Multi-agent motion planning using Bayes risk

We introduce a novel approach to controlling the motion of a team of agents so that they jointly minimize a cost function utilizing Bayes risk. Bayes risk is a useful measure of performance for applications where agents must perform a classification task, but is often difficult to compute analytically for many applications involving agent state variables. We use a particle-based approach that allows us to approximate Bayes risk and express the optimization problem as a mixed-integer linear program. By minimizing Bayes risk, agents are able to account explicitly for the costs associated with correct and incorrect classification. We illustrate our approach with a target interception problem in which a team of mobile agents must intercept mobile targets that are likely to enter a specified area in the near future. We show that the cooperative agent motion that minimizes a cost function utilizing Bayes risk is an efficient way to achieve selective interception.