MEASUREMENT OF LINEAR VELOCITY USING A MOBILE ROBOTIC PLATFORM WITH COMPUTER VISION

Abstract

This article presents an approach to integrating computer vision algorithms into the control system of traction electric drives in rail transport. It demonstrates the utilization of computer vision algorithms for calculating linear velocity as an alternative to conventional sensors like wheel odometers, GPS, DGPS and inertial sensors, which may prove ineffective on slippery surfaces and at low speeds. As a result, this article focuses on employing linear velocity as feedback within the control system to enhance power efficiency during starting and stopping and to prevent wheel slip. The electric drive control system has been successfully implemented and tested on a robotics platform designed for simulating dynamic behaviors in real rail transports scenarios. The article details the development process of this robotics platform, which is employed to mimic real-world dynamic conditions in rail transport. The proposed control algorithm for speed estimation is assessed using a specially designed test bench, revealing its capability to predict speed with a relatively high degree of accuracy. Additionally, an optical flow algorithm for velocity estimation is introduced and evaluated through a specially designed test rig, indicating a strong correlation between the predicted vehicle speed and the measurements from precision optical encoders. The study also determines the optimal feature window size for real-time optical flow rate estimation. In summary, this approach exhibits significant potential for accurate speed estimation. Ongoing experiments are being conducted under various real-world conditions, with future research aimed at developing a dependable autonomous system for speed measurement. The integration of modern digital computer vision technologies not only enhances the traction characteristics of electric drives but also extends the capabilities of traction electric drives to meet the rigorous demands of industrial equipment.