Adaptive Algorithm For Target Tracking In Wireless Sensor Networks

Wireless sensor networks (WSNs) are formed from hundreds of low-cost sensor nodes that work with self-organizing technology. It is widely used in military, environment monitoring, health nursing, home automation, disaster rescuing, and so on. As one of the most important applications of WSNs, target tracking gained much attention from the very beginning. Due to the severe resource constraints of sensor nodes, the research of target tracking becomes a challenging issue.Considering the disadvantages of present target tracking algorithms, such as the high computation complexity and the unreasonable using of energy, we propose an auction-based adaptive sensor activation algorithm (AASA) for target tracking in this thesis. AASA mainly solves the following three problems. First, AASA solves the uneven energy depletion phenomenon of the network. An efficient clustering algorithm is proposed in this thesis, which consists of a prediction method and an auction mechanism. Using the prediction method, we can determine where to form a cluster. Through the auction mechanism, appropriate sensor nodes are chosen to participate in tracking from all the nodes in a predicted region by competitions. The auction mechanism considers fully the remaining energy of a sensor node and the distance between the sensor node and the target. Second, AASA solves the interference problem of localization caused by detection errors. Original trilateration algorithm has the assumption that the network is ideal. That is to say, there aren’t any detection noises in the network, which is not consistent with the practical situation. In order to gain high localization accuracy in a network with detection noises, an improved trilateration algorithm is proposed in this thesis. Third, AASA solves the problem of making a trade-off between energy efficiency and tracking quality. In order to achieve energy-efficient tracking while keeping high tracking quality, we propose an adaptive sensor activation algorithm, which adaptively adjusts the radius of predicted region and the number of cluster members according to current tracking quality. Meanwhile, tracking interval is also dynamically modified based on the instantaneous velocity of the target in the adaptive sensor activation algorithm.We analyze the proposed algorithm AASA deeply through simulations. First, we study the influences of the important parameters in AASA on the tracking performance, and then we compare our algorithm with another existing tracking algorithm. Simulation results show that AASA solves the uneven energy depletion phenomenon, obtains significant energy savings and prolongs the network lifetime, while keeping acceptable tracking quality.