Walking robots, control algorithms, adaptation, vision system

The article presents the concept of trajectory generation algorithms for a reconfigurable walking robot based on a vision system and real-time image analysis. A developed six-legged robot design and a three-level microprocessor control system were used. A control algorithm is proposed that enables the robot to adapt to its environment based on image analysis and sensors (proximity and contact sensors in the legs) by selecting the appropriate direction of movement, avoiding/crossing obstacles, and choosing postures and gait types. Image analysis was performed using the OpenCV library based on a constructed simplified statistical Haar feature classifier. Field tests were conducted under various environmental conditions, yielding promising results.THE CONCEPT OF VISION SYSTEM BASED TRAJECTORY GENERATION ALGORITHMS FOR THE HEXAPOD ROBOT
Abstract
The paper describes the concept of trajectory generation algorithms for the reconfigurable walking robot using vision system and real time captured picture analysis. The design of the insect based robot with six legs and three-level control system is used. Based on vision analysis and contact sensors a control algorithm is proposed. The algorithm allows the robot to adapt to the environment by selecting the direction of movement, passing / crossing obstacles, a selection of posture and gait. Image analysis are carried out using the OpenCV library and constructed on the basis of a simplified statistical classifier Haar’a features. Several tests for different environmental conditions are made, giving promising results. (…)

Walking robots, control algorithms, adaptation, vision system – PDF file for download