An autonomous parking positioning method in parking lot based on vehicle vision and motion information

Abstract

The invention relates to a parking lot autonomous parking positioning method based on vehicle vision and motion information. The method comprises the following steps that 1) an intelligent vehicle located in a underground garage collects image data through a vehicle-mounted binocular camera, the position information of the intelligent vehicle in a local point cloud map is acquired by adopting a binocular vision-based synchronous positioning and mapping method, and primary positioning is completed; 2) the intelligent vehicle continuously detects garage position corner points around an automobile body in the running process, and whether the intelligent vehicle passes the garage position or not is judged; 3) after the garage position is detected, whether the current garage position is a garage position allowing parking is judged; and 4) according to the information of corner point coordinates of the garage position, the speed and the steering wheel angle, expansion Kalman filtering is adopted for fusion to obtain the accurate relative position of the intelligent vehicle relative to the garage position corner points, and secondary positioning is completed. Compared with the prior art,the method has the advantages of being capable of adapting to the GNSS-free signal environment, high in real-time performance and robustness, high in positioning precision, low in cost, high in practicability and the like.

Description

technical field [0001] The invention relates to the technical field of vehicle autonomous parking, in particular to a parking lot autonomous parking positioning method based on vehicle vision and motion information. Background technique [0002] In recent years, technologies related to autonomous driving have developed rapidly and have become a hot spot for investment and scientific research. Autonomous parking technology has become one of the best ways to implement and industrialize autonomous driving technology because it can eliminate cumbersome parking operations and troubles caused by complex parking environments to drivers, attracting investment from many companies and universities. A lot of money for research and development. [0003] However, for the full utilization of space, more and more parking lots are built underground. Since most of the current smart car positioning relies on GNSS, the underground parking lot where the GNSS signal is missing usually must rel...

AI Technical Summary

Problems solved by technology

Existing indoor parking lot positioning solutions mainly rely on wireless devices such as WiFi or UWB. On the one hand, a large number of base station equipment needs to be deployed, which is costly; on the other hand, the positioning accuracy (meter level) is difficult to meet the needs of unmanned driving.

However, the cost of indoor SLAM solutions based on lidar is too high to be equipped in mass-produced models.

In addition, the determination of the position of the self-vehicle relative to the storage location in the process of parking into the storage location generally only depends on dead reckoning, and the cumulative error is large, resulting in a large deviation between the final posture of the vehicle and the ideal posture.

Images