AGV导航——二维码导航总体设计：该系统由以陀螺仪导航系统、视觉系统、AGV子系统、电源管理系统、传感器系统和装置机械结构五部分组成。导航采用陀螺仪导航为主，视觉导航为辅，最大化融合和利用各导航的优势，提高系统的可靠性和导航精度。
二维码导航运行原理：AGV在接收到工作中心的指令后，由导航系统将其指引至货物装载处，装载完毕后，按照预设指令，其分析起点-终点路径后，规划出最佳行走路径，行走至指定位置。该过程中不断利用导航系统识别周围特征标志信息，以实时利用AGV子系统计算分析其所处位置，之后利用无线通信方式发送至工作中心电脑，以管理和规划工业现场的总体物流运行进度，避免相互干涉，提高运输效率。
二维码导航项目技术归纳：
（1）
陀螺仪导航与视觉联合导航：本系统采用陀螺仪导航系统专用模块，主要实现技术为差分定位，并结合工业现场的地图，利用车载控制系统实时分析系统地图坐标数据，之后与地图信息对比以获取定位信息。 项目采用图QR码扫描自适应阈值算法的视觉技术识别运动过程中的关键标志物，辅以航位推算系统以达到路径自动辨识和规划，从而最终达到对AGV导航的目的。通过视觉定位QR码技术导航的图像获取、摄像机标定、特征提取和深度恢复等过程，以达到对物体的位置精确定位。
AGV QR Code Navigation
QR码（二维码）
（2）路径规划：AGV运行路径规划分为全局规划和局部规划。全局规划中采用切线图法，即将路径中关键点作为特征点，将该特征点的切线表示弧，这样可以获取AGV起始点和目标点的最短路径，提高AGV路径进行规划的速度；局部规划中采用人工势场法，其设计思想是将AGV在工业现场作业视为一种抽象人造受力场中的运动，通过建立人工势场的负梯度方向指向系统的运动控制方向，目标点对AGV产生引力，障碍物对AGV产生斥力，其驱动结果使其在势场合力作用下控制AGV运动方向并计算AGV位置，为防止工业现场AGV在到达目标位置前陷入局部小点而无法达到预设位置，系统利用模拟退火算法使势函数跳出局部极小点，以使AGV顺利到达目标位置。
（3）多任务分解及协调：为解决多个AGV间任务分配、路径规划和相互协调，系统采用模糊动态数学模型的方法，该方法基于专家辨识系统的设计思路，将任务分配分解为“最重要、重要、一般、次要”四个等级，并将路径规划为“最近、较近、合理、备选”四个等级，之后利用模糊动态数学模型进行建模和分析，输出最佳的任务分解和路径规划。具体应用中，利用工业现场工作中心对多个AGV提前预置任务和目标路径，提供给系统的初始输入和输出，由系统自动完成对任务和路径的分析，并将指令传送至各AGV车载控制系统，以达到AGV间的任务协调和路径选取。需要指出的是，为了解决实际应用过程中由于任务的不断更新和调整，以及路径过程中可能出现障碍物等情况，在工作中心中设有一套预测控制系统和报警系统，在满足单一作业完成的前提下，对工作中心的任务和路径进行实时建模，以满足现场变化的需求，当路径出现死角时，利用报警系统进行现场提示，由现场人员进行介入解决。
（4）基于视觉的QR码定位技术:对于二维码子系统，其中存储该二维码正对的地面的坐标信息。AGV通过自身的高分辨率长焦摄像头对二维码进行读取和识别，同时通过二维码图像在摄像头坐标中的旋转情况，与电子罗盘数据进行融合，可以确定AGV的精确朝向。综合二维码编码的信息和其在图像中的位置信息，可以对AGV小车进行完整的定位。
AGV Navigation - QR Code Navigation Overall Design: The system consists of five parts: gyroscope navigation system, vision system, AGV subsystem, power management system, sensor system and device mechanical structure. Navigation mainly uses gyroscope navigation, supplemented by visual navigation, maximizes the integration and utilization of the advantages of each navigation, and improves the reliability and navigation accuracy of the system.
QR code navigation operation principle: After the AGV receives the instruction from the work center, the navigation system guides it to the cargo loading place. After loading, according to the preset instruction, it analyzes the starting point-destination path and plans the best walking. path, walk to the specified location. In this process, the navigation system is continuously used to identify the surrounding feature sign information, and the AGV subsystem is used to calculate and analyze its location in real time, and then it is sent to the work center computer by wireless communication to manage and plan the overall logistics operation progress of the industrial site to avoid Interfere with each other to improve transportation efficiency.
Technical summary of QR code navigation project:
1. Gyroscopic Navigation and Vision Joint Navigation: This system adopts the special module of gyroscope navigation system, the main realization technology is differential positioning, combined with the map of the industrial site, uses the vehicle control system to analyze the system map coordinate data in real time, and then compares it with the map information to obtain location information. The project uses the visual technology of the QR code scanning adaptive threshold algorithm to identify the key markers in the movement process, supplemented by the dead reckoning system to achieve automatic path identification and planning, so as to finally achieve the purpose of AGV navigation. The process of image acquisition, camera calibration, feature extraction and depth recovery through visual positioning QR code technology navigation, in order to achieve accurate positioning of the object's position.
2. Path planning: AGV operation path planning is divided into global planning and local planning. The tangent diagram method is used in the global planning, that is, the key points in the path are used as feature points, and the tangent of the feature point represents the arc, so that the shortest path between the starting point and the target point of the AGV can be obtained, and the speed of the AGV path planning can be improved; in local planning Using the artificial potential field method, the design idea is to regard the AGV's operation in the industrial field as a motion in an abstract artificial force field. By establishing the negative gradient direction of the artificial potential field, it points to the motion control direction of the system, and the target point generates Gravity, obstacles produce repulsive force on AGV, and the driving result makes it control the direction of AGV movement and calculate the position of AGV under the action of potential field force, in order to prevent the AGV from falling into a local small point before reaching the target position in the industrial field and unable to reach the preset position, The system uses the simulated annealing algorithm to make the potential function jump out of the local minimum point, so that the AGV can reach the target position smoothly.
3. Multi-task decomposition and coordination: In order to solve the task allocation, path planning and mutual coordination among multiple AGVs, the system adopts the method of fuzzy dynamic mathematical model. There are four levels of "important, important, general, and minor", and the path is planned into four levels of "nearest, nearer, reasonable, and alternative", and then the fuzzy dynamic mathematical model is used for modeling and analysis, and the best task is output. Decomposition and path planning. In specific applications, the industrial field work center is used to preset tasks and target paths for multiple AGVs in advance, and provide initial input and output to the system. The system automatically completes the analysis of tasks and paths, and transmits the instructions to each AGV vehicle-mounted control. system to achieve task coordination and path selection between AGVs. It should be pointed out that, in order to solve the situation of continuous update and adjustment of tasks and possible obstacles in the process of the actual application process, there is a set of predictive control system and alarm system in the work center. Under the premise, the tasks and paths of the work center are modeled in real time to meet the needs of on-site changes. When there is a dead angle on the path, the alarm system is used to prompt on-site, and the on-site personnel will intervene to solve the problem.
4. Vision-based QR code positioning technology: For the two-dimensional code subsystem, the coordinate information of the ground facing the two-dimensional code is stored. The AGV reads and recognizes the QR code through its own high-resolution telephoto camera. At the same time, the rotation of the QR code image in the camera coordinates is combined with the electronic compass data to determine the precise orientation of the AGV. Combining the information encoded by the QR code and its position information in the image, the AGV car can be completely positioned.
021-67708588
