Automatic driving systems require urban rail transit systems to have a high level of interconnectivity, safety, speed, and comfort. Since the 1990s, with the development of communication, control, and network technologies, high-capacity, two-way information transmission between trains and subways has become possible, making it possible for high-density, high-capacity subway systems to become truly automatic driving systems.
Principles of autonomous driving systems
1. Driver assistance system: It mainly provides the driver with important information and a favorable driving mode and environment, and can provide clear and concise warnings when danger occurs, such as lane departure warnings.
2. Partially automated systems: These systems allow for human intervention in dangerous situations, such as automatic emergency braking and emergency lane assist.
3. Highly automated system: The vehicle can be automatically controlled by a computer for a long or short period of time, but the driver can perform corresponding monitoring activities.
4. Fully automated system: It can perform corresponding driverless operation, and other people in the vehicle can carry out other activities during this period.
Autonomous driving system architecture
1. LiDAR: It can scan within a radius of 60 meters and display the results in 3D on the vehicle's computer.
2. Front-facing camera: Primarily used for traffic light recognition, with assistance from a computer.
3. Left rear wheel sensor: Computer positioning is performed during lateral movement to determine the correct position on the road.
4. Front and rear radars: A total of 4 such devices are installed on the vehicle, 3 in the front and 1 in the rear, mainly to measure obstacles on the road during driving.
5. Main control computer: This is the brain that receives and issues instructions.
An autonomous driving system for automobiles mainly consists of two parts: hardware and software. The hardware part comprises a binocular camera system, an image processing platform, a software library, a control interface, and actuators. The software part mainly includes target recognition, motion parameter measurement, and image processing. The image processing platform uses image processing software to perform preliminary recognition of the left and right images based on a standard road model in a knowledge base, segmenting the road area for the vehicle to travel on.
The autonomous driving system uses a binocular camera system to automatically collect road condition information within 1 to 200 meters in front of the vehicle, including ground markings, median strips, guardrails, lane markings, road obstacles, number of motor vehicles, lane distribution, distance, relative speed, and road sign information.
Based on this information, the system software automatically decides on vehicle speed, direction and position, automatic overtaking, emergency braking, horn activation, and lighting control. The autonomous driving system uses a closed-loop system for steering and speed control. It can be used for autonomous driving on highways to avoid traffic accidents, and it is simple in structure, low in cost, and highly reliable.
