A bayesian approach for constrained multi-agent minimum time search in uncertain dynamic domains

Abstract

This paper proposes a Bayesian approach for minimizing the time of finding an object of uncertain location and dynamics using several moving sensing agents with constrained dynamics. The approach exploits twice the Bayesian theory: on one hand, it uses a Bayesian formulation of the objective functions that compare the constrained paths of the agents and on the other hand, a Bayesian optimization algorithm to solve the problem. By combining both elements, our approach handles successfully this complex problem, as illustrated by the results over different scenarios presented and statistically analyzed in the paper. Finally, the paper also discusses other formulations of the problem and compares the properties of our approach with others closely related.

Multi-UAV target search using decentralized gradient-based negotiation with expected observation

“…In this paper, we presented a novel approach to reducing the myopicity of finite action horizon decision making controllers for solving the minimum-time search [30,20,21] and the lost target search problem [33,9,5] using a team of UAVs. We have modeled the future expected reward as a heuristic based on a range sensor model that estimates the expected future observation.…”

“…This reduces the vulnerability to central server failure, which further enhances the overall system robustness. In particular, there is an active research community proposing mobile sensor decision making solutions for target search [37,5,12,4,16,27,30,20,13,21,7] and tracking [15,8,17] problems. These solutions intrinsically exploit the action-perception loop [28,10] to design autonomous cognitive agents based on recent computational and decision models of the human brain [1,24].…”

Minimum Time Search in Real-World Scenarios Using Multiple UAVs with Onboard Orientable Cameras

“…One of the main goals of MTS planners is to reduce the target detection time, which can be achieved by optimizing the expected time of target detection [5][6][7][8][9][10]. Other PTSP approaches optimize alternative criteria, such as maximizing the probability of target detection [11][12][13][14] or minimizing its counterpart probability of nondetection [15,16], maximizing the information gain [17], minimizing the system entropy [18], minimizing its uncertainty (areas with intermediate belief of target presence) [19], or optimizing normalized or discounted versions of the previous criteria [4,[20][21][22].…”

“…The NP-hard complexity of PTSP [23] is tackled with suboptimal algorithms and heuristics, such as gradientbased approaches [13,[15][16][17]19], greedy methods [8,12,20], cross-entropy optimization [4,7], Bayesian optimization algorithms [5], ant colony optimization [6,9], or genetic algorithms [10]. Besides, streamlined formulations of the problem are typically accepted in order to further simplify the problem complexity.…”

“…Besides, streamlined formulations of the problem are typically accepted in order to further simplify the problem complexity. They range from considering static targets [8,10,13,15,16,19,20] instead of dynamic ones [4-7, 9, 11, 12, 14, 17, 21, 22] to modeling the sensors ideally [4][5][6]8] instead of realistically (e.g., as radars [9][10][11][12] or downward-looking cameras [13,14,17,19]) or to assuming that the UAVs fly following straight-lines according to the eight cardinal directions [4][5][6][7][8] or optimized waypoints [17,21,22] instead of considering the physical maneuverability constraints induced by the dynamic models of the UAVs [9-16, 19, 20]. Additionally, in some cases (e.g., in [10,13,16,17,19,20]) the approach uses a receding horizon controller to divide the UAVs trajectory into sequentially optimized sections, narrowing the optimization search space at the expense of constructing suboptimal myopic solutions.…”