机器人与自动化的进步

A Robotic Path Planner Contender

Abstract

Kamkarian P and Hexmoor H

This article presents a novel offline path planner method to yield a collision-free trajectory among groups of obstacles in a static workspace. It enables a single holonomic point robot acting in a static environment including a fixed initial and goal configurations to achieve its goal toward a collision-free trajectory. In developing our novel path planner, we focused to elevate features that help the planner to route in a wide variety of different situations in regards to lowering the processing time needed for analyzing the workspace and determination of ultimate trajectory. Unlike to some other planners that are able to be applied on some certain obstacle shales such as polygonal, our planner is skillful enough to analyze any types of obstacles, such as circular, spiral, and curved edged obstacles successfully. To increase the performance of our proposed offline path planner, we defined and benefitted from introduction of parameters that help to achieve the best possible results among different scenarios in workspace components arrangements as well as reducing the planner needing to access system resources such as memory or processing unit to analyze the workspace while computing the shortest collision-free path from start point to the goal configuration. The novel planner analyses and transforms the two dimensional representation of the environment into a roadmap consisting a graph of nodes along with all possible routes from the moving robot’s initial point into the goal configuration. In order to manage some of the popular problems such as being trapped inside a U-shaped obstacle or routing in narrow pathways among obstacles, that challenge other offline path planners such as Potential Field planner are facing, we used a multi-layer approach in form of different stages to help the planner considering all possible circumstances and hence, compute the best possible route.