Design Of A Multi-sensor Fusion-based Motion Sensing System For Lower Limb Rehabilitation Robots

With the progress of society and the development of human technological civilization,lower limb rehabilitation robots have rapidly developed in the field of medical rehabilitation.The demand for rehabilitation medicine has also increased.Due to the relatively late development of rehabilitation medical equipment in China,although most hospitals have rehabilitation departments,they do not have professional rehabilitation training equipment and rely solely on physicians to manually bend patients’ joints to assist them in rehabilitation training.This method is inefficient and the rehabilitation techniques of the treating physician are difficult to achieve uniformity,resulting in uneven rehabilitation outcomes.Although a few rehabilitation departments have simple lower limb rehabilitation trainers,these traditional trainers have no sensors added,or the addition of sensors is only for optimizing a certain function of the robot and does not have a motion perception system,so intelligent rehabilitation cannot be achieved.The main content of this study is to accurately identify the patient’s movement intention,determine the active/passive rehabilitation training mode,and help physicians accurately obtain various data of the patient’s body to achieve intelligent rehabilitation effects through the addition of multiple sensors.Based on the above issues,this paper conducts the following research:1.Research on the interaction between the exoskeleton of a mobile lower limb rehabilitation robot and the patient’s legs.The auxiliary force during patient rehabilitation training walking is provided by the exoskeleton.Based on the structural characteristics of the exoskeleton,the interaction mode between the exoskeleton and the patient’s leg,the driving principle of the joint motor,the arrangement position,structure,and function of the leggings are clarified.2.Establish an exoskeleton dynamics model,use the Lagrange equation to establish a lower limb exoskeleton dynamics model,and obtain the driving torque of the motion joint and the human-machine coupling force model.By analyzing the force simulation between the exoskeleton and the human body,the position of the interaction force is obtained and the design scheme of the human-machine interaction force sensor is planned.3.Design of rehabilitation robot sensors.The acquisition of human-machine interaction force is an important foundation for the motion perception of lower limb rehabilitation robots,and is the early stage for determining the active/passive rehabilitation training mode.In order to accurately collect the pressure data of patients’ legs,a leggings bar type interactive force sensor based on strain gauge is designed to achieve the acquisition of human-machine interaction force.Sensors located at the large and small legs collect the interaction forces at different positions of the legs for data complementarity,and cooperate with weight loss sensors to achieve data fusion.Use four tension sensors located above the patient’s body to collect the patient’s tension value data,and determine the patient’s gait cycle based on the variation pattern of the tension value,achieving the acquisition of weight loss and gait cycle.Associate and fuse gait cycle data with interaction force data,and use neural network algorithms to determine active/passive modes.4.Design of signal processor.To reduce the electromagnetic interference and crosstalk of the sensor’s output signal,a hardware filtering circuit is used to perform primary filtering on the sensor’s output signal,and then a Kalman filtering algorithm is used to perform secondary filtering on the signal.The smooth output of the sensor signal is achieved through a combination of software and hardware filtering.5.Design a motion sensing system with multi-sensor fusion and conduct experimental verification.Integrate interaction force data with gait cycle data.Design a motion perception experiment based on multi-sensor fusion.By collecting experimental data from volunteers in the laboratory and conducting clinical tests in hospitals to collect experimental data from patients,a control group is established to verify the accuracy of active/passive mode judgment.