Accurate positioning method of driverless cars in underground parking

Abstract

The invention discloses an accurate positioning method of driverless cars in underground parking. The accurate positioning method of driverless cars in underground parking comprises following steps: laser point cloud ground segmentation module processing, cylinder characteristic extraction and characteristic coupling, non-ground effective point selection module processing, laser effective point merging and downsampling module processing, grid map generation and three dimensional map coupling module processing, and initial pose module processing. According to the accurate positioning method, cylinder characteristics are adopted to reject dynamic obstacles in laser point cloud scene, and realize automobile accurate positioning using map coupling technology. The sensitivity of the accurate positioning method on initial pose is much lower than that of conventional method, relatively positioning accuracy is achieved, the accurate positioning method is suitable for positioning and navigationof driverless cars in underground parking, and is especially suitable for positioning when there are a large amount of dynamic obstacles in underground parking.

Description

technical field [0001] The invention belongs to the field of unmanned vehicle research, in particular to a precise positioning method for an unmanned vehicle in an underground garage. Background technique [0002] In recent years, with the rise of a new generation of artificial intelligence technology, unmanned vehicles have emerged as an important algorithm verification platform for artificial intelligence technology, and have shown a blowout development, especially a large number of start-up companies have emerged in China. Directly promote the rapid development of this research field. An unmanned vehicle is a special wheeled mobile robot. It relies on various sensors carried by itself to sense the road environment, automatically obtain vehicle positioning information, plan driving routes, and control vehicle movement, so as to realize autonomous driving of the vehicle. Features. [0003] Positioning technology plays an important role in the field of unmanned vehicle res...

AI Technical Summary

Problems solved by technology

In view of the high precision and high reliability of RTK-GPS, most unmanned vehicles choose to use GPS for positioning. However, there are many tunnels, bridges, tall trees and other obstacles in the current urban roads. GPS signals are easily interfered and blocked in this environment, and the vehicle positioning accuracy is seriously insufficient. In order to solve this problem, people propose a combined positioning method using GPS+IMU, but its disadvantage is that GPS signals are affected by long-term tunnels, etc. In an obstructed environment, it is easy to fail, or the vehicle starts in an environment where the GPS signal is blocked, and the vehicle pose cannot be determined

[0004] For this reason, the researchers proposed to use various sensors carried by the vehicle itself for perceptual positioning. This method first establishes a priori map, and then uses the vehicle motion model and inertial navigation devices for positioning. This method is in the environment without GPS signals. The effect is obvious, but the only disadvantage is that this method is easily affected by noise or dynamic obstacles in the perceived environment data, thus affecting the positioning accuracy

Images