A new robust visual servoing algorithm: theory and experiment

In the paper, a new kind of robust visual servoing algorithm is put forward. The method is based on the nonlinear proportional control law, with additional term to compensate for the parameter uncertainties. Theoretical analysis proves the stability and robustness of the system. When the real image Jacobian deviates from the nominal value too much, parameter estimation is performed based on Kalman filter. Application of this method to the visual tracking task shows it can successfully fulfill tracking task where traditional method cannot even work.