Application of TMS320F2812 in Vehicle Four-Wheel Steering Control System

Digital Signal Processor (DSP) is a microprocessor suitable for digital signal processing operations, which can realize real-time and fast digital signal processing algorithms. Usually, a DSP-based core, equipped with peripheral functional circuits required for measurement and control, is integrated into a single chip, which greatly reduces the chip price, reduces the size, makes the structure compact, convenient to use, and improves reliability. Therefore, the multi-function board with integrated DSP chip is an ideal choice for motor applications, excitation pulse control systems, and power protection systems. This paper uses the CS4U9813 programmable intelligent multi-function board of Chaotuo Industrial Control as the controller of the servo motor to realize the rear wheel steering function of the car. The board integrates TI's high-performance DSP chip TMS320F2812 .

1 Structural features of TMS320F2812[1][4]

TMS320F2812 is the most advanced and powerful 32-bit fixed-point DSP chip on the market launched by TI. It has both digital signal processing capabilities and powerful event management capabilities and embedded control functions, and is particularly suitable for industrial automation, motor, and motor servo control systems. The TMS320F2812 chip structure adopts an improved Harvard structure that separates program and data storage, 8-level pipeline operation, 128-bit key protection, almost all instructions are completed within 6.67ns (150MHZ), up to 1MB external memory interface, and up to 56 independent programmable, multi-purpose input/output (GPIO) pins. It is the best choice for realizing motion control systems. The functional block diagram is shown in Figure 1, and its main performance is as follows:

Figure 1 Functional block diagram

(1) High-performance 32-bit CPU

• Main frequency 150MHZ (clock cycle 6.67ns), low power consumption (core low voltage 1.8V, I/O port 3.3V)
• 16-bit × 16-bit and 32-bit × 32-bit multiply-and-accumulate operations and 16-bit × 16-bit dual multiply-and-accumulate
• Unified register programming mode, up to 4M words of linear program address and data address

(2) On-chip memory

• 8Kx16-bit Flash memory
• 1Kx16-bit OTP read-only memory
• L0 and L1: two 4Kx16-bit single-port random access memory (SARAM)
• H A 8Kx16-bit single-port random access memory
• M0 and M1: two 1Kx16-bit single-port random access memories

(3) Clock and system control

• Supports dynamic change of phase-locked loop frequency (PLL)
• On-chip oscillator
• Watchdog Timer Module
• A control system combining CPU-level and peripheral-level interrupts

(4) Rich peripheral devices

• Two event managers (EVA, EVB)
• Serial Peripheral Interface (SPI)
• Two serial communication interfaces (SCI), standard UART
• Enhanced Controller Area Network (ECAN)
• Multi-channel buffered serial port (MCBSP)
• 16-channel 12-bit digital-to-analog converter (ADC)

3 Implementation of four-wheel steering control system

Four-Wheel Steering (4WS) system means that during the steering process of the vehicle, the front and rear four wheels can steer as needed, which can effectively improve the vehicle's maneuverability and handling stability, and is being continuously developed and applied. When the 4WS car turns at low speed, the front and rear wheels steer in opposite phases, which can reduce the turning radius of the vehicle; when turning at high speed, the front and rear wheels mainly steer in the same phase, which can reduce the vehicle's center of mass side slip angle β, reduce the steady-state overshoot of the vehicle's yaw rate, etc., and further improve the vehicle's handling stability.

The four-wheel steering platform built in this paper is composed of a DC servo motor with a worm gear reduction mechanism, an electromagnetic clutch and an ordinary front wheel steering mechanism. The motor power supply is DC12V, the reduction ratio is set to 10: 1 (adjustable), and the electromagnetic clutch power supply is DC48V. The experimental platform is shown in Figure 2. The control strategy adopts direct yaw rate closed-loop control combining feedforward and feedback, in which the front wheel angle signal is collected by an absolute angular displacement sensor; the rear wheel angle is obtained by a high-precision incremental photoelectric encoder; the vehicle speed is obtained by a five-wheel instrument; the rear wheel steering is achieved by a DC servo motor connected to a mechanical steering mechanism through an electromagnetic clutch; the core of the ECU hardware electronic system is composed of TMS320F2812, and the structural block diagram is shown in Figure 3:

Figure 2 Experimental platform

Figure 3 Structure diagram

4 Control System Hardware Design and Simulation

According to the control strategy designed above, the ECU unit needs to collect the front wheel angle signal, rear wheel angle signal, and wheel speed signal and output the driving voltage of the motor through CPU calculation to realize automatic steering of the rear wheel. This article uses C language to compile programs to implement the designed control algorithm, and completes hardware-in-the-loop simulation debugging in the CCS (Code Compose Studio) environment when the simulator is connected. CCS2000 is a Windows-based DSP integrated development environment provided by TI for the TMS320C2000 series DSP, and is also the best DSP development software at present. In the CCS environment, program development, debugging, compilation, and linking can be carried out, and software development in assembly and C/C++ is supported. Powerful debugging tools such as breakpoints, probes, analysis, and graphic display can be used, and the output target file can be burned in the end.

4.1 Front wheel angle signal acquisition

The front wheel angle is obtained by an absolute angular displacement sensor, and the output voltage is linearly proportional to the front wheel angle, with a range of 0~12V. This article uses the ADCHA0 pin of the CS4U9806 board for sampling, and the sampling peripheral circuit DSP chip is integrated internally. This board has a single-channel sampling time of 200nS, and the input signal range can be selected by jumpers -5~+5V, 0~10V, 0~20V, and the default is -5~+5V. Due to the limited range of the front wheel angle, the jumper selects 0~10V and the sampling frequency is set to 96HZ. Part of the program design process is as follows:

4.2 Rear wheel angle signal acquisition

The rear wheel angle signal is obtained by a high-precision incremental photoelectric encoder. It converts the angle of the worm into corresponding increase or decrease count pulses according to the direction of rotation. Each rotation generates 2048 pulses, and the output is a clock signal and a direction signal. This paper uses the DICH0 (CAP2) pin to capture the clock signal, and the input pin DICH19 (GPIOF12) to obtain the direction signal. The design of the switch input circuit with optical coupler is shown in Figure 4, and the capture unit circuit is integrated inside the DSP chip.

Figure 4 Input circuit block diagram

Part of the programming process is as follows:

4.3 Algorithm Design and D/A Output

The basic idea of ​​the algorithm is: when 4WS is started, a front wheel angle is input, and through the yaw rate feedback, it is compared with the ideal yaw rate steady-state response gain G0 related to the speed, and then the rear wheel angle is controlled by the controller G1 to achieve four-wheel steering, and the center of mass slip angle, yaw rate, and lateral acceleration are output for monitoring. The control block diagram is shown in Figure 5. The drive voltage of the motor is generated by the comparison register of the DSP to generate a PWM signal, which is then output through the D/A conversion circuit. This paper uses timer 4 to generate PWM, which is output by the DA4 pin, and the first-stage amplification D/A conversion circuit is shown in Figure 6.

Figure 5 Control block diagram

Figure 6 D/A conversion circuit

Part of the programming process is as follows:

This paper applies the control algorithm designed above to simulate the 4WS vehicle based on TMS320F2812 DSP hardware in the loop. The vehicle parameters are: m=1740kg; I= 3214kg.m2; a=1.058m; b=1.756m; K1=29000N/rad; K2=60000N/rad; and for simplicity, the interception speed u= 30m/s. The results are shown in Figure 7. The results show that the yaw rate tracking control of the 4WS vehicle can maintain better dynamic and steady-state performance than two-wheel steering in the high-speed range of steering control. At the same time, CCS2000 is used for debugging and development, which has a short cycle and low cost. Therefore, the hardware control system based on DSP has broad prospects in the fields of industrial control and automotive control.

Figure 7 Simulation results