Font Size: a A A

Gait Detection System Of Lower Extremity Exoskeleton Robot

Posted on:2017-03-26Degree:MasterType:Thesis
Country:ChinaCandidate:H JiFull Text:PDF
GTID:2308330485469442Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
With the development of society, there’s more needs for advanced equipment to improve human’s life. Exoskeleton robot is a power assist device that can supply a great help in daily life and work. In order to design an effective lower extremity exoskeleton robot that can really assist people, a lot of research on gait detection system has been done in this paper. Gait detection is the key of robot control that determines the assistance effect directly.In the present exoskeleton system, the precision of gait detection is not high enough for the control requirements. The main reasons include the simplification of gait information and the misuse of recognition method and other external factors. These problems have been studied in this article, include three main parts as follows.1. The design of gait information collection systemThe design of gait information collection system includes two parts:hardware design and software design. In the design of hardware, the selection of microcontroller, sensors and data transmission mode are needed to build the whole hardware circuit. In software design, information collection of sensors by microcontroller, data transmission between upper and lower computers, and software development of the upper computer acquisition interface are carried on.2. Gait information collection and preprocessingAfter completing gait collection system, ten volunteers are required to conduct gait information acquisition experiments. Then these gait data will be pre-analyzed for regular observation. According to the specialty of these data, appropriate data preprocessing methods have been applied, include data filtering, normalization and principal component analysis, and so on.3. Gait recognitionSupport vector machine and neural network are used for building the gait classifier model. In the model, grid search and genetic algorithm are applied for algorithm parameter selection. Four gait recognition methods are compared through experimental results. GA-SVM is proved to be the best gait detection method in exoskeleton system.
Keywords/Search Tags:exoskeleton, gait, recognition, SVM, NN
PDF Full Text Request
Related items