On The Robust Design Of Sensor Data Fusion Problems

Sensor data fusion have wide applications in military and civilian area. A sensor is to estimate the origin data in the sense of unbiased minimum-variance by receiving the uncertain data. In the case of multi-sensors, the data should be fused to improve the estimation accuracy. However, the sensor may be interrupted by external uncertainties such as noise, weather, etc. in practice, and then the accuracy of estimation is reduced. In order to improve the accuracy of estimation, the robust design of sensor data fusion problem should be introduced.In this thesis, we consider robust sensor data fusion problem in discrete time and continuous-time. This thesis is organized as follows.In chapter 1, we introduce the sensor fusion problem and review the literature of this topic. Then, we introduce the concepts of optimal control and Kalman filtering.In chapter 2, we consider the robust sensor fusion problem in discrete time. We esti-mates the state of an uncertain process based on measurements obtained by a given set of noisy sensors, where the measurements of sensors are subjected to external interference uncertainties. We formulate this problem into a minimax optimal control problem, which is equivalent to a semi-infinite programming problem with a dynamic system. The dis-cretization method is used to solve this problem, where the computation is very large scale in general. We propose an approximation method such that the number of constraints is decreased and the computational complexity is reduced. For illustration, two numerical examples are solved.In chapter 3, we consider the robust sensor fusion problem in continuous time. Sim-ilar to the discrete case, there are a group of sensor systems which are continuous with uncertain perturbation parameters. We estimates the state of an uncertain process based on measurements obtained by these sensors. We formulate the robust sensor fusion prob-lem by transforming it into an equivalent semi-infinite programming problem driven by a dynamic system. We propose an approximation method such that the number of con-straints is decreased and the computational complexity is reduced. Then this problem can be solved by the optimal control software MISER3. For illustration, two numeri-cal examples are solved to demonstrate the efficiency and effectiveness of the proposed method.In chapter 4, we summarize the research in this thesis and make a prospect for future research.