To achieve on-demand assistance for lower limb exoskeletons

a method for lower limb gait recognition based on piezoresistive sensors and inertial measurement unit(IMU)is proposed.A collection system is designed based on these two kinds of sensors

including hardware setup and software development

to realize real-time data acquisition and storage of sensor data.Simultaneously

a study on multi-sensor intention recognition is conducted.A comprehensive architecture of gait recognition system is established and offline training and online recognition algorithms are implemented.Experimental results demonstrate that the accuracy of offline model training achieves 99.64%

while the accuracy of online gait recognition is 89.92% with an average latency of 19.8ms.This method effectively distinguishes various lower limb gait patterns

achieving high-precision real-time recognition

and thus supporting the control of exoskeleton assistance.