A neural network model for route planning constraint integration

The ability to plan routes that avoid obstacles and achieve mission goals in a timely fashion is a requirement in ground, air, and undersea autonomous systems. A neural network solution to route planning based upon the Hopfield model is presented. The translation of the terrain information into the Hopfield network representation is discussed. Primary emphasis is placed upon the generation of an energy function capable of representing the planning and mission constraints common to all three operating domains. Utilizing the energy function based upon goal point distance, terrain gradients, and feedback of the vehicle´s altitude data, results are shown demonstrating the route planning capability on actual terrain database imagery. Ongoing efforts to integrate the route planner into an overall vehicle planning scheme are discussed