基于Raspberry Pi的拖拉机通用自动驾驶系统

为了改善现阶段拖拉机自动化驾驶系统主要针对单个特定型号的拖拉机进行开发的局限性，该文设计了一套采用分布式控制方式的拖拉机自动化驾驶系统。该系统的作用在于同SAFAR系统（software architecture for agricultural robot，农业机器人的软件体系结构）整合，成为一个能够进行实际工作的系统。系统采用Raspberry Pi微型计算机作为系统的电子控制单元（electronic control unit，ECU），在不同的ECU上分别处理不同的任务，利用通讯协议传输控制协议/互联网络协议来进行ECU之间的通讯。为了保障系统的安全，该系统采用激光测距传感器R2100来保障车辆和周围环境的安全，采用Heart Beat（心跳）检测机制来探测系统故障。该系统可以接收来自SAFAR的控制信号，并对拖拉机进行相应的操作来实现拖拉机的自动化驾驶。除此之外，该系统还提供了无线遥控的功能，能够方便操作人员进行简单、直观的操控。经过测试，该系统能够正确、快速地响应来自操作人员或SAFAR的控制，控制系统的响应时间低于0.5 s，电机转向角误差小于0.06；舵机角度偏差4°左右；系统各个ECU之间通讯稳定可靠，可在试验设定的4 h内稳定工作，每0.25 s检测一次工作环境并给予响应。该研究对农业机械自动化的实现提供了参考。

Universal autopilot system of tractor based on Raspberry Pi

Abstract: With the process of industrialization, young workers pour into the city, the number of farm labor declines. The average age of farmers is increasing. The tractor, as a modern agricultural production machine, is one of the most important driving force sources of agricultural production. However, some drawbacks come with the manual control, such as, the uncomfortable working condition, the inaccurate routine and the operator fatigue. However, automated driving can improve the field operation standard, the utilization rate of the existing machinery, the mechanical efficiency and labor comfort. Therefore, designing a set of tractor automatic system has a significant meaning in terms of agriculture. At present, the majority of the tractor automated driving systems are designed for specified tractors. When a new type of tractor is used, the developers often need to upgrade their system in terms of software and hardware. And the monotonous work can be reduced by developing a universal system for all types of tractors. Professor Simon Blackmore has been working on a software system for tractors for more than 10 years, and this system can be used as a universal tractor control software system. However, a hardware system is still needed to cooperate with it. In the project, a distributed control system was developed for tractors. Raspberry Pi is used as ECU (electronic control unit) in this system, and TCP/IP (transmission control protocol/internet protocol) is used as the communication protocol in this system; laser distance sensors are used to monitor the environment of the tractor; a Heart Beat system is used to check every important part of this system to make sure the system runs properly. This system receives control commands from a Xbox wireless controller or SAFAR (software architecture for agricultural robot), and it responses to the control commands rapidly and correctly. The John Deere X534 tractor was transformed into automated tractor in the test, R2100 laser scanning distance sensor was selected to safeguard the security of information input system. Since John Deere X534 uses a light tractor CVT automatic transmission, the control strategy is to control the size of the throttle by a servo, and use a motor to control the tractor steering. Experimental results showed that the system was stable and reliable. Due to high precision and high frequency, R2100 laser scanning sensor made the security of the system greatly improved. In the test, when an object was close to the tractor, the tractor could timely stop to protect life and property safety. During the test, in the case of normal operation, disconnecting directly the TCP/IP interface, turning off the router and using other methods to make a fault in the system, the Heart Beat detection mechanism could immediately find the system fault, and Arduino Mega could make the speed set to 0 m/s and the steering angle set to 0 when other tractor control devices failed. The angle of the steering gear was changed in the experiment by changing the width of input signal (square wave signal, the frequency 50 Hz). The test results showed that this system could response to the operator or SAFAR rapidly and correctly, and the response time was about 0.5 s, which was highly acceptable for an off-road vehicle system. The communication between 3 ECUs was stable, dependable and fast. The safety system worked perfectly when any obstacle was found or any system error occurred. The experiment result shows that this system meets the requirement of this project significantly.