Title:Artificial Potential Field-Based Path Planning for Cluttered Environments

Abstract:In this paper, we study path planning algorithms of resource constrained mobile agents in unknown cluttered environments, which include but are not limited to various terrestrial missions e.g., search and rescue missions by drones in jungles, and space missions e.g., navigation of rovers on the Moon. In particular, we focus our attention on artificial potential field (APF) based methods, in which the target is attractive while the obstacles are repulsive to the mobile agent. In this paper, we propose two major updates to the classical APF algorithm which significantly improve the performance of path planning using APF. First, we propose to improve an existing classical method that replaces the gradient descent optimization of the potential field cost function on a continuous domain with a combinatorial optimization on a set of predefined points (called bacteria points) around the agent's current location. Our proposition includes an adaptive hyperparameter that changes the value of the potential function associated to each bacteria point based on the current environmental measurements. Our proposed solution improves the navigation performance in terms of convergence to the target at the expense of minimal increase in computational complexity. Second, we propose an improved potential field cost function of the bacteria points by introducing a new branching cost function which further improves the navigation performance. The algorithms were tested on a set of Monte Carlo simulation trials where the environment changes for each trial. Our simulation results show 25% lower navigation time and around 300% higher success rate compared to the conventional potential field method, and we present future directions for research.