The Design And Implemetation Of Gait Recognition System For Biped Robot With Heterogeneous Legs

Biped robot can simulate normal walking of human being comparatively well, but for abnormal walking like that of amputees, it can't do well. The intelligent prosthesis (IP) can restore walking function for above-knee amputee. During the research of IP, plentiful, various and repeated walking test of amputees fixed with IP is necessary. But in view of the physical limitation of amputees, it is very difficult to do such experiments. Biped robot with heterogeneous legs (BRHL) is a novel robot mode, which is integration of the research of biped robot and IP from the points of improving humanoid walking characteristics, enhancing walking function and providing research test-bed for IP.Following the description of the significance, contents and background of BRHL study, this paper gives detailed introduction of hardware designs of various parts of Under-bit machine which is detection system and design principles and design process software designs. Then this paper collects a lot of gait data in the basis of realization of this collection, and studies gait recognition in different movement states and new gait planning deeply.In order to make intelligent bionic leg follow the gait of artificial leg, gait identification of bionic leg is necessary. This paper simulates the biped robot gait data information using the information of inertial sensors and speed sensors that embedded in human bodies. It's the input signal of gait identification system. Respectively, this paper uses of Fast Fourier Transform method combined with BP Neural Network and Support Vector Machine to identify the input samples and obtains a better recognition result.IP should be used to amputee eventually and the training of its gait data is a important part of IP gait planning. So designing a good gait planning method which is suitable for the application of amputee is significant. This paper presents a new proof of the method of symmetry of the human gait. PCA gait data reconstruction method not only proves that the symmetry of human gait data, but also shows it is a good IP gait planning approach.