Dynamic graph-based search in unknown environments

A novel graph-based approach to search in unknown environments is presented. A virtual geometric structure is imposed on the environment represented in computer memory by a graph. Algorithms use this representation to coordinate a team of robots (or entities). Local discovery of environmental features cause dynamic expansion of the graph resulting in global exploration of the unknown environment. The algorithm is shown to have O(k.nH) time complexity, where nH is the number of vertices of the discovered environment and 1 <= k <= nH. A maximum bound on the length of the resulting walk is given.