A new neural network for robot path planning

This paper presents a new methodology based on neural dynamics for robot path planning. The target activity is treated as an energy source injected into the neural system and is propagated through the local connectivity of neurons in the state space by neural dynamics. The elegant properties of harmonic functions are incorporated in the neural system by formulating the local connectivity of neurons as a harmonic function. An improved Hopfield-type neural network model is established for propagating the target activity among neurons in the manner of physical heat conduction, which guarantees that the target and obstacles remain at the peak and the bottom of the activity landscape of the neural network, respectively. The real-time collision-free robot motion is planned through the dynamic neural network activity without any prior knowledge of the dynamic environment, without explicitly searching over the global free workspace or searching collision paths, and without any learning procedures. Examples are presented to demonstrate the effectiveness and efficiency of the proposed methodology.