Research And Design Of Auxiliary Eating Manipulator System Based On OpenMV

With the emergence of the aging problem,the elderly and disabled robot,as an important branch of the service robot,has become a hot research topic.In this paper,aiming at the dining problem of the elderly with inconvenient hands and feet or the special group of upper body amputation,a design scheme of auxiliary eating manipulator based on machine vision is put forward.According to the design requirements of the system,the kinematics analysis and joint space trajectory planning of the manipulator,face detection algorithm,template matching algorithm and PID algorithm based on face tracking are studied and analyzed.The software and hardware design of the system are analyzed in detail,and the implementation of the algorithm is verified.The research contents include:1.The model of manipulator is modeled by D-H method,then the kinematics analysis is done,and the results of forward and inverse kinematics analysis are simulated and verified.Then the joint space trajectory planning is carried out by cubic polynomial interpolation and the point-to-point(PTP)motion experiments are carried out to verify that the motion stability of the manipulator is satisfied.2.The organization structure and detection principle of face detection classifier based on HAAR feature Adaboost are analyzed.The training method of classifier based on HAAR feature face detection classifier is presented and transplanted into embedded system.Then the NCC template matching algorithm is studied and analyzed.Then the face pixel-based PID controller algorithm is designed and the test results are analyzed.Finally,a debugging software is designed and two teaching modes,point-to-point instruction and multi-stage trajectory planning,are developed.3.The hardware of the control system is introduced.The hardware circuit of the system is designed,including power circuit,voice module,camera module,steering gear module and WIFI module based on ESP8266.The main function modules are debugged in stages.4.The software design of the system mainly includes the transplantation of FreeRTOS system on STM32F103ZET6,the driving of voice module LD3320 A and the software scheme to improve the recognition rate.According to the training method of face classifier proposed in chapter 3,the software realization of face detection function is given,and the software implementation flow of vision-based PID face tracking algorithm and custom frame communication protocol on different platforms is given.Then,a detailed scheme of manipulator access to intelligent cloud Internet of things platform is given.Finally,the whole system is debugged and verified.