Dynamic Perception And Identification Of Human Motion Imbalance

Since human gait behavior contains information such as human kinematics,dynamics,and physiology,gait analysis is widely used in individual identification,lower limb rehabilitation monitoring,and human balance evaluation.The gait behavior of patients with stroke,head injury and spinal injury,as well as the elderly with weak physical fitness,will show varying degrees of balance dysfunction.In clinical practice,the balance ability of patients with dyskinesias is usually evaluated by visual inspection,scale or instrumental observation.These testing methods usually require patients and their families to go to the rehabilitation center and be carried out with the assistance of professional doctors and special equipment.Related technologies of wearable devices have been continuously developed in recent years,and their wide application has provided great convenience.Gait analysis based on wearable devices can be used to measure the balance ability of the human body.This article monitors human balance ability by estimating gait information and uses inertial sensors as the data source.Since the data collection process does not rely on special environments and equipment,the detection process of human balance ability is simpler.This article mainly consists of two parts: The first part is the construction of the gait analysis system,including the selection of data acquisition equipment and data receiving equipment.We selects inertial sensors as data collectors,smart phones as data receivers,and WIFI for data transmission.The mobile phone application can realize the functions of online display of motion curve and data storage,and the visual data collection process avoid the sudden interruption of data collection and other similar problems.The second part is to choose the signal and signal preprocessing scheme used for gait stage division.We analyzes the common division method of gait stage and divides the gait cycle into four stages.Since the angular velocity signal contains less noise,this article divide the input data of the model into the 15-dimensional temporal characteristic gait phase of the smoothed and filtered angular velocity.Then,HMM and parameter adaptive model are used to divide the gait stages,and analyzes the advantages of HMM compared to other commonly used classification models.Finally,according to the results of gait stage division,the gait space-time parameters related to the balance ability of the human body and the symmetry of the bilateral lower limbs are calculated,and the balance ability of the human body is analyzed according to the calculation results.The experimental results show that the model after HMM and parameter adaptation achieves a good recognition effect in the gait stage,and the final recognition rate reaches up to 91%.The analysis results of balance ability based on the gait space-time parameters and bilateral lower limb symmetry are consistent with ground truth.