UNCALIBRATED 2D ROBOTIC VISUAL TRACKING BASED ON ARTIFICIAL NEURAL NETWORK

摘要

In this paper, without explicit external and internal calibration, we propose a nonlinear visual mapping model for the eye-in-hand robotic visual tracking problem, which connects the image feature space with the robotic work space tightly. Moreover, a new visual control scheme based on artificial neural network is designed, and the visual tracking problem is converted into a servo problem in image feature space. Simulation results show that this method can drive the static tracking error to zero quickly and keep good robustness and adaptability at the same time. Additionally, the algorithm is very easy to be implemented with low computational complexity.