Extending radar micro-Doppler analysis to various types of gait abnormalities

In recent years, the recognition and analysis of human gait abnormalities have attracted significant attention due to their critical applications in healthcare diagnostics, rehabilitation, biometrics, and security. This paper extends existing radar-based gait analysis by investigating and comparing micro-Doppler signatures (MDS) of various abnormal gaits, including normal walking, limping of varying severities, Parkinsonian gait, and hemiplegic gait. Using electromagnetic modeling and advanced signal processing methods, we simulate and analyze these gait types through their micro-Doppler signatures. We employ the Hilbert-Huang Transform (HHT) and the Continuous Wavelet Transform (CWT) to extract distinctive time-frequency characteristics unique to each gait pattern. Our analysis reveals that Parkinsonian gait exhibits significantly lower instantaneous frequencies, reduced bandwidth, and distinctive energy distribution across intrinsic mode functions, while hemiplegic gait demonstrates characteristic asymmetric patterns and intermediate complexity metrics. Heavy limping shows elevated bandwidth and frequency components compared to mild limping, enabling severity assessment. The combined application of micro-Doppler analysis with HHT and CWT provides complementary insights into the biomechanical differences between these gait patterns, suggesting potential applications for non-contact, objective assessment of movement disorders in clinical settings.