AGV forklift truck autonomous positioning mobile SLAM navigation system

Abstract

An AGV forklift autonomous positioning mobile SLAM navigation system comprises instant positioning and map construction, a three-thread structure is adopted, a tracking thread, a local optimization thread and a global closed-loop detection and optimization thread are synchronously carried out, and an AGV forklift adopts a laser + SLAM navigation system of a reflector-free laser AGV forklift. An artificial beacon of a pinhole camera model and a robust beacon recognition algorithm are adopted for instant positioning and map construction, and the artificial beacon adopts a visual SLAM algorithm architecture. A binocular odometer based on ORB feature points is designed at an algorithm front end, the influence of illumination change on the system is reduced, and the positioning error of the system is reduced by combining a RansacPnP algorithm and a local optimization algorithm in the pose tracking algorithm. An algorithm rear end establishes global plane optimization for the motion model of the AGV and the actual working scene, and the artificial beacon closed-loop detection optimizes the global pose graph, thereby improving the positioning precision and operation efficiency of the SLAM algorithm.

Description

technical field [0001] The present invention relates to a forklift navigation system, in particular to an AGV forklift autonomous positioning mobile SLAM navigation system using three-thread structure, tracking thread, local optimization thread, global closed-loop detection and optimization thread synchronously. Background technique [0002] SLAM, the English full name is simultaneous localization and mapping, real-time positioning and map construction, and the sensors used in SLAM are mainly divided into two categories: lidar and camera. Accordingly, the industry also divides SLAM into two categories: laser SLAM and visual SLAM. SLAM was first proposed by Smith, Self and Cheeseman in 1988 and has a history of 30 years. The existing forklift navigation system generally adopts ORB-SLAM feature method or DSO direct method. However, the ORB-SLAM feature method has the following disadvantages: the built map point cloud is sparse, and in terms of running speed, because there is...

AI Technical Summary

Problems solved by technology

However, the ORB-SLAM feature method has the following disadvantages: the built map point cloud is sparse, and in terms of running speed, because there is a bottleneck in the time to extract feature points, the fastest running speed should not exceed 30 frames / s, measured on this machine (i7-6600U) The speed is basically around 20frames / s, so for a camera with a high frame rate, the frame rate needs to be reduced to use

The DSO direct method also has the following disadvantages: it has high requirements for scene lighting, and it is required to keep the exposure time as stable as possible.

The code scalability is relatively poor. At present, only the monocular version is open source. This version does not do scale restoration, so it cannot be directly used in practice.

Images