An Explorative Study of Applying Reichardt-Model in Visual Servo Control

In this paper, a biologically inspired motion detector (Reichardt-Model) is introduced into visual servo control to ensure stability of the system with high gain and timedelay in its feedback. As a consequence of the specific velocity dependence of the Reichardt-Model, the stability margin of the visual servo control is increased and high overall gains, thus, better performance are achievable. As the first step of the study, the control performance of visual servo control with different sampling and different time-delay in its visual feedback loop, and velocity control performance with Reichardt-Model are investigated. A high-speed visual servo control system with up to 500Hz framerate is built up on a 1-DOF linear axis. With the results of simulation, the inner velocity control loop with Reichardt-Model is demonstrated to be more stable than standard model without Reichardt-Model when feedback gain increases and time-delay is presented. The results of real-time experiments show an improved performance by utilizing highspeed visual feedback or when the overall time-delay is small.