A Hybrid Estimation Algorithm for Tracking an Adversarial Team

This paper is motivated by the problem of two adversarial networked teams that collect information about each other and make decisions based on this information. This problem has applications in network security, economic decision making, and robotic soccer. From the perspective of one team, the quality of decision making can be improved with better methods of aggregating measurements of the other team and estimating unknown information about that team. This paper presents an approach for estimating the underlying strategy of an opposing team based on limited observations. As the opposing team has continuous and discrete states, it can be modeled as a hybrid system with continuous and discrete measurable outputs. A sequential estimation algorithm is developed which simultaneously estimates both continuous and discrete states. An example is provided which illustrates the application of this algorithm to estimating the formation of an opposing team from incomplete information.