The virtual wall approach to limit cycle avoidance for unmanned ground vehicles

https://doi.org/10.1016/j.robot.2007.11.010Get rights and content

Abstract

Robot Navigation in unknown and very cluttered environments constitutes one of the key challenges in unmanned ground vehicle (UGV) applications. Navigational limit cycles can occur when navigating (UGVs) using behavior-based or other reactive algorithms. Limit cycles occur when the robot is navigating towards the goal but enters an enclosure that has its opening in a direction opposite to the goal. The robot then becomes effectively trapped in the enclosure. This paper presents a solution named the Virtual Wall Approach (VWA) to the limit cycle problem for robot navigation in very cluttered environments. This algorithm is composed of three stages: detection, retraction, and avoidance. The detection stage uses spatial memory to identify the limit cycle. Once the limit cycle has been identified, a labeling operator is applied to a local map of the obstacle field to identify the obstacle or group of obstacles that are causing the deadlock enclosure. The retraction stage defines a waypoint for the robot outside the deadlock area. When the robot crosses the boundary of the deadlock enclosure, a virtual wall is placed near the endpoints of the enclosure to designate this area as off-limits. Finally, the robot activates a virtual sensor so that it can proceed to its original goal, avoiding the virtual wall and obstacles found on its way. Simulations, experiments, and analysis of the VWA implemented on top of a preference-based fuzzy behavior system demonstrate the effectiveness of the proposed method.

Introduction

Navigation in unknown, very cluttered environments is a difficult task in mobile robotics. Behavioral approaches have been shown to be very successful in these environments [2], [6], [12]. Robot navigation in these types of environments has motivated the development of behavioral architectures that can tolerate information uncertainty and the ability to simultaneously accommodate the needs of the different behaviors. Recently, Dunlap, Selekwa and Collins [6] developed and implemented a fuzzy preference-based behavioral system which achieved smooth navigation in cluttered environments.

However, experimental results have shown that behavioral systems that are goal directed (i.e., systems in which one of the behaviors is goal seeking) tend to fail when the robot enters an enclosure that has its opening in a direction opposite to the goal. In scenarios like this, behavioral systems usually cause the robot to become trapped in deadlocks. In particular, the robot exhibits limit cycle behavior in which it retraces the same path indefinitely, thereby failing to accomplish its mission of reaching the goal.

One way of escaping these deadlock enclosures has been by inclusion of deliberative algorithms that use a stored map of the environment to detect and retract from the deadlock regions. While these strategies have been effective in some applications, many of them have been computationally demanding, resulting in a slow retraction from the deadlock enclosures. This paper presents an algorithm that uses local maps and virtual sensors to solve the problem of deadlock detection and avoidance. Although the proposed limit cycle negotiation approach is independent of the navigation algorithm, the fuzzy preference-based behavioral system for navigation in cluttered environments [6], [12] was selected to test the VWA due to its ability to generate smooth trajectories.

This paper is organized as follows. Section 2 presents the background necessary to better understand the proposed limit cycle negotiation approach. Section 3 details the proposed approach to solve the limit cycle problem. Section 4 gives an overview of the fuzzy preference-based behavioral system that is used to navigate the robot. Section 5 shows simulation results. Section 6 presents experimental results. Concluding remarks are given in Section 7.

Section snippets

Background

This section provides the background necessary for a better understanding of the proposed approach for limit cycle avoidance.

Proposed approach: The virtual wall algorithm (VWA)

This section describes a new simple deliberative approach for detection and avoidance of navigational limit cycles in cluttered indoor or outdoor environments with various sizes and shapes of potential deadlock regions. The basic flowchart of the proposed algorithm is given in Fig. 3. The strategy was conceived trying to emulate the reasoning of a human being in a similar situation. Imagine a human in a very cluttered or maze-like environment seeking to reach a goal that is at a known location.

The preference-based fuzzy behavior control system used to test the VWA

This section presents a brief description of the control system used to test the VWA. However, it is important to notice that the proposed approach is general and it could be implemented on top of any behavioral or reactive system.

The behavioral system used to navigate the robot from a starting point to a desired target is a fuzzy preference-based behavioral system. The system has a heading control action that is achieved using four behaviors: (1) goal seeking, (2) obstacle avoidance, (3) left

Simulation results

The proposed method was simulated on top of the preference-based fuzzy behavior control system described in Section 4. Two type of simulations were performed as described in the following two subsections. In Section 5.1 the VWA was compared against some of the alternative methods developed by other researchers, while in Section 5.2 the VWA was tested using cluttered environments that emulate dense forests.

Experimental results

After satisfactory simulation performance, the proposed strategy for limit cycle negotiation was implemented and tested in a laboratory environment on a Pioneer 2 robot equipped with a SICK laser range finder (Fig. 27). This robot, which is manufactured by ActivMedia Robotics, is a differentially driven platform configured with two drive wheels and one swivel caster for balance. Each wheel is driven independently by a motor with 19.5:1 gear ratio, which enables the robot to drive at a maximum

Conclusion

A method for detecting and avoiding navigational limit cycles, which is one of the major shortcomings of behavioral systems, has been proposed. This method uses dynamic local spatial memory to keep track of places visited in order to detect limit cycles. Whenever a point in spatial memory records that one place has been repeatedly visited in excess of the allowable levels, a limit cycle is said to have been detected. Once the limit cycle is detected, a virtual wall is built to close the

Disclaimer

The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government.

Camilo Ordonez is a Ph.D. candidate in Mechanical Engineering at Florida State University. He received his M. S. degree in Mechanical Engineering from Florida State University in 2006 and also holds a B.S. degree in Electronics Engineering from Pontificia Bolivariana University in Colombia. During his undergraduate studies Camilo was an exchange student in the Department of Electrical Engineering at Concordia University in Montréal, Canada. As part of his exchange program, he served as a

References (14)

  • H. Maaref et al.

    Sensor-based navigation of a mobile robot in an indoor environment

    Robotics and Autonomous Systems

    (2002)
  • ActivMedia Robotics, 44 Concord St., Peterborough NH, 03458. Pioneer 2/PeopleBot Operations Manual, 9 edition, October,...
  • R.C. Arkin

    Behavior-Based Robotics

    (1998)
  • T. Balch, R. Arkin, Avoiding the past: A simple but effective strategy for reactive navigation. In: Proceedings of the...
  • H. Choset et al.

    Principles of Robot Motion: Theory, Algorithms, and Implementations

    (2005)
  • Auto Identification Division, Technical Description: LMS 200 Series Laser Measurement Systems. SICK, Inc., 6900 West...
  • D.D. Dunlap, Implementation Of Multi-Valued Fuzzy Behavior Control For Navigation In Cluttered Environments, Master’s...
There are more references available in the full text version of this article.

Cited by (37)

  • Sensor data fusion for optimal robotic navigation using regression based on an IOT system

    2022, Measurement: Sensors
    Citation Excerpt :

    The length of the robot's trip to the target point, as evaluated by the study, reveals good performance. In Ref. [20], a different approach to solving the local minima problem is suggested. To locate and measure the local minima, the suggested method in Ref. [21] constructed a grid map that matched the grid and presented the examined occupied cells in the grid while traveling.

  • Reactive versus cognitive vehicle navigation based on optimal local and global PELC<sup>∗</sup>

    2017, Robotics and Autonomous Systems
    Citation Excerpt :

    In general, this school of thought (the reactive one) does not require high computational complexity since the robot’s actions must be given in real time according to local perceptions [34]. Obviously while taking this approach for robot’s navigation, the overall robot movement (from its current robot location to the final desired location) cannot be considered as optimal, mainly if the environment is complex with a multitude of obstacles and maybe trapped regions [35]. In fact, since the robot’s actions are guided only by local perceptions, the global optimality cannot be demonstrated in advance.

  • A new hybrid navigation algorithm for mobile robots in environments with incomplete knowledge

    2012, Knowledge-Based Systems
    Citation Excerpt :

    Nevertheless, as pointed out before, most existing reactive navigation methods may get stuck in traps without the assistance of a deliberative layer, which is the well-known local minima problem. While many efforts [28–30] have been made to enhance the robustness of the reactive navigation in the face of trap-situations, to our knowledge, only a family of algorithms termed the Bug algorithms [31] have strictly proved their convergence and therefore thoroughly solve the local minima problem in theory. The Bug algorithms navigate the robot by two reactive motion modes: moving toward the goal if no obstacle is encountered, otherwise following the obstacle boundary until a leaving condition is satisfied.

  • Robotic path planning in static environment using hierarchical multi-neuron heuristic search and probability based fitness

    2011, Neurocomputing
    Citation Excerpt :

    Shibata et al. [30] used Fuzzy Logic for fitness evaluation of the paths generated. Various other approaches [6,26] have also been proposed. Zhu and Latombe [38] used the concepts of cell decomposition and hierarchal planning.

View all citing articles on Scopus

Camilo Ordonez is a Ph.D. candidate in Mechanical Engineering at Florida State University. He received his M. S. degree in Mechanical Engineering from Florida State University in 2006 and also holds a B.S. degree in Electronics Engineering from Pontificia Bolivariana University in Colombia. During his undergraduate studies Camilo was an exchange student in the Department of Electrical Engineering at Concordia University in Montréal, Canada. As part of his exchange program, he served as a research assistant at the Groupe de Recherche en Perception et Robotique (GRPR). After receiving his B.S. degree, Camilo joined Total Seguridad Ltda as a support engineer in the area of electronic security systems. His research interests include mobile robotics and control systems.

Emmanuel G. Collins Jr. is a John H. Seely Professor of Mechanical Engineering and Director of the Center for Intelligent Systems, Control and Robotics (CISCOR) at the joint College of Engineering of Florida A&M University and Florida State University. He received his Ph.D. in Aeronautics and Astronautics from Purdue University in 1987 and also holds B.S. degrees from Morehouse College and the Georgia Institute of Technology. He spent 7 years in research and development at the Harris Corporation, prior to joining the Department of Mechanical Engineering as an Associate Professor in August 1994. Dr. Collins teaches courses in controls, robotics and dynamics. His current research interests include intelligent control systems for autonomous vehicles, human–robot interaction, fault detection and isolation, control in manufacturing, applications of advanced robust control techniques, numerical algorithms for control law design, and control of propulsion systems.

Majura F. Selekwa is an Assistant Professor of Mechanical Engineering and Director of the Mechatronics Laboratory at North Dakota State University. He received his Ph.D. in Mechanical Engineering from Florida A&M University in 2001, an M.Eng.Sc degree in Mechatronics from the University of New South Wales in Australia and a B.Sc. in Mechanical Engineering from the University of Dar es Salaam, Tanzania. His research interests include numerical algorithms for control system design, application of artificial intelligence in controls and robotics, behavior robotics, and mechatronic systems design. Before taking his current position he worked as a Visiting Assistant Research Professor in robotics at the Center for Intelligent Systems, Control and Robotics (CISCOR) in the joint College of Engineering of Florida A&M University and Florida State University. Dr. Selekwa has also held academic positions in the Department of Mechanical Engineering of the University of Dar es Salaam, in Tanzania and has practiced as a process control engineer at Process Control Solutions Inc. in Tallahassee, Florida.

Damion D. Dunlap is currently a Ph.D. student in the Mechanical Engineering Department of Florida A&M University. He is an active researcher in the Center for Intelligent Systems, Controls and Robotics (CISCOR) where his primary research interest include kinodynamic path planning and mobile robot navigation in cluttered environments. He received his M.S. and B.S. degrees in Mechanical Engineering from Florida A&M University in 2004 and 2002 respectively.

Prepared through collaborative participation in the Robotics Consortium sponsored by the U. S. Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD 19-01-2-0012. The U. S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation thereon.

View full text