IMU-Based Real Time Four Type Gait Analysis and Classification and Circuit Implementation

Nowadays, foot gait analysis has been widely used in the diagnosis and treatment of neuromusculoskeletal diseases and has received much more attention. The motion of the human hip joint can be simplified into three degrees of freedom: flexionextension, exhibition-adduction, and internal-external rotation motion. We can encompass these types of hip movements by using four types of walking gait: normal walking, tandem walking, toe walking, and heel walking. In this paper, we proposed a real-time and applicable system, which has low computational complexity while maintaining high accuracy for four kinds of gait analysis and recognition with a single Inertial Measurement Unit (IMU) sensor. The orientation is estimated by fast complementary filter (FCF) and Attitude and heading reference system (AHRS). Then the linear acceleration is integrated twice and calculated many spatial-temporal parameters of gait can be obtained, and the zero velocity update (ZUPT) method is used to suppress the integral drift. Experimental results show that the error rate of the four type of walking stride length can reach 1.34%, 1.65%, 2.26%, 2.13% respectively. The system has a recognition accuracy rate of 83.3%. Implemented with TSMC 0.18 µm process. To achieve a low power design, set the clock frequency to 45 kHz. The performance of the chip achieves an area of 2.47 x 2.47 mm2 and a power consumption of 0.1198 mW.