| Lower limb exoskeleton robots have been a hot area in recent years,and they have extensive application and broad prospects in military and rehabilitation filed.One basic requirement for the exoskeleton systems is they must follow human's movement stably,so it is essential to recognize human's gait correctly for their control system.Generally,different sensors are used for gait recognition,which makes the sensor system of exoskeleton more complex and heavier burden on data processing.Interactive force between foot and ground is a kind of very important gait data when one is walking.In order to cut down sensor quantity,the paper analyses the relation between the foot pressures and joint angles.To recognize human gait with foot pressure,the gait is redefined according to the relation in this paper.And because of time-delay,the exoskeleton must predict the human gait so that to track his action timely.This paper presents a set of foot pressure measurement system powered by battery,which consist of a pressure insole and a signal module.It can collect,amplify,sample and store the pressure signal without influencing normal walking.Working with attitude transducers,this system can gather the pressure and angle data simultaneously.And the gait features are got from pressures and angle data respectively,and this paper also analyses the correlation between the two kinds of gait signal with correlation coefficient and monotonicity.Based on this correlation,the gait cycle is redefined,and here also defines two gait characters,which are used to do the recognition.Then the prediction of pressure can be completed with adaptive algorithms LMS and RLS.And the algorithms' best parameters are decided by simulation.The predicted data is used in gait recognition to achieve the goal of gait prediction.Lastly,both recognition and LMS prediction algorithms are implemented in signal module.And design some experiments to prove the correction of algorithms.From the experimental results,they works well. |
PDF Full Text Request