the real world. The universal existence and ability to sense target making it very suitable for target tracking. At present, based on the radar and wireless sensor network(WSN) ,many effective target tracking algorithm was proposed, but it needs the target are sensed by equipment at any time, so when the equitment is few and scattered, or the equipment is out of work, the accuracy of locating and tracking the discount.In the process of target tracking, a factor is especially useful - geographic information. To some degree, it can atone errors caused by the low equipment precision and the lack of equipment. Therefore, this paper is to discuss the target tracking methods based on geographic information, especially here, the geographic information can restrain the target's moving pattern, such as road network, mine laneway, etc.Firstly this paper studied raod network modeling, puts forward a modeling nethod, which can express the actual network restrictions and attribute, support localization and tracking algorithm to find the road sections effectivly. Then based on network model,we propose a road segment matching algorithm, which can correct positioning results, through comprehensive consideration of the target's motion model, speed,direction and other data. By using the algorithm we can project the position to correct road segment and display on digital map, also it can reduce the locating and tracking error.Secondly, according to the characteristics of IoT equipment, we put forward a locazation algorithm based on regional intersection, which combined the advantage of trilateral survey localization algorithm and weighted centroid localization algorithm. Through the calculation of intersection arc, it can locate the target in a more smaller area. Simulation experiment verify that compared with centroid localization algorithm it can distinctly improve the location accuracy.The target's the trajectory presents repeatability, so we can consider the moving pattern in the process of locating and tracking. According to the special road network constraint, we put forward a novel frequent movement pattern mining algorithm, which also support incremental database. Then we can use the moving pattern to predict target's current state .It can solve the uncertain problem in target tracking.Finally we realized target tracking system which based on geographic information, describe the system structure and function. Through this system, we can config geographical information, manage deployed IoT equipment, view tracking trajectory real-time. In the system, we realize the road matching algorithm, frequent moving pattern mining algorithm and localization and tracking algorithm, system runs efficient, which shows our method is feasible and work efficiency.