Research On The Perception System Of Wearable Gait Orthosis

Wearable gait orthosis is a type of lower limb rehabilitation robots, and it was ahot research area in medical rehabilitation in recent years. Wearable gait orthosisrelates to a lot of discipline domains, such as human physiology, ergonomics, controltheory, information processing, etc, which is aimed at helping the old and the patientsto train and improve their lower limbs athletic ability.Researching on the perception system of wearable gait orthosis is a veryimportant part of lower limb rehabilitation robot system. Automation of perceptionsystem directly affects the mechanical control of gait orthosis and the trainingeffectiveness. Body weight support treadmill training (BWSTT) is a traditional wayof lower limb rehabilitation training. It depends on doctor’s hands-on experience. Sothe automation level and efficiency of rehabilitation training by this way is rather low.Different from BWSTT, wearable gait orthosis depends on intelligent perceptionsystem with multi-sensor. It can solve the problems of traditional ways and improvesthe efficiency of rehabilitation training.In order to acquire the gait information of lower limbs motor disturbance patientsand help their gait to recover, a perception system with multi-sensor and high speeddata transmission network was designed. The perception system get the hydrauliccylinder force information by force sensors, get angle and angular velocityinformation by rotary encoder on joint, get acceleration information by accelerometer,as well as get real-time walking gait information by the pressure sensors under foot.After condition circuitry processing, the sensor signal collected by data acquisitioncard and transferred to the host computer by a high-speed synchronous datatransmission network. Generally those information contain high frequency noise, sodifferent Butterworth low-pass filter to remove high-frequency noise for differentsignal characteristics was designed to restore the sensor truthful information. Throughanalyzing the information after digital filtering and the human real-time motions, thefeasibility of perception system was verified. At the same time, the human gait cyclewas divided by analyzing the changes of human plantar pressure when walking.Based on different gait dividing, there are different gait recognition methods: method based on the plantar pressure information to determine and classification methodbased on Euclidean distance. The actual data of walking experiment can be used tovalidate the feasibility of the plantar foot pressure information can be used to detecthuman gait and the rationality of the layout of pressure sensor under foot. Finally itcan monitor and recognize real-time gait by the plantar pressure information. Theseprovide a basis for gait orthosis to plan normal gait and control it to walk normally.