Compared with traditional warehouses, a distinctive feature of smart warehouses is the application of various intelligent equipment technologies at the system execution level, including automatic guided transport vehicles, unmanned forklifts, shelf shuttles, smart wearable devices, etc. The process links of goods, shelves, storage, picking, collecting and shipping can effectively improve the operating efficiency in the warehouse and reduce the logistics cost.
Technologies in receiving goods
The pallet handling operation is to transport the whole pallet cargo from the transport vehicle to the receiving area for quality inspection and storage. Currently, the automation solution uses a laser or inertial navigation unmanned AGV for pallet handling operations. These unmanned AGVs are usually equipped with reflectors, magnetic nails and other markers or reflectors in a fixed area. The working environment is fixed and the working condition is stable. It is the most mature and perfect warehouse automation solution. Based on the traditional laser navigation AGV, it is dedicated to the development of SLAM (simultaneous positioning and map creation) technology to realize the natural navigation of unmanned AGV. It does not need to install markers or reflectors, just need to install the environment-aware sensor. The human AGV starts from a certain location in an unknown environment, and performs self-positioning according to the sensing information acquired by the internal and external sensors during the movement process, and gradually establishes a continuous environment map, and then can realize the unmanned forklift on the basis of the map. Accurate positioning and path planning to complete navigation tasks. The natural navigation unmanned AGV has the characteristics of short installation time, low input cost, and free creation of new paths. It is an excellent research and development direction for the next generation of intelligent unmanned AGV.
(2) Intelligent demolition robot
The unpacking operation is to carry a box of goods placed on the transfer tray to the conveyor line. The traditional automation solution is completed by the industrial robot arm grabbing or sucking. Because the industrial robot arm operation control is based on the box size and palletizing rules stored in the computer system database, and the on-site job object is not recognized online, it can only be realized. The removal of the same size of the box from the same tray, the maintenance of the database is a very arduous task when faced with thousands of types of cargo boxes. The size of the cargo boxes on the same pallet received by the e-commerce company is different, and there is no fixed rule in the pallet. The traditional industrial robot arm is difficult to operate. The intelligent demolition robot uses 3D vision and deep learning algorithms to realize self-training and self-correction of industrial robot arm operations, without the need for box and 的 type database maintenance. The industrial robot recognizes the top cargo contour through a 3D depth camera. When picking up a box for the first time, it builds a model of the shape of the box and accelerates the identification of the next box based on this model.
Technologies in Storage
In the traditional centralized operation process, goods of the same product are imported into and out of the warehouse in large quantities in trays. The most common automation solution for the storage chain is the Autonomous Vehicles Storage and Retrieval System (AVS/RS). The unit is stored on the high shelf, and the pallet loading and unloading operation is completed by the stacker. In the process of fragmentation, in order to facilitate the selection of thousands of order items in the late mass items and thousands of order lines, the goods are stored in the bin mode, and the intelligent dispatching algorithm is used to command the car group to complete the goods entering and leaving the warehouse.
(1) KIVA robot system
The KIVA robotic system consists of hundreds of machine carts that lift the shelving unit. After the goods are unpacked, they are placed on the shelf unit, and the goods are bound to the information of the shelf unit through the bar code at the bottom of the shelf unit. The bar code grid is arranged on the warehouse floor, and the two cameras of the machine trolley respectively read the bar code of the ground bar code and the bottom of the shelf unit. Cargo handling navigation is done with sensors such as encoders, accelerometers and gyroscopes. In addition, the machine trolley does not support movement and steering synchronization, and it needs to be fixed in place when steering. The core of the system is a centralized multi-agent scheduling algorithm that controls the cart.
(2) Automatic shuttle warehouse system
The KIVA robot system is limited by the height of the shelf unit and can only store the goods in the plane space. The automatic shuttle warehouse system uses the three-dimensional bin-type shelf to realize the storage of the goods in the three-dimensional space in the warehouse. Before entering the warehouse, the goods are stored in the bin after being unpacked, and the bin is sent to a certain layer through the hoist at the front end of the racking lane, and then the shuttle in the layer stores the goods into the specified cargo. When the goods are out of the warehouse, the cooperation between the shuttle and the hoist is completed. The core of the system is to optimize the algorithm of the location allocation optimization algorithm and the trolley scheduling algorithm, balance the workload between each lane and each layer in the single lane, improve the parallel working time between equipments, and maximize the working efficiency of the equipment.
(3) Cell unit system
The automatic guided trolley in the KIVA robot system realizes ground handling, and the shuttle in the automatic shuttle warehouse system realizes the handling on the shelf rail. The new cell unit system is the fusion of the above two technologies.
When the cell unit trolley is on the shelf or the hoist, it moves on the track according to the working mode of the traditional automatic shuttle; when leaving the shelf to reach the ground, it can switch to the operation mode of the automatic guided trolley to run on the ground. The navigation method on the ground is different from the KIVA robot system. It uses sensor fusion technology based on wireless sensor network ranging, laser range finder measurement and speculative navigation method. The wireless sensor network realizes information communication and global positioning. The range finder measurement and speculative navigation method achieve position tracking and positioning accuracy correction, which is more flexible than the KIVA robot system ground tag with inertial navigation. The cell unit system seamlessly links the three-dimensional shelf storage space with the ground plane storage space, representing the future development direction of the scalable and highly flexible car group technology.
Technologies in Picking
(1) AR assisted picking technology
Traditional manual picking solutions use handheld RF (Radio Frequency) picking, Pick to Light or Voice Directed Picking. The picker picks up the goods in the shelf based on the lights on the shelf or the hand-held RF and the prompts in the wearable device. Although the traditional manual picking method improves the accuracy of the operation, it requires the picking personnel to be familiar with the layout of the warehouse. Augmented Reality (AR) integrates the information of the real world and the virtual world “seamlessly”, automatically identifies the warehouse environment through AR glasses, locates the position of the goods to be sorted, and automatically plans the picking path to establish route navigation. Guide the operator to reach the target picking position in the shortest time, automatically scan the barcode of the goods through the AR glasses, guide the operator to accurately obtain the goods, and liberate the hands, which can greatly improve the efficiency of the picking operation.
(2) Array type automatic picking technology
The traditional automatic sorting equipment is a channel type picker represented by the A-frame system. The same item of goods is neatly stacked in the vertical channel, and the cargo is sorted onto the conveyor line by the ejection mechanism at the bottom of the channel. Parallel arrangement along the conveyor line, single cargo item sorting area is large, and equipment costs are high, mainly suitable for distribution centers with large selection and concentration in limited items, when faced with large selection and involving many items Order processing tasks are often not available due to space layout and equipment cost constraints. The array automatic picker was designed and developed and successfully applied. The Matrix Automated Order Picking System (MAOPS) is a new type of automated zero-removal picking system that is arranged in a space by a large number of horizontal tilting picking channels. All picking channels are installed on the equipment at a certain angle of inclination. The bottom of the channel is filled with fluent strips. The goods are placed on the fluent strips and slide down to the front of the channel under the influence of gravity. An ejecting mechanism is installed at the front end of each sorting channel. Each time the ejecting mechanism moves, the single piece of goods at the forefront of the channel is sorted out, and the sorted goods are slid along the baffle to the conveying line, and the remaining goods in the channel are subjected to gravity. It is continuously added to the ejecting mechanism to ensure the continuity of the picking of goods. The width of the channel can be adjusted within a certain range to suit different cargo sizes, but only the goods of the same item are placed in each channel, and the goods with a large selection amount can be stored in multiple picking channels at the same time. The number of picking channels corresponding to the same length of conveying line is about 5~7 times that of the A-frame system, and the equipment cost is only 1/10~1/5 of the A-frame system.
(3) Delta robot picking technology
The array picking technology is mainly suitable for the packaging of standard packaging products, which cannot meet the picking requirements of other packaging types such as bags and bottles. Professor Reymond Clavel of Sweden proposed the Delta parallel robot in the 1980s. The drive motor of the manipulator is designed on the frame. The slave arm can be made into a light rod, so the end can obtain high speed and acceleration, especially suitable for light goods. High-speed sorting operation type. Based on cameras and computers to simulate human visual functions, Delta robots can perform dynamic picking, and the robot can change the picker according to the different sizes and types of products, so the applicable packaging types can be varied. In addition, in order to ensure the accuracy of the capture, the Delta robot needs to use artificial intelligence technology to train the recognition accuracy of the same product in different postures.
Technologies in Collection and Delivery
In the delivery area, different blocks are divided according to the delivery line. The order sorting operation is to place the order completed in the corresponding delivery line. Traditional automation solutions use automated sorting lines based on skewer splitter, slider sorter or cross-belt machine. The sorting line can only solve the order according to the delivery line classification, but can not fulfill the order according to the delivery line. In the fixed order, only when the shipment is shipped, the delivery personnel will pick out the corresponding order items in the order of the far and near customers according to the delivery line, and then load them into the compartment, which seriously affects the loading efficiency.
The intelligent delivery sorting system adopts the automatic shuttle technology. The order of the picking is stored in the three-dimensional shelf. The access of the shuttle can be adjusted according to the size of the box, so it can be applied to different sizes of goods and different. Type of packaging. When receiving the delivery loading instruction, the order goods will be taken out from the shelf in the order of the far and near customers according to the delivery line, and sent to the loading area through the conveying line. If the telescopic belt machine is used, the direct loading can be realized. The intermediate secondary handling link is reduced, which can greatly improve the loading efficiency. In addition, since the order is stored in the shipping area shelf, the space utilization rate is significantly improved compared to the ground plane storage mode of the traditional shipping area.
In summary, the development of intelligent warehouse equipment technology has two significant features: one is the rapid integration of robots based on data mining, artificial intelligence algorithms and automatic perceptual recognition technology; the second is friendly and efficient collaboration between human and machine. For picking and picking operations such as demolition and picking, robots equipped with 3D machine vision and artificial intelligence algorithms can gradually deal with all kinds of weights of goods; for handling operations such as unloading, warehousing, delivery, and collection, all kinds of equipment are available. The navigation device and the automatic guided trolley robot of the scheduling algorithm can be completed in cooperation. The integration of robots makes the traditional warehouse logistics system from rigid to flexible, while artificial intelligence makes logistics operations more efficient and precise. At the same time, due to the diversity of products and the high construction cost, all unmanned warehouses using robotic operations will only appear in individual industries or enterprises, and have no general applicability. Therefore, they will be adopted in more application scenarios. Human-machine collaboration mode. For example, in the order picking process, the KIVA robot system or the automatic shuttle warehouse system sends the selected cargo bins to the picking station, and the picking operation is completed manually. The improvement of the manual picking efficiency depends on the design of the picking station, and the number of people needs to be increased. The friendliness of the machine interface makes people work more pleasant and comfortable, and makes the work station more human and intelligent, which can effectively avoid mistakes.