SENSOR SELECTION AND TASK ALLOCATION IN SENSOR NETWORK by Ma Yifan Thesis Supervisor: Dr. Hou Fen MSc in Electrical and Computer Engineering Wireless sensor network is the wireless network composed of spatially distributed sensor devices to collect and report information about the surrounding environment. In recent years, wireless sensor network has attracted much attention. Its applications cover different fields such as environment monitoring, biomedical health monitoring, border security, traffic planning, urban dynamics mining and military target tracking, which keep our daily lives convenient and safety. Nowadays, as the rapid development of communications and electronic technology, sensors are becoming smaller, cheaper and more intelligent, such that they can be integrated into mobile devices like mobile phones and wearable devices easily. Users equipped with sensors embedded mobile phones can be a part of WSN to contribute data, so asto enhance the coverage of sensor network and collect surrounding measurements more specifically and accurately. Such mobile sensors have features of mobility and flexibility. Due to the aforementioned characteristics of the mobile sensor, it is complimentary with the stationary sensor. For the stationary sensor, it is easy to control the sensing coverage due to the fixed locations. Therefore, the main concern is the limited battery since it has a self-contained battery and is difficult to frequently maintain and charge the fixed sensors. On the contrary, the power concern is relaxed for mobile sensors since the charging can be regularly conducted for mobile devices. However, the sensing coverage of mobile sensors varies with their locations, which makes it difficult to accurately predict the sensing coverage due to the randomness of user mobility. Thus a heterogeneous sensor network consisted by the two kinds of sensors will be considered in our research. In this thesis, a key question is how the service provider selects the sensors to conduct the sensing task considering the heterogeneity of sensors in terms of location, mobility pattern, energy constraint, and sensing cost. First, we formulate the dynamic sensor selection as a sequential decision problem and propose an optimal sensor selection algorithm OSSA using dynamic programming. Then we proposed a greedy selection algorithm GSSA to reduce the computational complexity and compare the expected social welfare with OSSA and random selection. Simulation results show the nice performance of the proposed OSSA algorithm in terms of social welfare efficiency and coverage ratio. In addition to sensor selection, sensing task allocation is another important issue in the wireless sensor network. In terms of this issue, we consider the network scenario where mobile users are both the data contributor and service consumers. We address how to fairly and efficiently allocate the sensing task among mobile users. The work focuses on the situation that the total sensing tasks are less than mobile users’ total service consumption. A bankruptcy game based task allocation among mobile phone users is proposed. The allocation of four different bankruptcy rules are discussed to solve the proposed problem and analyze the satisfaction of users and the fairness of each rule. In summary, the efficient sensor selection algorithm is proposed in the thesis considering users’ location, mobility pattern, energy constraint, and sensing cost. In addition, taking the service consumption of mobile phone users into consideration, we also design a novel task allocation scheme to achieve better performance in terms of fairness and efficiency.