This paper employs a combination of radio frequency identification (RFID) technology, infrared distance sensing technology, and a geomagnetic induction electronic compass to design a navigation and control system for an autonomous mobile robot based on DSP control and information fusion. During the design process, some improvements were made to the traditional infrared obstacle avoidance sensor according to actual needs. Furthermore, the design of the drive frame for the DSP system in the information fusion state was researched and discussed. The design of the motion control drive circuit for a small robot is presented. Finally, the robot designed under this architecture was subjected to actual stability testing in a supermarket. 1. Introduction With the development of science and technology and the improvement of people's living standards, robots have begun to enter people's lives. The advent of this era has brought about various new types of robots, such as cleaning robots and security robots. One of the most important parts of a mobile robot—the navigation system—has attracted even more attention in the field of robotics. Robot navigation systems are crucial for many robot application areas, such as intelligent warehouses, supermarket guides, home robots, automated libraries, and intelligent hospitals. Line-following robots are relatively common. During my graduate studies, the author participated in several robot competitions and designs of this type. In actual design, it was found that due to the limitation of fixed lines, this system cannot achieve truly omnidirectional autonomous operation. Furthermore, this method is greatly affected by light and cannot be practically applied in daily life. Lines on the ground also affect the aesthetics of the surface. After thorough literature review and consideration, a new robot navigation system was proposed, integrating RFID, geomagnetic induction, DSP, and other technologies. Actual system hardware and software design and stability testing were conducted. 2 Navigation Hardware System Our developed hardware system mainly consists of: a DSP core board, an RFID board, a system motherboard, an electronic compass, and a driver board. 2.1 DSP Core Board Our self-designed DSP system uses the TMS320F2812 as its core. The 2812 is a high-performance, multi-functional, and cost-effective 32-bit fixed-point DSP chip from TI for control. The core uses a high-speed processor, mainly considering future secondary development of the system and convenient algorithm porting. The instruction set can operate at a maximum frequency of 150MHz and has 18k of on-chip SRAM with zero wait cycles and 128k of on-chip FLASH with an access time of 36ns. Its on-chip peripherals mainly include an ADC, dual SCI, SPI, McBSP, eCAN, etc., and it also has event management modules (EVA, EVB), including 6-channel PWM/CMP, 2-channel QEP, 3-channel CAP, and 2-channel 16-bit timers. The dual SCI can connect one channel to a sensor and the other to a PC for timely output of debugging information. The 16-bit PWM enables fine-tuning of speed. CAP and other components facilitate interfacing with sensors. The 16-channel, 12-bit ADC with a maximum input of 3V and a conversion time of 80ns can simultaneously connect to 16 distance sensors. The 2812 has a 16×16-bit dual multiplier-accumulator, providing sufficient processing speed for RF and geomagnetic direction signals. Due to the special characteristics of the 2812's core voltage (1.8V) and oscillation frequency, this core board uses a passive crystal oscillator. The core board uses a TPS767D318 dual-output low-dropout LDO. Dual power supplies are provided to ensure a stable 1.8V core voltage supply. The TPS767D318 has a high-speed transient response, is dedicated to DSP, and can provide a maximum current of 1A. It can realize time-sharing reset of IO and core. It has power-on reset function, low voltage protection function, and its reset delay time is 200ms. It has overheat protection function. The maximum power can be calculated using Equation 1. (1) Where TJMAX is the maximum allowable temperature, which is generally around 125 degrees according to experience. TA is the ambient temperature. RθJA is the connection impedance. For 28 pins, it is usually 27.9°C/W. The actual power consumption can be calculated using Equation 2, where VI and VO are the input and output voltages, respectively, and IO is the output current. (2) In terms of external voltage filtering, a multi-stage tan capacitor (104 and 220uF) is connected in parallel. This design can produce a low ESR effect. According to experience, in the frequency domain, 104 can greatly filter out high-frequency interference, while 220uF can best reduce low-frequency interference of voltage. 2.2 System Mainboard and Sensor Module Regarding power supply, the system uses a 12V lithium battery, regulated by a three-terminal positive voltage regulator. Internal power protection is integrated. The output current can reach 1A. The input withstand voltage can reach 30V. A charging interface is provided, controlled by a switch; the circuit board has an interface for charging the lithium battery. The positioning device uses radio frequency identification (RFID) technology. RFID has become a popular technology. Recently, Walmart passed a resolution requiring its top 100 suppliers to use RFID technology when sending pallets and boxes to its distribution centers before January 2005, and to gradually use this technology on individual items after 2006. From the basic principle of information transmission, RFID technology is based on a transformer coupling model (energy and signal transfer between primary and secondary windings) in the low-frequency band and a spatial coupling model (radar detects targets and returns target information to the radar receiver after the radar emits electromagnetic wave signals that encounter the target) in the high-frequency band. The most basic RFID system consists of three parts: tags, readers, and antennas. Based on their effective range, obstacle avoidance cards can be categorized into closely coupled cards (effective range less than 1 cm), near-coupled cards (effective range less than 15 cm), sparsely coupled cards (effective range approximately 1 meter), and long-range cards (effective range from 1 meter to 10 meters, or even further). This system uses a near-coupled card. The RF module communicates with the SCIA port of the 2812. By decoding the data stream, the robot's position is determined. Next is the infrared obstacle avoidance module. Generally, robots can avoid obstacles using infrared reflection, which is common in robot competitions. GPIO controls the emission of infrared light, and if an obstacle is encountered, it will be reflected back. The receiving tube receives the light, causing a change in resistance, and detecting this change in resistance determines whether an obstacle is present. However, this method is susceptible to light noise interference, resulting in a relatively short range, typically only 2-3 cm. Through research and investigation in numerous competitions, I have proposed a circuit that uses a standard high-frequency signal of 38kHz infrared light for obstacle detection. Because it uses a high-frequency signal and a high-frequency operational amplifier, it has a certain anti-interference capability, and the maximum detection range is increased to 8cm. First, infrared light is emitted via a 555 timer. Next, the signal passes through the infrared receiver tube and then through the DC blocking capacitor, before being sent to the high-frequency operational amplifier LM318N, as shown in Figure 4. Then, it is amplified by 50 times. As shown in Figures 5 and 6. Figure 1 555 transmitting circuit [align=center] Figure 2 Frequency identification circuit [/align] [align=center] Figure 3 Detection amplifier circuit [/align] The LM567 is a phase-locked loop circuit, with an 8-pin dual in-line package. The resistors and capacitors connected to pins 5 and 6 determine the center frequency of the internal voltage-controlled oscillator. Pins 1 and 2 are usually grounded through a capacitor, generating an output filter network and a single-stage low-pass filter network for the loop. The capacitor connected to pin 2 determines the capture bandwidth of the phase-locked loop: the larger the capacitance value, the narrower the loop bandwidth. The center frequency and filter bandwidth of the voltage-controlled oscillator can be determined by equations 3 and 4. (3) (4) (where Vi is the input voltage) Then there is the electronic compass module. The CMPS03 electronic compass is a plane angle sensor. By detecting the angle between the current sensor and the Earth's magnetic field, the electronic compass can obtain an absolute rotation angle with a resolution of 0.1 degrees. This electronic compass module is specifically designed for robots to provide them with appropriate directional navigation signals. It can generate a unique code for any direction. The sensor uses the PHILIPS KMZ51 geomagnetic sensor chip, which is highly accurate. There are two output methods: the first is I2C, output via Pin 2 (SCL) and Pin 3 (SDA). Pins 7 and 5 must be left floating. Pin 6 is used for calibration. These pins are all connected to the motherboard. Since the module is powered by 5V while the DSP is 3.3V, a 74LVC245 is needed for level conversion. The system is calibrated via Pin 6 through the 2812's GPIOB. Calibration only needs to be performed once, as the data is stored in the EEPROM of the PIC microcontroller within the electronic compass. Pin 6 has a pull-up resistor. Calibration only requires sending a negative pulse to Pin 6 via GPIO, and because of the pull-up resistor, this pin can also be disconnected from the system. Finally, the design of the system's driver module is discussed. The L298 chip is used. The L298 is a product of SGS, with the 15-pin Multiwatt packaged L298N being a common example. Internally, it also contains 4 channels of logic driver circuitry. The power output devices within the L298 are designed and fabricated on a quartz substrate. Due to the uniformity of the manufacturing process, it boasts performance parameter consistency and stable operation that is incomparable to discrete component combination circuits. Pin 15 is the output current feedback pin, otherwise identical to the L293. In normal use, these two pins can also be directly grounded. It is a high-voltage, high-current dual full-bridge driver chip. It can directly accept standard TTL logic levels. It can drive various loads such as motors and relays. It has two enable inputs, which control the validity of the PWM wave. The L298 integrates two energy output blocks, A and B. Additionally, our designed board includes a freewheeling diode. 2.3 Driver Design The InitSysCtrl() function is used to configure the watchdog control register WDCR, where the WDFLAG bit is the watchdog reset status bit. Setting this bit indicates a watchdog reset. Writing 1 to the WDDIS bit disables the watchdog module, while writing 0 enables it. The WDPS bit primarily determines the clock rate of the watchdog counter. Due to the numerous loops in the program, special attention must be paid to the watchdog settings. Then, the PLLCR is configured to set the system's phase-locked loop. Note that the program needs to include a 5000-cycle wait for phase-locked loop stabilization. This is not necessary in the 2407, but is important for the 2812 system. High-speed and low-speed peripherals are then matched using HISPCP and LOSPCP. Since interrupts are used in the program, peripheral interrupt extensions (PIEs) need to be configured. The direction sensor is connected to the DSP core via the capture unit. Capturing a single pulse requires two interrupts, and the pulse code for direction information is retrieved by controlling and reading the FIFO register. In addition, the infrared obstacle avoidance module is connected to the interrupt pin after level conversion via a 74LV245 module. Therefore, the interrupt function needs to be programmed. Communication with the RFID sensor is through the 2812's two-wire asynchronous serial port. The SCI module supports digital communication between the CPU and asynchronous peripheral devices using the non-return-to-zero (NRZ) standard format. The 2812's SCI receiver has a 16-level deep FIFO, which reduces idling. The program checks the TxRDY bit to determine if there is an RFID interrupt. This allows for timely detection of whether the robot has reached a new position. Then, the SCIRXBUF is read. Finally, the PWM pulse is set through the EV unit to control the robot's movement. 3 Conclusion and the Implemented System Finally, the entire system was debugged near Lianhua Supermarket. After repeated debugging and program modifications, navigation to RF path points was achieved. The final system operation is shown in Figure 6: Figure 4 Actual operation of the robot in the supermarket References: [1] Song Zhenhua, Zhan Xingqun, Zhang Yanhua. Underwater robot navigation system based on multi-sensor fusion[J]. Microcomputer Information, 2008, 1-2: 240-241. [2] Ding Yi, Qiu Baijing, Zhou Ning, Cui Jun, Qian Guohong. Research on GPS/DR combined navigation system of lawn mowing robot[J]. Agricultural Mechanization Research, 2006, 4-4: 68-71. [3] Wang Zhi, Liu Jianye, Liu Bo. Design and implementation of SINS/Two-Antenna GPS combined navigation system (English)[J]. Transactions of Nanjing University of Aeronautics & Astronau, 2006, 3-1: 41-46. [4] Ziparo, VA; Kleiner, A; Nebel, B; Nardi, D. RFID-based exploration for large robot teams [J]. 2007 IEEE International Conference on Robotics and Automation, ICRA'07.2007, 4-1,: 4606-4613.