Analysis And Modeling Of Weight-bearing Gait Based On Multi-sensor

Walking with asymmetric load-bearing is one of the common daily behaviors.Office workers carrying briefcases to work,students carrying bags to school,athletes holding training equipment to train and people holding shopping bags to shop are all involved in the behaviors of asymmetric load-bearing.The presence of heavy objects on the human body affects posture,gait,and over time even changes the structure of the foot and damages the musculoskeletal system.The purpose of this study is to establish a model to evaluate the gait performance and analyze the gait difference under asymmetric loading,so as to explore the kinematics rule of lower limb segments.The main research work is as follows:1.Aiming at the problem that there is no public available wearable data set for asymmetric load-bearing,a multi-sensor based data set for load-bearing gait is established in this paper.The data set was carried out on a 12-meter-long laboratory corridor.Gait data were collected from eight subjects under five loading conditions.Each experiment was repeated three times,and four wearable sensors were placed in the middle of the subjects' shanks and thighs.2.In order to solve the problems of low data utilization and large random error in gait evaluation,a Regression Angle Offset Model(RAOM)was established in this paper.The model is optimized from the preliminary establishment of the model to the data diagnosis.The offset angle under different load-bearing states is used to measure the symmetry of gait,and the position span of subjects in each group of load-bearing modes is used to measure the movement consistency,which provides a new idea for the evaluation of symmetry and consistency in gait analysis.3.Gait symmetry,movement consistency and waveform similarity in the shanks and thighs are contrastively analyzed in this paper.In addition,the influence of weight-bearing side and weight-bearing degree changes on the step length parameters under five loading states was analyzed,and seven important discrete parameters in the cycle period were analyzed from the perspectives of variability and symmetry,so as to explore the movement differences between lower limbs' shanks and thighs.4.The proposed model is validated by applying the gait symmetry feature,motion consistency feature and waveform similarity feature to gait recognition.The recognition results of two sensor features in the lower shank segment and four sensor features in the lower shank and thigh segments on different feature sets were analyzed and compared.The results show that when the gait analysis features are applied to the load-bearing gait,the maximum recognition accuracy is increased by 14.39%,and the final classification accuracy reaches 92.64%.