Research On Localization Prediction With Mobility Ocean Currents Situation

With the development marine resources exploitation, underwater sensor network is getting more and more attention. It has a bright future for marine monitoring, marine information collection,disaster prevention and auxiliary navigation. As a key part of the underwater sensor networks, localization plays a key role on the effectiveness of underwater sensor. The localization methods are different from traditional ones, because of the underwater channel with high latency,low bandwidth,severe loss of multipath propagation,etc.This paper first studied the different movement models of underwater sensor network. The motion model is divided into two parts: motion model of shallow sea and Lagrange motion model of deep sea. According to different characteristics,the different localization mechanisms were proposed.For the shallow water areas, the Particle filter Algorithm is taken into consideration. This paper proposed algorithms of CWPFA (Coastal Waters Particle filter Algorithm) and ICWPFA(Improved CWPFA). The basic principle of two algorithms is same,but the initial sampling is different. CWPFA used measure information of all perception nodes to track by anchor nodes,while ICWPFA algorithm needs to select a specific node for sending information. The results show that the localization accuracy of CWPFA is greater than that of ICWPFA, while the response time and communication cost can be less than ICWPFA.For the deep sea marine sports, considering that SLMP (Scalable Localization with Mobility Prediction based on movement Prediction positioning) didn't refer to other nodes movement prediction information,a method of CWP (Base on the Current Situation of Mobile Prediction)algorithm is proposed. The algorithm combines precursor node movement prediction information and other information to complete the prediction. In this paper,we did a large number of simulation experiments. And with different parameters, the node localization coverage,positioning error and average communication costs were analyzed separately. The results show that even the CMP algorithm takes positioning error as sacrifice,but greatly improves the effect of average localization coverage and communication costs, which is always more important and more difficult to improved. The result also explains the reason why CMP algorithm can better adapt to the large-scale mobile environment on the deep sea marine spots.