Error modeling and improved position estimation for optical incremental encoders by means of time stamping

Optical incremental encoders are extensively used for position measurements in motion systems. The number of slits on the encoder disk defines the resolution of the encoder and bounds the accuracy of the position measurement. The encoder position measurements suffer from quantization errors. Moreover, encoder imperfections occur due to manufacturing tolerances. In this paper a method, which is based on time stamping, is proposed to obtain more accurate position and velocity estimations. Time stamping makes use of stored events, consisting of the encoder counts and their time instants, which are captured at a high resolution clock. Low order polynomial fitting through these encoder events is combined with active compensation of the encoder errors. The proposed method is applied in real-time experiments to a motion system. The results show an improvement in the position accuracy of up to 87% in terms of the maximum error. The estimated velocity is also much more accurate than the differentiated quantized encoder output signal.