Loading...

A Fuzzy-Behavior Based Approach to “Dynamic and Static Obstacles” Avoidance in an Unknown Space

Khan Mohammadi, Hadi | 2013

1266 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45185 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Sayyadi, Hassan
  7. Abstract:
  8. In this article, research has been carried out on the control technique of autonomous mobile robots to navigate in unknown environments, avoiding static and moving obstacles using eight ultrasonic sensors for obstacle recognition.
    In this article we used simple three wheeled robots. Each robot has a passive caster wheel and two active fixed wheels which can act independently. In addition robots have holonomic constraints like maximum acceleration, maximum speed and maximum angular velocity, so robot’s speed and orientation change independently and these changes are small in each time step. Each robot has eight ultrasonic sensors, each of them cover 22.5 degree of space from robot‘s point of view. Each ultrasonic sensor is capable of detecting distance from nearest obstacle in its region of coverage and the component of “relative velocity of robot-nearest obstacle” in direction of the line which passes through robot and obstacle centers.
    A fuzzy behavior-based approach has been used in robot’s controller. Fuzzy system is useful in uncertain environments but the number of its rules increases exponentially with increase in the number of variables. To overcome this problem, behavior based structure has been used. In this structure the main task of robot divide into elementary tasks called behaviors. Each behavior has access to sensors data and produces an output relevant to its goal disregard of other behaviors’ goals, thus the problem of designing a complex controller splits into designing several simpler controllers and design of a behavior-coordinator. The devised controller has seven behaviors which are responsible for achieving simple goals like “dynamic consistency”, “Goto”, “obstacle avoidance”, etc. Devised behavior-coordinator task is combining output of all behaviors to pursue robot’s main task. Each behavior in the controller gives a fuzzy vote for points in speed-angle space using its internal logic and regarding to its own goal. At the end behavior-coordinator combines these votes using behavior priorities and type of combination which devised before.
    The robot main task is to navigate between two predefined points in the space in an unknown complex environment with static and moving obstacles. To avoid obstacles, controller uses information from ultrasonic sensors to detect possible obstacle zones in next time step. Using these information Obstacle-Avoidance Behaviors of controller generate their votes to guide the robot through risk free areas.
    To investigate proposed controller ability to guide the robot through desired environments several simulations has been performed. Simulation’s results showed the good ability of devised controller to guide the robot toward final goal through complex, dynamic and unknown environment with reasonable speed
  9. Keywords:
  10. Moving Obstacles ; Navigation ; Obstacle Avoidance ; Behavior-Based Control ; Cooperative Movement

 Digital Object List