Saifang Technology teaches you how to use RISC-V development board to realize automatic line patrol driving of small cars

Recently, DFRobot has developed a line patrol car exhibit, which uses a RISC-V main control development board equipped with the Saifang Technology JH7100 (Fang Jinghong 7100) chip and runs the Fedora Linux operating system. With the support of the PinPong Python library, the Mind+ programming software is used for GPIO control. The car uses the HUSKYLENS (Erha Image Recognition) visual sensor to perform intelligent route recognition and realize the function of automatic line patrol.

1. Exhibit Description

1. Saifang Technology JH7100

Fang Jinghong 7100 is equipped with dual-core U74, shares 2MB of L2 cache, operates at up to 1.2GHz, and supports Linux operating system. Among them, the ISP independently developed by Saifang Technology is compatible with mainstream sensors, and the built-in image and video processing subsystem supports H265/H264/JPEG codecs. The integration of high-performance, low-power Vision DSP and NNE makes application development smarter and more efficient. Fang Jinghong 7100 can complete various complex video image processing and intelligent visual computing to meet a variety of real-time visual processing requirements at the edge.

2. PinPong Library

The PinPong library is a Python library for controlling open source hardware main control boards. It is based on the Firmata protocol and is compatible with MicroPython syntax. You can use Python to control open source hardware in 5 minutes. With the help of the PinPong library, you can program various common open source hardware directly with Python code. The principle is to burn a specific firmware to the open source hardware so that the open source hardware can communicate with the computer through the serial port and execute various commands. The design of the PinPong library is to allow developers to focus on the business implementation of software without being bound by complicated hardware models during the development process. Even if the program is developed with Arduino in the early stage of writing and changed to the main control development board during deployment, it can run normally by modifying the hardware parameters, realizing "write once and run everywhere". With the joint efforts of engineers from DFRobot and Saifang Technology, the PinPong library has been fully adapted to the Saifang Technology RISC-V main control board, and the PinPong library can be used smoothly to directly read various sensors and operate actuators.

3. Mind+

Mind+ is a domestically produced youth programming software with independent intellectual property rights. It integrates various mainstream main control boards and hundreds of open source hardware, supports artificial intelligence (AI) and Internet of Things (IoT) functions, and can not only drag graphical building blocks for programming, but also use advanced programming languages ​​such as Python/C/C++, allowing everyone to easily experience the fun of creation. Mind+ not only supports graphical programming, but also supports Python programming. By simplifying a series of tasks such as environment configuration, we can achieve the state of being able to work as soon as we open it. At present, DFRobot has completed the adaptation work for Saifang Technology's JH7100 SoC, and can use a low-code approach to develop Saifang Technology's RISC-V main control board, further reducing the difficulty of developing Saifang Technology's VisionFive single-board computer.

4. HUSKYLENS (Husky Image Recognition) Vision Sensor

This line patrol car uses HUSKYLENS (Erha Image Recognition) as a visual sensor. As an easy-to-use artificial intelligence camera (visual sensor), HUSKYLENS has 6 built-in functions: face recognition, object tracking, object recognition, line patrol tracking, color recognition, label recognition, object classification, QR code recognition, and barcode recognition. AI training can be completed with just one button, getting rid of tedious training and complex visual algorithms, allowing developers to focus more on the conception and implementation of the project. HUSKYLENS has an onboard UART/I2C interface that can be connected to mainstream development boards such as Arduino, Raspberry Pi, LattePanda, Microbit, etc., to achieve seamless hardware docking and directly output recognition results to the controller. You can create very creative projects without having to toss complex algorithms. This car exhibit has achieved the adaptation of HUSKYLENS and the RISC-V main control development board of Saifang Technology.

HUSKYLENS is very easy to use. You can switch algorithms by toggling the function button; press the learning button and the HUSKYLENS can learn new things. It comes with a 2.0-inch IPS display. When adjusting parameters, there is no need for computer assistance. The debugging process and recognition results are directly displayed on the screen, and what you see is what you get. The built-in machine learning technology enables it to have the ability to recognize faces and objects. By long pressing the learning button, it can learn from different angles and ranges, and the more it learns, the more accurate it becomes.

2. Implementation steps

1. Function

a) Use PinPong to drive the car and use Erha to recognize images to patrol the line;

b) Erha View communicates with Saifang Technology JH7100 using I2C bus;

c) Use 2 IO ports to control the motor direction;

d) Use 2 PWM ports to control the motor speed;

e) Start up and run automatically.

2. Features

a) Automatic operation;

b) Scalable development;

3. Usage steps

a) After the motherboard is powered on and enters the system, /root/linefollower.py will be automatically executed immediately;

b) The learning of line patrol recognition needs to be controlled manually. Click the button in the upper right corner of HUSKYLENS to learn the black line to be recognized. When the white line with an arrow turns into a blue line, it means that the learning is completed;

c) After learning is completed, you can put it on the line patrol track and start patrolling the line.

This case shows the huge potential of the RISC-V development board of Saifang Technology in the field of robotics, and it can be well adapted to various sensors for IoT control. In the next stage, Saifang Technology will cooperate with DFRobot to jointly develop more IoT industrial control applications based on Saifang Technology's VisionFive single-board computer and the new generation of JH7110 chip, and promote the application of IoT.