A Novel Approach For Path Planning Based On Reactive Behavior-artificial Potential Filed

Path planning problem is one of the most important robotic problems for autonomousmobile robot. And the path planning is aimed at enabling robots with the capabilities ofautomatically deciding and executing a sequence of collision-free and safety motions in orderto achieve a certain tasks in a given environment. In substance, this field for autonomousmobile robot is an optimization problem that involves computing an optmal path betweeninitial location and goal location. In the current work, the main objective of this paper is tofocus on the local minima and the goal nonreachable with obstacles nearby (GNRON) issuesencountered in path planning by the artificial potential field (APF) method, and propose a novelapproach to plan a trajectory adaptive for the environment that the obstacles are randomlydistributed. The contents of this paper are organized as follows:1. Depicting the classical artificial potential field approach and presenting the twosiginificant inherent linitations although it is simiple and elegant. Investigating the currentlyexisting approaches that be related with this work, also the multiple algorithm fusion methodsof path planning. According to the main factors affecting the algorithm of relevant potentialfield method, made a brief analysis and rough classification.2. Introducing a modification of Newton’s method which overcoming inherentoscillation problem of potential field method (PFM) the presence of obstacles and in narrowpassages. This is a systematic approach to solving the oscillation problems in PFMs. Bycomparing its performance with the gradient descent method in obstacle-avoidance tasks withdifferent potential models and parameter changes.3. Elaborating the train of thought of reactive behaviors (RB)-artificial potential field(RB-APF) in detail. By combining the efficiency of the APF with the simplicity of the RB, so itcan be suitable for real-time application in mobile robots. The switch conditions and optimalselection equations are reasonably designed with the consideration of the differentcircumstances of the robot located in.4. Analysing the performance of simulations between the standard particle swarmoptimization (PSO) and PSO-immune algorithm. Also presenting the simulation result of theRB-APF. By comparing the two types of path planning algorithm, that based on evolutionaryalgorithm and reasoning algorithm, analysis their individual advantages, disadvantages andapplications.