AGV model characteristics and Selection Guide
Summarized in terms of AI Jiwei current AGV product library is divided into two categories, one is no reflector laser AGV forklift, the other is the car AGV magnetic navigation. The two AGV product categories, the biggest difference is that used in navigation in different ways, AGV car using the magnetic navigation mode, belong to marked navigation technology without reflector laser forklift AGV is used without reflector laser navigation technology AI Jiwei original, no sign belongs to the navigation technology.
The main components of the laser forklift without reflector AGV are the forklift body, the laser navigation module without the reflector board, the safety module and the vehicle control module. Edgeway's forklift truck AGV does not need to build any auxiliary signs including reflective plates, magnetic strips, magnetic nails, two-dimensional codes and so on. Combined with vehicle borne laser scanner, we scan the current AGV operation environment of forklift truck, and determine the current location of forklift AGV based on SLAM algorithm.
Without reflector, laser self guided forklift truck AGV can be divided into stacked forklift truck, mobile forklift truck and counterweight laser forklift according to the structural characteristics of forklift truck. Each type of forklift is based on the different parameters of the load and height, and the different types of forklift bodies in the same category are used in the same category. If the material height is generally used forklift stacking AGV; the material is heavy, basically do not need to improve handling forklift AGV; the difference between the balance heavy forklift AGV and the two AGV forklift is fork lower without legs, when the lower part of non space such as swastika tray, use forklift trucks AGV car.
The AGV trolley includes a tractor, a lurking car and a backcar. The tractor style is relatively simple, so there is no more segmentation, main feature is when car chassis height is lower than 300mm, AGV car unable to drill into the car below, can be installed in the car in front of the traction mechanism, traction type AGV traction mechanism to drive the car through the car running, to complete the specified action potential models; the most widely used, in accordance with the number of drivers the walking direction and then divided into single driving type, single driving bidirectional unidirectional latent latent latent type bidirectional and dual drive. This type of application features is when the car chassis is higher than the height of 300mm, installation of the traction mechanism in car bottom type AGV drill to skip car lurking below the traction mechanism through the body lifting rod is connected with the drive, skip work. The car needs to run the process, the space size, material load comprehensive judgment, using single drive single or double drive two-way one-way, two-way drive this car, the space is saved; the car with the most complex, because the burden of car load mechanism can according to the customer the actual material the myriads of changes, with appropriate state of transplanting mechanism. More typical back vehicles include drum type, chain type, lift type peace platform type. Another feature of Edgeway is a heavy load model, and a few tons of material can be carried automatically through a heavy load back AGV.
Whether it is no reflector laser or forklift AGV AGV car, the basis for selection should consider the factors of customer site or skip tray type and size, space size, aisle width, material load, work flow, docking level etc.. The selection of AGV need to fully consider the application environment, AI Jiwei project engineer will generally do models recommended after a comprehensive assessment of the application environment.
1. what is SLAM?
Slam (Simultaneous Localization And Mapping, simultaneous localization and simultaneous mapping), that is, use sensors to locate the robot (such as AGV) and to map the surrounding environment. When the robot comes to a strange environment, it needs to quickly establish the relationship between itself and the environment. "Where am I now?" "Where do I go?" "Where do I come from?" "What do I have to do now?" "What is the world I see now" "can I locate my position in the existing abstract world?" In a series of problems, based on these questions, the robot can make a perfect answer, that is, the SLAM method needs to be solved.
4 elements of the implementation of 2.SLAM method
When the SLAM method is implemented, the following 4 aspects should be considered:
The map represents a problem, which needs to be chosen according to the actual scene needs
The problem of information perception needs to consider how to comprehensively perceive the external environment. In general, SLAM uses ambient sensing sensors to be divided into visual navigation sensors, such as cameras and laser navigation sensors, such as laser scanner.
The data types, timestamps and coordinates of different sensors are different and need to be handled in a unified way.
The problem of location and composition is how to achieve pose estimation and modeling, which involves many mathematical problems, including physical modeling, state estimation and optimization.
SLAM can be divided into two categories: laser SLAM (also 2D and 3D) and visual SLAM.
About SLAM 2D laser is SLAM, positioning, only use a single laser sensor, two-dimensional positioning in a plane scanning laser sensor on the two-dimensional positioning precision, in the acquisition, based on the solution of 3D laser point cloud, to become a complete three-dimensional data.
Similarly, laser SLAM 3D uses three-dimensional laser sensor to get 3D data, then localize it through feature points matching of 3D data, then calculate and match the complete 3D data on the basis of 3D positioning, and finally get its pose.
The visual SLAM system is divided into four modules: the camera, the back end, the building diagram and the loop detection. The process is visual SLAM, which relies on a camera to perceive the world around and evaluate its pose. It needs to use visual odometer to pass two adjacent photos.