Corridor planning for natural agents

Certain agents and systems of agents, such as animals, possess innate path-planning capability, both locally and globally. If we wish to influence their motion, we may be able to take advantage of their local navigation ability and provide only some simple imposed constraints rather than computing and enforcing a detailed path. In this work, we consider the problem of generating a set of constraints, or "corridor", under which such an agent (or group of agents) produce an appropriate trajectory. The set of constraints should be both small in number and efficient in terms of length. When producing the corridor, we explicitly consider two different types of obstacles - those which the agents themselves would naturally avoid, such as dense vegetation, and those which they must be forced to avoid, such as sensitive stream environments. This allows both for simplification of the corridor and more natural use of the underlying navigation ability. An implementation and several examples of planned corridors are also presented