This article is the second in a series of driverless technologies. After the last part of the analysis of optical radar (LiDAR) technology, this paper focuses on the unmanned system based on the robot operating system ROS. This article will introduce ROS and its advantages and disadvantages in the unmanned scene, and discuss how to improve the reliability, communication performance and security of the driverless system based on ROS.
Driverless: integration of multiple technologies
Unmanned technology is the integration of multiple technologies. As shown in Figure 1, an unmanned system consists of multiple sensors, including long-range radar, laser radar, short-range radar, camera, ultrasonic, GPS, gyroscope and so on. Each sensor continuously generates data while it is running, and the system has strong real-time processing requirements for the data generated by each sensor. For example, the camera needs to reach a frame rate of 60 FPS, which means that the processing time for each frame is only 16 milliseconds. But when the amount of data increases, allocating system resources becomes a problem. For example, when a large amount of lidar point cloud data enters the system and fills up the CPU resources, it is likely that the camera data cannot be processed in time, causing the driverless system to miss the identification of the traffic light, causing serious consequences.
Figure 1 Example of an unmanned driving system
As shown in Figure 2, the driverless system integrates multiple software modules (including path planning, obstacle avoidance, navigation, traffic signal monitoring, etc.) and multiple hardware modules (including calculations, controls, sensor modules, etc.), how to effectively deploy Software and hardware resources are also a challenge. Specifically, it includes three problems: First, when the data of the software and hardware modules increases, it is inevitable that some modules will abnormally exit during the operation, and even cause the system to crash. How to provide system self-repair ability at this time? Second, because there is a strong connection between modules, how to manage effective communication between modules (communication between key modules, information can not be lost, can not have excessive delay)? Third, how to isolate resources between each functional module? How to allocate calculations and memory resources? How to confirm higher priority execution when resources are insufficient?
Figure 2 Unmanned hardware and software integration
A simple embedded system does not meet the above requirements of an unmanned system. We need a mature, stable, high-performance operating system to manage each module. After detailed investigation, we feel that the robot operating system ROS is more suitable for unmanned scenes. The advantages and disadvantages of ROS and how to improve ROS make it more suitable for driverless systems are described below.
High Voltage Flat Style Resistor
High Voltage Flat Style Resistor refers to SHV RF series which are designed in planner shape. It's widely used in the market for the low TCR & VCR and excellent high-frequency characteristics and high reliability.
Axial Polyester Capacitor,hv divider,High Voltage Flat Resistor,Flat Thick Film Resistor
XIAN STATE IMPORT & EXPORT CORP. , https://www.capacitorhv.com