Propose a creative approach to identify wind direction using microphone of smartphone and attitude sensor. Build a model for relationship between noises generated by wind blowing microphone and incident angle of air flow. The approach relies on observation that the noises generated at two symmetrical incident angles by the same wind are similar. While incident angles change uniformly

the recorded wind noises has character of symmetry and symmetrical axis corresponds to angle of wind direction. Thus

propose a convolution-based circular autoconvolution（ CAC） algorithm to locate the symmetrical axis in sequence. Based on this algorithm

propose an approach

called Wind Compass

to calculate wind direction from wind noise signals. Implement the algorithm on Matlab and test effect of measurement.