Shulou(Shulou.com)06/02 Report--

The iterative research and development of intelligent driving technology needs the support of a variety of sensors, massive data and massive scenes. At present, synchronous data acquisition of multiple sensors at Gbit/s level, rapid analysis and evaluation of massive data, slicing and extraction of key scenes are recognized as thorny problems in the industry.

In order to solve the above thorny problems, Jingwei Hengrun launched the intelligent driving test system-VDAS. VDAS is mainly composed of data acquisition equipment, environment awareness system and data analysis software VDA.

Data acquisition equipment

The data acquisition equipment is composed of vehicle industrial control computer, a variety of data acquisition boards, sensors, vehicle display screen and so on. Through the supporting data acquisition software, to achieve all kinds of vehicle bus data, perceptual data acquisition, adapt to CAN/CANFD, Ethernet, vehicle Ethernet, USB, serial port and other common interfaces.

Perceptual system

Through lidar, high-definition webcam, integrated navigation sensor, combined with advanced environmental perception algorithm and sensor fusion algorithm, we can accurately sense the information of the surrounding environment, including the category, distance, speed, acceleration, heading angle of the surrounding traffic participants, as well as the distance, type and color of the surrounding traffic signs. For lane detection, we can also detect whether to cross the line. And count the number of pressing lines, and so on.

The following dynamic picture shows the perception effect of lidar output through multi-target tracking, machine learning classification and other algorithms in the urban expressway scene.

Lidar perception

In order to reduce the detection uncertainty of a single sensor, the system adopts the information fusion scheme of laser radar + camera, combined with the characteristics of high precision of lidar and rich camera information, the final fusion effect is shown in the following figure.

Lidar + camera sensor fusion

Through camera calibration and image feature extraction, the system can identify the lane information, including the lane type, the distance from the lane, the fitted lane equation and so on.

Lane recognition effect

VDA data analysis

VDA is a real car test data analysis software independently developed by Hengrun. With the data acquisition system, it can support automatic extraction of key indicators, data statistics and automatic generation of test reports for all kinds of site testing and open road testing. Compared with the traditional manual analysis, it can greatly improve the efficiency of data processing.

VDA Target Information Analysis Interface

Regulatory evaluation package and road test evaluation package supported by VDA

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.