Real-Time Path Planning in Unknown Environments Using a Virtual Hill

Abstract

The artificial potential field based path planning has been most wisely used for local path planning because it provides simple and efficient motion planners for practical purposes. However, this approach has a local minimum problem which can trap a robot before reaching its goal. The local minimum problem is sometimes inevitable when a mobile robot moves in unknown environments, because the robot cannot predict local minima before it detects obstacles forming the local minima. The avoidance of local minima has been an active research topic in the potential field based path planning. In this study, we propose a new concept using a virtual hill to escape local minima that occur in local path planning for a mobile robot. A virtual hill is located around local minimum to repel a robot from local minimum. Copyright © 2005 IFAC Keywords: mobile robot, path planning, artificial potential field, virtual hill, extra potential, navigation. ITRODUCTIONArtificial potential field methods provide simple and effective motion planners for practical purposes (Lee and Park, 1991). These approaches have been widely applied to path planning of a mobile robot and a manipulator (Borenstein and Koren, 1989;Chuang, 1998;Chuang and Ahuja, 1998;Chuang et al., 2000;Guldner and Utkin, 1995;Haddad et al., 1998;Hwang and Ahuja, 1992;Tsai et al, 2001;Vadakkepat, 2001;Veelaert and Bogaerts, 1999). The applications of artificial potential field for obstacle avoidance was first developed by Khatib (Khatib, 1985;Khatib, 1986). This approach uses two types of potential, which are a repulsive potential field to force a robot away from obstacles or forbidden regions and an attractive potential field to drive the robot to its goal. The robot moves under the action of the artificial force which is proportional to the negative gradient of artificial potential. The robot is driven from the positions with the higher potential to that with the lower potential.However, the path planning by the artificial potential field approach has a major problem, a local minimum problem, which can trap a robot before reaching its goal. The local minimum problem is sometimes unavoidable in local path planning, because the robot can detect only local information on obstacles. In other words, the robot cannot predict local minima before experiencing the environment. An avoidance of a local minimum has been an active research topic in potential field based path planning (Chang, 1996;Cho and Kwon, 1996;Connolly et al., 1990; Connolly, 1992;Janabi-Sharifi and Vinke, 1993;Kim and Khosla, 1992; Lee and Park, 1991;McFetridge and Yousef-Ibrahim, 1998;Park and Lee, 2003; Rimom and Koditschek, 1992;Volpe and Khosla, 1990). However, the previous solutions were limited to simple formations of obstacles or available for path planning in known environments.

Real‐time motion planning for mobile robots by means of artificial potential field method in unknown environment

“…The concept of wall following has been widely used for the navigation of mobile robots. It also can be applied as an LME method for solving local minima issue in the APE method (Borenstein and Koren, 1989; Yun and Tan, 1997; Park et al , 2005; Mabrouk and McInnes, 2008). These methods believe that the key obstacle causing local minimum is the obstacle lying between the robot and the goal.…”

“…Furthermore, Park et al (2005) have proposed an implicit wall‐following method for solving the local minima issue of the APF method. The concept of this method is similar to the VO method (Park and Lee, 2003).…”

“…These factors lead to the advantage that most LME methods can be utilized on‐line in complete unknown environments. There are a lots of LME methods, such as the combination of searching algorithm and the APF method (Sharifi and Vinke, 1993; Barraquand et al , 1992; Barraquand and Latombe, 1991), virtual obstacle (VO) method (Park and Lee, 2003; Li et al , 2000), sub‐goal method (Bell and Weir, 2004; Weir et al , 2006), wall‐following method (Borenstein and Koren, 1989; Yun and Tan, 1997; Park et al , 2005; Mabrouk and McInnes, 2008), etc. Nevertheless, since heuristic strategies only based on local environment information are usually applied, many LME methods cannot guarantee to converge in general.…”

Bidirectional Potential Guided RRT* for Motion Planning

“…However, its disadvantage is that it is easy to fall into a local minimum point [2], where the resultant force of the robot is 0 and the target pose cannot be reached. To solve this problem, many scholars have proposed improvements, such as introducing virtual target points [3] or making robots try to move randomly [4], using the simulated annealing algorithm [5], using the adding extra control force method [6], applying the genetic algorithm into artificial potential field method [7], introducing the gain factor [8], applying a virtual obstacle concept [9] or virtual obstacle method [10], [11] to get rid of the current minimum. Jinseok Lee proposed an internal state model to solve the local minimum problem with computational cost which is low relatively [12].…”

A New Algorithm of Path Planning Based on Local Data

“…A lot of methods are proposed to solve the local minimum problem; the super-quadratic potential field method [6] and the harmonic functions method [7] are used to remove the local mina with improved potential function. Some methods try to solve this problem by escaping from local mina such as virtual obstacle method [8,9], sub-goal method [10,11], wall-following or bug method [12,13] and the artificial moments method. Among these improved ones, the bug method gives a better solution to guide the robot to escape from the trap, but the methods proposed above have a less optimal procedure to the known environment, which will be investigated in this paper .…”