Manshi Limbu

This paper studies a team coordination problem in a graph environment. Specifically, we incorporate “support” action which an agent can take to reduce the cost for its teammate to traverse some high cost edges. Due to this added feature, the graph traversal is no longer a standard multi-agent path planning problem. To solve this new problem, we propose a novel formulation that poses it as a planning problem in a joint state space: the joint state graph (JSG). Since the edges of JSG implicitly incorporate the support actions taken by the agents, we are able to now optimize the joint actions by solving a standard single-agent path planning problem in JSG. One main drawback of this approach is the curse of dimensionality in both the number of agents and the size of the graph. To improve scalability in graph size, we further propose a hierarchical decomposition method to perform path planning in two levels. We provide both theoretical and empirical complexity analyses to demonstrate the efficiency of our two algorithms.

We are interested in designing coordinated group motion, where the safety or cost for one agent to move from one location to another may depend on the support provided by its teammate. For example, consider a scenario where a team of human and robot must traverse an environment with some “risk” edges as shown in Fig. 1. Those risks might represent actions such as going up a ladder, crossing a "shaky" bridge, or walking through a dark tunnel. In these situations, a human (or robot) teammate can support the other by holding the ladder, stabilizing the bridge, or lighting up the tunnel. We capture the feasibility of these "supporting" actions in the green dashed arrows in Fig. 1, extending from the nodes from which the support can be provided. The core questions we seek to answer are: (i) when such support/coordination is beneficial, and (ii) how to best coordinate the actions as a team to minimize the overall cost. We formulate a problem that incorporates support actions to a minimum-cost graph traversal problem. We then propose a solution approach based on the notion of joint state graph (JSG) formulation, converting the problem into single-agent path planning. To address the curse of dimensionality, a hierarchical decomposition method based on Critical Joint State Graph (CJSG) is introduced for two-level planning. Complexity and statistical analyses demonstrate the efficacy of our algorithm.