A new hopfield-type neural network approach to multi-goal vehicle navigation in unknown environments

A Hopfield-type neural networks (HNN) algorithm associated with histogram navigation method is proposed in this paper for real-time map building and path planning for multiple goals applications. In real world applications such as rescue robots, service robots, mining mobile robots, and mine searching robots, etc., an autonomous vehicle needs to reach multiple goals with a shortest path that, in this paper, is capable of being implemented by a HNN method with minimized overall distance. Once a global trajectory is planned, a foraging-enabled trail is created to guide the vehicle to the multiple goals. A histogram-based local navigation algorithm is employed to plan a collision-free path along the trail planned by the global path planner. A re-planning-based algorithm aims to generate trajectory while an autonomous vehicle explores through a terrain with map building in unknown environments. In this paper, simulation and experimental results demonstrate that the real-time concurrent mapping and multi-goal navigation of an autonomous vehicle is successfully performed under unknown environments.