Focus on AGV industry for 18 years
Focus on AGV industry for 18 years
Robot application has been one of the important means for the manufacturing industry to improve production efficiency and increase profits for a long time. In the process of applying robots to the manufacturing industry, the application performance of logistics robots is the focus of the current market. In fact, logistics robot has gone through many stages from research and development to improve, and is fully entering the stage of visual autonomous mobile robot.
The research and development of autonomous mobile robots benefits from the development of sensor and artificial intelligence technology. It has rich environmental perception ability, on-site dynamic path planning ability, flexible obstacle avoidance ability, global positioning ability and so on. It can adapt to more complex work scenes, so as to effectively solve the pain points of logistics industry and manufacturing industry.
The development of logistics robot has experienced four generations of products: magnetic Automatic Guided Vehicle (AGV), two-dimensional code AGV, laser Autonomous Mobile Robot (AMR) and visual AMR.
In 1953, the first AGV was invented to solve the problem of unmanned handling in the field of industrial logistics. The early AGV moved along the guide wire laid on the ground. After more than 40 years of development, AGV has developed electromagnetic induction guidance, magnetic Automatic Guided Vehicle, two-dimensional code guidance and other technologies. AGV belongs to automatic equipment, which needs to perform tasks along the preset track and according to the preset instructions, and can not flexibly respond to field changes. When there is an obstacle on the guide line, it can only stop and wait. During multi machine operation, it is easy to block on the guide line and affect the efficiency. In a large number of scenarios requiring flexible handling, this kind of AGV can not meet the needs of the application end.
With the development of sensor and artificial intelligence technology, people began to introduce more and more sensors and intelligent algorithms to wheeled mobile devices. Its ability of environmental perception and flexible movement is continuously enhanced, and gradually a new generation of AMR, autonomous mobile robot, has been developed. AMR refers to a robot that can intelligently understand the environment and move autonomously in it.
AMR vehicle perceives the field environment through multi-modal sensors (lidar, camera, ultrasonic radar, etc.), and uses intelligent algorithms to analyze the perceived data, so as to form an understanding of the field environment. On this basis, the most effective way and path to perform the task can be independently chosen. AMR generally has rich environmental awareness, site-based dynamic path planning, flexible obstacle avoidance, global positioning, etc.
Although AMR and AGV are both automatic handling equipment, they have essential differences in many important aspects. The biggest difference is autonomy: AGV needs to follow the preset route and complete the task according to the preset instructions. In the process of task execution, it cannot change its behavior according to the changes of the on-site environment while AMR has the ability of environmental perception and independent planning, and can deal with complex on-site environmental changes. Based on the ability of intelligent perception and autonomous mobility, AMR can more flexibly plan routes between various locations in warehouses or factories. In a highly dynamic operating environment, AMR can better cooperate with humans to perform tasks and make the workflow more efficient.
AMR can be divided into laser AMR and visual AMR. Compared with traditional AGV, laser AMR initially has the ability of perception and independent planning, but its perception ability is very weak. For example, it can not distinguish the types of obstacles, so it can not give flexible obstacle avoidance strategies according to different obstacles, and its positioning method based on contour matching can not effectively solve the impact of highly dynamic scene changes.
