Research And Implementation Of Self-balancing Robot Based On ARM And FPGA

As an important member of wheeled robots,the two-wheeled self-balancing robot has the characteristics of simple structure,multiple control methods,and flexible movement.The two-wheeled structure enables it to gyrate in place,and it can play a great role in completing tasks in a small space.By adding various sensors and machine structures to the fuselage,it can also be applied to a wider range of fields,such as hazard investigation,fire alarm,industrial handling,intelligent logistics,short-distance transportation.The development prospects are very impressive.At the same time,the two-wheeled robot is an under-actuated and unstable nonlinear system,which can be used as a platform for verifying various motion control algorithms,so it is of great significance to the research of algorithms.In order to improve the operating efficiency and control ability of the system,this article proposed a master-slave control system based on ARM and FPGA,and STM32F103ZET6 and EP4CE6F17C8 N are used as the core control chips,which realized the detection operation of data and the parallel output control of the signal according to their own advantages.ARM is the main controller,which completes data acquisition,human-computer interaction and angle fusion tasks;FPGA is a slave structure,which receives angle data and motion commands from the main chip through SPI communication,completes error processing,parameter storage and fuzzy PID control output tasks.The two PWM and direction signals output by the FPGA are used as external input commands for the two servo drives.After the setting of current loop and the speed loop,the motor will rotate at the speed corresponding to the control quantity.Therefore,by adjusting the controller parameters,the output control quantity can be adjusted reasonably to the speed,and the dynamic balance of the two-wheeled robot is finally realized.This paper established a state space model based on the structure and motion characteristics of the two-wheeled robot,which proved that the self-balancing robot is an unstable and controllable system,and the key to maintaining balance lies in the design of the controller.Therefore,by studying the principles of traditional PID,fuzzy adaptive PID and LQR controllers,and building a simulation model on Simulink,it is concluded that under the same initial conditions,fuzzy PID control has both good speed and performance,while the PID and LQR control have overshoot or response speed problems.At the same time,the strategy research on its motion control is carried out.In this design,the motion amount is superimposed on the balance controller as an additional control amount in the manner of parallel control,so that the balance can be maintained during the motion.In addition,in response to the jitter problem that exists when the MPU6050 sensor collects data,a comparative experiment on complementary filtering,moving average filtering and Kalman filtering is designed.By observing the filtering effect of each filter in static and dynamic conditions,it is known that the Kalman filter has better dynamic and static characteristics,which can be used as a basis for data filtering and fusion.In the specific design,C language and Verilog HDL were used to complete the module design on the ARM and FPGA controllers,which realized the conversion from algorithm to program.Finally,the assembly of the two-wheeled robot and the joint debugging of software and hardware were completed,and the effectiveness of the filter and control algorithm and the anti-interference performance of the system were tested through balance,motion and robustness experiments.The experimental results prove that the master-slave controller based on ARM and FPGA designed this time can quickly adjust the balance of the robot,complete the user-specified movement,and has good stability and adaptability.