Improving the accuracy of laser self-mixing interferometry for velocity measurement

Self-mixing interferometry (SMI) employs fewer components, leading to a compact, robust and low cost system. In this paper, the main error sources of SMI for velocity measurement are analyzed and a simple error model is proposed. The dependence of signal amplitude and relative error on the incident angle is acquired by experiment. With the comparison of simulations and experiments, the proposed model is proved that relative errors could be reduced by decreasing the incident angle. To reduce the error of Doppler frequency estimation, the block splitting and averaging techniques are used. Finally, with a commercial laser Doppler velocimeter (LDV) the comparison between sensors demonstrates that the accuracy of SMI is close to that of LDV.