A novel approach to reduce oscillations in path planning based on potential field approach, by applying Nesterov's accelerated gradient
A. Asadi, Sadjaad Ozgoli · 2017
This paper presents a method to reduce the oscillations of potential field-based path planning approach in presence of obstacles and in narrow passages. The oscillation issue might impede path planning progress and cause problems in implementation. The proposed method uses Nesterov's accelerated gradient (NAG) which is a first order optimization method. Compared to other first order methods used in potential field approach, NAG generates smoother paths with faster convergence rate. To determine the effectiveness of the proposed method, a comparative study has been conducted between performance of basic gradient descent, momentum and proposed method.