Research On Gait Recognition Of Lower Limb Exoskeleton Robot Based On Support Vector Machine

Lower Limb Exoskeleton Robot is an anthropomorphic mechanical system for being worn on human's body parallel to the lateral side to enhance the wearer's strength and endurance,which integrates the humans' intelligence and the machine's physical power.At present,exoskeleton has many practical applications.In the field of military affairs,exoskeleton can help soldiers march in high speed with bearing more weapons,and can reduce the human resource losses caused by injuries or non-combat loss of members which were due to huge body energy consumption.In the field of medical treatment,exoskeleton can assist the medical staff to conduct rehabilitation training on the patients of stroke or lower limb injury with higher precision and higher repeatability.In order to achieve the Lower Limb Exoskeleton Robot's fast,effective and stable control,one needs to identify the gait information.Whether the recognition is fast and accurate will directly affect the whole exoskeleton robot control system.Therefore,designing an accurate gait recognition system is of great significance to the exoskeleton robot control.The emphases of this dissertation are as the follows: Firstly,On the basis of analysis of human lower extremity kinematics,the DoFs of each joint of the exoskeleton are configured to meet the requirements of the common gaits.By comparing the typical human gait phase model,a “five-phase” model is put forward.Then,the SVM and TWSVM working principle and the characteristics of three typical multi class SVM are analyzed.Multi-classification methods for nonparallel dual hyperplane SVM are discussed.A multi-class LSTSVM classifier based on optimal directed acyclic graph is given.A class separability criterion based on distance measure and the corresponding category number rearrangement method are provided in order to determine the class divisibility,its purpose is to overcome the cumulative errors cased by the traditional DAG structure.Finally,we design a set of Lower Limb Exoskeleton Robot mechanical structure,and then a sensing system which detects movement acceleration and foot pressure is designed to get human movement information,and filtering algorithm,feature extraction is used to process the motion information.Experiments prove that the method has impressive performance in the test accuracy and execution speed.