Walking robot method control using artificial vision

  • Authors:
  • Stefan A. Dumitru;Danut A. Bucur

  • Affiliations:
  • Robotics and Mechatronics Department, Institute of Solid Mechanics of the Romanian Academy, Bucharest, Romania;Robotics and Mechatronics Department, Institute of Solid Mechanics of the Romanian Academy, Bucharest, Romania

  • Venue:
  • Proceedings of the 15th WSEAS international conference on Systems
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

The paper presents an intelligent control algorithm of a walking robot for avoiding obstacles in the work environment. Algorithm is based on artificial vision and is composed of four modules: capture and image analysis, neural controller, Bezier interpolation and generation of new support points. The first module processes captured image and provide data for a neural controller. Neural controller provides according to received data the coefficients of rotation to left and to right used to calculate the angle of rotation that robot must made on next step. Bezier interpolation module calculates the deviation from the old path and recalculate a new the robot path. The last module provides data to a robot leg control algorithm. The results obtained by applying this algorithm indicates that you can to control the robot in a fast feedback loop and that it can generate the part to the target point on the shortest path with avoiding obstacles in the robot environment.