Photogrammetric calibration of a mobile robot model

A non-intrusive methodology to calibrate kinematic models for wheeled mobile robots is proposed. The calibration process begins with photogrammetric measurement of robot trajectories for known control inputs. Measurement is achieved with a low cost CCD camera. The camera is calibrated using a modified form of Tsai's direct linear transform based algorithm. Robot kinematic state is determined from a sequence of video data using a modified version of Lowe's pose recovery algorithm A derivative free form of the extended Kalman filter is used to estimate parameters for a known model from the time series kinematic data. Experimental results are presented for differential wheeled mobile robot configuration.