Background and purpose: High cost and complication of some instruments assessing functional capabilities in athletes with anterior cruciate ligament (ACL) deficiency have caused increasing attentions towards using inexpensive and user-friendly instruments as accelerometers and gait analyzing based on signal processing. Therefore, gait phases detection seems to be necessary while accelerometer is mounted on tibia. The aim of this study was to present a novel algorithm based on wavelet functions and dynamic time warping in order to gait detecting and gait breakdown from tibia acceleration raw data.

Materials and methods: A semi-experimental study was performed in which swing and stance phases of 20 healthy individuals and 20 ACL-deficient individuals’ gait were detected using up to order 32 of Mexican hat function of wavelet transformation. The acceleration signals were divided o peak searching parts. Time domain analyses were performed using peak detectors and dynamic time warping functions.

Results: In normal individuals and ACL-deficient subjects, 376 and 392 strides were detected from 391 and 415 strides, respectively. Also, in normal individuals from 17 distorted strides 14 and in those with ACL-deficiency all distorted strides were completely corrected.

Conclusion: The suggested algorithm was found to be capable of detecting swing and stance phases of the gait in both normal and ACL-deficient patients by accuracy rates of 94% and 95%, respectively. These rates are higher compared with previous studies. This novel algorithm could be applied as a useful method for gait evaluation in healthy people and individuals with ACL-deficiency.