I. Introduction
Multi-robot teams have been proposed for planetary surface exploration missions. These missions typically require team members to coordinate, plan motions, and move to specific locations of interest (targets) in communication-restricted environments. Often, these robots have partial or no information about the environment at the mission onset. Moreover, the robots may have constrained dynamics making it hard or impossible to compute the two-point boundary value problem (BVP) solution [1] for their motion planning. In this work, we propose a multi-robot feasible motion planning framework that considers partially known planetary surface environments and complex robot dynamics to coordinate robots to visit multiple unlabeled targets without solving the two-point BVP. The framework choice is motivated by limited power, computation and communication – constraints that are typical for a multi-robot planetary surface mission.