Walking Intention Recognition Method Based On Proximity Sensor Array

With the acceleration of global aging,the contradiction between the medical care needs and the pension and nursing resources has intensified.It makes the daily assistance and nursing intelligence of the elderly become a hot spot in the field of robotics.As an important requirement in the daily life of the elderly,walking assistance determines the quality of their life directly.It is great significance to complete the daily walking training of the elderly through intelligent robots to improve the participation of their life.The optimization of the hardware environment and the recognition strategy are used to improve the interactive experience and the compliance of the user's speed intention recognition during in walking training based on the omnidirectional walker(ODW).First,the structure of the ODW is introduced.Based on the mechanical structure,the VL53L0 X laser unit and the STM32F103C8T6 minimum system are used in the independent development of eight-channel laser array sensors.The effectiveness of the system has been verified in lab-environment.Through the analysis of walking data,a gait detection system based on a vertically distributed three array laser sensor was built.The motion feature detection mechanism is established by calculating the coordinates of the user's body center position in the local coordinate system.The improved particle filtering based on back propagation neural network(BPNN)is proposed as a compliant recognition strategy.The back propagation characteristic of BPNN is used to optimize the distribution of particle weights during the re-sampling process of particle filtering,which solves the problem of particle degradation in the particle filtering algorithm of importance sampling.Finally,the effectiveness of the recognition strategy was proved by the analysis of walking training experimental data.Then,in the results of the comparison experiment and the subjective evaluation,the accuracy of speed intention recognition was improved by 8.27%,and the subjective interactive experience was also improved.