This work presents a new solution for coordinated search with a team of heterogeneous robots executing a time-critical mission. It is challenging to specify and represent search locations (targets) in known but dynamic environments as well as to find robotic paths to visit the locations. We propose a technique to construct an information map that includes locations of uncertain targets, and generate optimal paths. We especially focus on combining a satellite map that has global coordinates with local images gathered from an aerial robot. Specific targets are represented on a homogeneous coordinate system, so that different types of robots, capable to gather necessary information, may cooperatively conduct a mission. Once a homogeneous map is constructed, a centralized pathfinding algorithm can be applied. Our path-finding algorithm is to choose a set of paths, suggesting a proper number of robots along with their initial locations. In our work, robots can independently travel search locations, which may have dynamics or changes, but collaboratively cover all target locations. Through the experiments with real robotic platforms, we validate the generation of a map including targets and a choice of paths, and compare with existing algorithms.
