Fixed-point computing with variable fraction length for motion profile generation

This paper proposes a new fixed-point computing method for motion profile generation. When parameters have wide ranges, floating-point computing usually is used for accurate computation; however, this method increases execution time. In the proposed method, by considering input data, optimum fraction lengths are determined for different parameters. In this way, the fixed-point computing offers an accurate computation in a minimum execution time. Experimental results show the performance of the proposed method.