| With the rapid development of logistics industry in the direction of high speed,automation and high throughput,autonomous mobile robots have become an important means of indoor logistics operations due to their flexibility and high carrying capacity.The autonomy of robots depends heavily on their positioning and navigation capabilities.At present,there are mature positioning and navigation solutions based on electromagnetic guide rails or QR codes,which have problems of high cost and being easily constrained by the environment,so and it is difficult to meet the application requirements of indoor logistics operations.Visual SLAM which perceives rich environmental information whit low cost,can greatly improve the robot's autonomy and environmental adaptability.In addition,multiple information fusion technologies can make up for the shortcomings of single sensors.Therefore,visual SLAM based on multi-information fusion should be an effective means for indoor robot positioning and navigation.However,due to the characteristics of dynamic and changing environment,high operating efficiency requirements and limited costs,there are still many challenges in large-scale applications of indoor logistics operations.Based on the above background,this paper proposes to research the key technology of indoor autonomous mobile robot navigation for logistics operations.Based on understanding the current status and development trends of related technologies at home and abroad,and clarifying the theoretical model of multi-sensor fusion,it focus on the key technologies of visual inertial fusion positioning based on RGB-D and 3D map reconstruction for indoor robot navigation,which improves the robot positioning performance in the indoor complex environment,and sloves the technical problems of high-precision dense map reconstruction.At the same time,we combine multi-sensor fusion SLAM and mechatronics technology to develop an indoor autonomous mobile robot system for logistics operations,and relevant experimental research was carried out to verify the feasibility and effectiveness of the technology developed in this paper.The main research contents and innovations include:In the first chapter,I discussed the important role of autonomous mobile robots in logistics applications and the importance of conducting multi-sensor fusion SLAM technology research,summarized the current research status of related technologies and their development trends,analyzed the challenges and countermeasures of multi-sensor fusion SLAM technology using in logistics operations,which pointed out the direction for subsequent research.At the same time,the research content of each chapter is arranged.In the second chapter,we established the theoretical basis of multi-sensor fusion technology.Based on the analysis of vision and IMU measurement models and their internal and external parameter calibration methods,the establishment of SLAM state estimation theory,IMU pre-integration model,and multi-sensor fusion optimization model were established.The overall technical scheme of multi-information fusion positioning and navigation for indoor autonomous mobile robots was clarified.In the third chapter,we carried out the research of visual inertial fusion positioning technology based on RGB-D sensor.We Introduced RGB-D sensors into the framework of monocular vision and IMU integration,which implements visual feature tracking based on depth information,initialization for different motion conditions,nonlinear optimization based on sliding windows,loop detection and pose map optimization,and completed the technology of indoor global positioning based on the known maps.In the fourth chapter,we carried out the research of 3D map reconstruction technology for indoor robot navigation.Based on multi-sensor fusion obtaining high-precision pose estimation,a single-frame RGB point cloud was generated based on the down-sampled RGB-D images,and an octree map structure was used to implement real-time dense building maps for real-time avoidance of robots.At the same time,the global dense map was constructed offline and optimize by cloud points filtering.Multi-map fusion and update are implemented using relocation.High-precision reconstruction of the global 3D dense map is innovatively implemented to meet the visual navigation application of indoor robots.In the fifth chapter,through the above-mentioned theoretical and technical research,we integrated the multi-sensor fusion SLAM and the mechatronics technology.Based on the completion of the system design and the integration of software and hardware modules,supplemented by path planning and control technology,a set of indoor autonomous mobile robot systems for logistics operations had been developed,and relevant experimental research was carried out to verify the feasibility and effectiveness of the technology developed in this paper in the application of indoor logistics operations. |
PDF Full Text Request