A vehicle environment perception and control method based on cloud machine learning

Abstract

The invention relates to a vehicle environment perception and control method based on cloud machine learning, which includes: A1. Collecting image data of the front, rear and left and right rear blind areas of the vehicle, and millimeter-wave radar data of obstacles in the front area of ​​the vehicle; A2. Extract the feature vector set of the image data; process the obstacle millimeter wave radar data to obtain the distance of the obstacle in front of the vehicle; A3. Perform machine learning on each feature vector set to identify the front, rear and both sides of the obstacle; A4. Carry out target tracking and obtain the target's environmental perception data and vehicle current status data; A5. Calculate the safe distance and perform active vehicle safety control accordingly; A6. Upload the target's environmental perception data to the cloud service system for learning, and finally update the target recognition Parameter sets and algorithms control parameter sets. The invention effectively improves the accuracy of environmental perception, vehicle control accuracy and ranging distance, and overall improves the performance of the vehicle active safety system.

Description

technical field [0001] The invention relates to the technical field of vehicle active safety, in particular to a vehicle environment perception and control method using cloud big data learning. Background technique [0002] With the development of the automobile industry, more and more attention has been paid to the safety of vehicles. Due to the existence of many types of safety hazards such as high speed, side view blind spots, and rear-end collisions, collision accidents often occur. Therefore, in order to avoid vehicle collision accidents as much as possible, vehicle environment awareness and active safety have become urgent problems to be solved. [0003] The purpose of the vehicle environment perception and control system is to perceive the surrounding environment data of the vehicle through various sensors such as visual sensors, radar sensors, ultrasonic sensors, etc., such as: static and dynamic obstacles in front of the vehicle, moving objects in the blind spots on...

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to propose a vehicle environment perception and control method based on cloud machine learning to solve the problems caused by the short vehicle environment perception detection distance and the inability to update the environment perception parameters and control parameters in the existing automobile active safety technology. Insufficient environmental perception accuracy and vehicle control precision

Images