How to plan the path of the sorting robot

The external interaction points of the entire system come from the inbound and outbound requests, and the intelligent sorting robot replaces the work of the previous sorters and pick-up truck drivers. After receiving the task data, the back-end goods database retrieves whether the goods exist and their corresponding positions. The robot is then called to perform the corresponding operation and the corresponding logistics order is given a record and status feedback after the operation is completed.

The dynamic planning of the sorting operation is related to the environmental information. The environmental variables mainly include: the current position and expected time and space information of other picking robots, the state and position of the temporary storage area, and the shape of the warehouse layout. This paper will mention the possible impact of these factors on the picking robot, and will control the changes of these factors. On the basis of confirming the feasibility, the paper mainly studies the unified path planning problem of the picking robot under the single-width channel storage model.


The simulation system can easily simulate the robot movement by using discrete changes in the position of the points in the coordinate system. After all, the movement patterns of all the robots are based on the corresponding motion algorithms. In reality, positioning the robot requires a certain method. Common methods include zigbee signal positioning, GPS positioning, infrared recognition, and acoustic recognition.

For the application scenario of the simulation system in this paper, the application of zigbee or other communication methods for multi-point communication positioning, or combined with some other methods, such as RFID (radio frequency identification technology) is more advantageous, the signal is removed while the signal is removed to reduce noise. interference. GPS positioning is not applicable because of its weak positioning ability for indoor environment, and infrared recognition is not applicable to transparent objects. It can be combined with acoustic recognition to avoid positioning and obstacle avoidance, and can also use ground line and node optical barcodes. A series of ways to synchronize location information.


When the robot enters the storage space, it is bound to encounter conflicts, that is, the two robots are expected to arrive at the same location at the same time, or new cargo picking requirements appear. Here we use a non-anonymous unified scheduling method to deal with such problems, combining the location data of all robots and the time data of task data and environment with the road network status data for multi-dimensional integrated operations.


In view of the application scenarios of logistics picking robots, the wired communication method is very unfavorable for the activities of large-scale mobile robots. We need to use wireless communication. A series of problems such as signal, error correction, and handshake, which are common in wireless communication, can be solved by referring to the TCP/IP protocol commonly used on PCs, thereby completing communication between the upper computer and the lower computer in the system. That is to say, the data connection is first confirmed through a series of data intercommunication, and secondly, the communication and verification mechanism are established to ensure the correct transmission of data.

The use of ordinary communication protocols will generate a large number of polls, which will have high requirements on the performance and network quality of the communication chips in the whole system, and will not be able to complete the requirements well. Therefore, for the transmission of data packets, we can learn from the mechanism of the artificial neural network, that is, the data packet is transmitted only at the transmitting end that establishes a connection with the target terminal. At the same time, a signature mechanism is established to establish a corresponding pool of physical addresses and virtual addresses for the sender and the terminal. Message publishing and subscription-based messaging can usually be done initially using IBM's MQTT protocol.


That is, the problem of goods being passed between sorting robots. The sorting robot can exchange goods to a certain extent to solve the problem of simultaneous empty collision of the sorting robot. However, because the packet transmission will increase the instability in the process, the design difficulty of the robot and the time loss will be not small, so this paper does not use the packet transfer mode for the algorithm model design, but this method is also a solution, so enumerate Come out for comparison.


That is, how much cargo a robot can hold. Usually we can control in a variety of ways, such as the number of pieces, size, etc., mixed warehouses may even need a variety of different sorting robots to adapt to different types of goods, here we simplify the problem to the same type of goods and sorting robots to discuss , set the backpack capacity to a constant b.


That is, the granularity of operation, usually the finer the granularity, the greater the calculation pressure of the upper computer, and the better the optimization effect. Here we adopt the unit time concept, in which one vehicle width is one unit length n, then the time when the vehicle moves n is unit time t. We assume that we can complete the communication of communication instructions every unit time. Temporarily ignore the actual time of parking and cornering.


Different shelf placement methods will also directly affect the efficiency of the picking robot. This article uses the layout shown in the figure below to design. The method has high compatibility with other road network models, and the research results are more scalable. At the same time, the model can maximize the storage space, and is also a model for maximizing the space utilization of logistics vehicles and warehouses.

Reprinted from the network