An efficient multi-objective optimization approach for sensor management via multi-Bernoulli filtering

Intelligent sensor management is generally required for efficient and accurate data processing when the multi-sensor system is used for multi-target tracking (MTT). However, this is theoretically and computationally challenging. To deal with this problem, we propose a novel sensor management approach based on efficient multi-objective optimization for MTT under the framework of partially observed Markov decision process. The multi-Bernoulli filter is used in conjunction with two objective functions. To simplify the multi-objective optimization problem, we use the Euclidean distance (ED) between the feasible and utopian solution vectors as a measure of the objectives and then sequentially select sensors from the candidates. For the selected sensors, we rank them according to the obtained ED measure and implement the iterated-corrector fusion scheme after the ranking. Numerical studies demonstrate the effectiveness and efficiency of our approach in multi-sensor MTT scenarios.