3D MAP Construction Of Mobile Robot Based On Multi-sensor Fusion

With the rapid development of artificial intelligence and computer technology,intelligent mobile robots are more and more widely used in the fields of mass service,unknown environment exploration,and technology power.Therefore,researching the autonomous positioning and navigation of mobile robots has always been a hotspot and a difficult point in the field of mobile robots.The three-dimensional map construction of mobile robots in the process of autonomous positioning and navigation of mobile robots is one of the important basic problems.Effectively solving the problem of three-dimensional map construction is considered to be one of the key technologies for realizing intelligent mobile robots.In recent years,with the continuous development of multi-sensor fusion technology,the realization of three-dimensional map construction of mobile robot based on multi-sensor fusion technology has gradually become an important research direction.This paper first introduces the research status of multi-sensor fusion technology and 3D map construction of mobile robots,and analyzes the advantages and disadvantages of current research results.Then based on the current research results,a three-dimensional map construction scheme of multi-sensor fusion mobile robot is proposed.The scheme is constructed by a mobile robot equipped with a visual-inertial measurement device combined with sensor calibration,visual tracking,key frame extraction,closed-loop detection and map optimization algorithms to realize a structured indoor three-dimensional map construction.Finally,the shortcomings and difficulties in the current 3D map construction are proposed.In the visual tracking module,a ring matching tracking algorithm combining LK optical flow and feature points is proposed,which can quickly realize system initialization while ensuring tracking quality.Secondly,the key frame extraction module combines the disparity constraint and the feature matching quantity constraint with the interval frame number constraint to obtain a better key frame extraction effect.Then,the DBoW2-based word bag model is used to realize closed-loop detection in the map construction process.And then,the optimization of the map is realized by an optimization algorithm of inertial-visual sensor fusion.Finally,the experimental platform is built to verify the multi-sensor fusion mobile robot 3D map construction algorithm.In this paper,the algorithm of this paper is tested and compared using the standard test data set provided by EuRoC,and the TurtleBot2 robot platform is equipped with visual-inertial sensor for algorithm testing in real environment.The experimental results show that the multisensor fusion mobile robot 3D map construction algorithm is more accurate and robust than the visual SLAM,which proves the effectiveness of the proposed algorithm.