Latency Reduction for Telerobotic Applications

This paper tackles the problem of reducing the latency in telerobotic applications. The time delay between an action of the human operator and the corresponding reaction of the system has a negative impact on the synchrony, disturbs the operator's immersion, and yields adverse effects like simulator sickness or kinetoses. In this context we present an application to control a stereo camera system by measuring the operator's head motion. To decrease the latency, the head motion is described as Gauss‐Markov model and predicted with a Kalman filter. This estimation provides a reduction of the mean orientation error by 70% and the maximum error by 50% in contrast to an unpredicted control of the vision system. Experiments with subjects showed an improved compatibility due to the reduced latency.