Dynamic handling task coordination method based on real-time AGV warehouse robot model

Abstract

The present invention relates to a handling task coordination method based on a dynamic real-time AGV warehouse robot model, and belongs to the field of AGV warehouse robot control technology. The present invention constructs an AGV warehouse robot tracking model, and finally tracks the AGV warehouse robot of the target warehouse through the AGV warehouse robot tracking model, performs real-time collision detection on the dynamic model of the AGV warehouse robot, and generates the final handling task coordination planning result according to the collision detection result, and performs planning according to the final handling task coordination planning result. The present invention performs three-dimensional model visualization management of the AGV warehouse robot during the execution of the task, so that the manager can observe the actual working conditions of the AGV warehouse robot in real time, and through the dynamic analysis of the three-dimensional model, it can more accurately evaluate the execution of the AGV warehouse robot during the task, and improve the coordination of the handling task.

Description

Technical Field

[0001] The present invention relates to the technical field of AGV warehouse robot control, and in particular to a handling task coordination method based on a dynamic real-time AGV warehouse robot model. Background Art

[0002] Manufacturing is the driving force for the development of various industries. According to statistics, a lot of time and manpower are consumed in the loading and unloading, transportation and storage links during product production and processing, especially for large unmanned warehouses, ports, large manufacturing enterprises and other large scenes. The handling and dispatching of items requires more time and economic costs. As a highly intelligent automated material handling tool in the logistics industry, automatic guided vehicles (AGVs for short) have been widely used in various industries to reduce the cost of enterprise operations. The development of the logistics industry has greatly promoted the degree of opening up and economic development. However, for large-sized heavy industrial logistics such as containers, conventional single AGV forklifts cannot meet their logistics and transportation requirements due to the limited carrying capacity. It is necessary to adopt a multi-AGV collaborative handling method to improve the actual efficiency of logistics handling. At present, most AGV forklifts are used in a path planning mode and travel along a pre-planned path. However, with the increase in the use environment, the increase in the number of AGVs and the extension of working hours, hidden dangers such as path conflicts and collisions to road locks are inevitable, which directly affect the paralysis of the entire logistics system and reduce the transportation efficiency of logistics. Moreover, AGV forklifts are often equipped with different heavy objects, and different heavy objects often have different sizes. Only through path planning and path collision detection, the overall situation of the heavy objects and the AGV forklifts is often not considered. In the implementation process, collisions are still prone to occur, which can easily lead to reduced handling capacity and AGV work efficiency. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a dynamic handling task coordination method based on a real-time AGV warehouse robot model.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] The first aspect of the present invention provides a method for coordinating handling tasks based on a dynamic real-time AGV warehouse robot model, comprising the following steps:

[0006] By setting a position detection sensor in the AGV warehouse robot, and obtaining the load information of the AGV warehouse robot through the position detection sensor, a dynamic model of the AGV warehouse robot is constructed according to the load information of the AGV warehouse robot;

[0007] Obtaining heavy object information and spatial structure diagram data in each cargo area in the target warehouse, and constructing a target warehouse model in each timestamp in the target warehouse according to the heavy object information and spatial structure diagram data in each cargo area in the target warehouse;

[0008] A target warehouse-AGV warehouse robot combined dynamic model is constructed according to the target warehouse model in each timestamp in the target warehouse and the AGV warehouse robot dynamic model, and an AGV warehouse robot tracking model is constructed;

[0009] The AGV warehouse robot tracking model is used to track the AGV warehouse robot in the target warehouse, and the dynamic model of the AGV warehouse robot is subjected to real-time collision detection. The final handling task coordination planning result is generated according to the collision detection result, and planning is performed according to the final handling task coordination planning result.

[0010] Furthermore, in the handling task coordination method based on the dynamic real-time AGV warehouse robot model, the dynamic model of the AGV warehouse robot is constructed according to the load information of the AGV warehouse robot, which specifically includes:

[0011] Obtaining spatial coordinate information of the load-bearing object on the load-bearing object information of the AGV warehouse robot according to the load-bearing object information of the AGV warehouse robot;

[0012] Acquire the appearance feature information of the AGV warehouse robot, and construct an initial static model of the AGV warehouse robot according to the appearance feature information of the AGV warehouse robot, and construct a load-bearing object model based on the spatial coordinate information of the load-bearing object on the load-bearing object information of the AGV warehouse robot;

[0013] The initial static model of the AGV warehouse robot and the load-carrying heavy object model are integrated to construct an AGV warehouse robot-heavy object fusion model, and the AGV warehouse robot-heavy object fusion model in each timestamp is obtained, and a dynamic demonstration is performed based on the AGV warehouse robot-heavy object fusion model in each timestamp;

[0014] Through dynamic demonstration, the dynamic model of each AGV storage robot in the target warehouse is obtained, and the dynamic model of each AGV storage robot in the target warehouse is output.

[0015] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, the heavy object information and the spatial structure diagram data in each cargo area in the target warehouse are obtained, and the target warehouse model in each timestamp in the target warehouse is constructed according to the heavy object information and the spatial structure diagram data in each cargo area in the target warehouse, specifically including:

[0016] Obtain the heavy object information in each cargo area in the target warehouse, calculate the Euclidean distance value between the heavy object information in each cargo area in the target warehouse, and regard the heavy objects whose Euclidean distance value is not greater than a preset Euclidean distance threshold as heavy objects of the same form;

[0017] The heavy objects whose Euclidean distance values ​​are greater than a preset Euclidean distance threshold are regarded as heavy objects of different shapes, and the quantity information of heavy objects of the same shape and different shapes is counted to obtain the volume information of heavy objects of the same shape and different shapes;

[0018] Based on the volume information of the weights of the same shape and non-same shape, construct weight model images of the same shape and non-same shape, and copy the weight model image of the same shape, count all weight model images of the same shape and non-same shape, and obtain a plurality of weight model images of the same shape and non-same shape;

[0019] Obtain the spatial structure diagram data in each cargo area in the target warehouse, and the coordinate positions of heavy objects of the same shape and of non-same shape; build a target warehouse model in the target warehouse based on the spatial structure diagram data in each cargo area in the target warehouse, and obtain the target warehouse model in the target warehouse at each timestamp.

[0020] Furthermore, in the handling task coordination method based on the dynamic real-time AGV warehouse robot model, a target warehouse-AGV warehouse robot combined dynamic model is constructed according to the target warehouse model at each timestamp in the target warehouse and the dynamic model of the AGV warehouse robot, specifically including:

[0021] By sorting the target warehouse models at each timestamp in the target warehouse in chronological order and constructing the target warehouse dynamic model, the dynamic model of the AGV warehouse robot is obtained;

[0022] Obtaining the position information of each AGV storage robot in the target storage at each time stamp, and integrating the AGV storage robot dynamic model into the target storage dynamic model based on the position information of each AGV storage robot in the target storage at each time stamp;

[0023] Through fusion, a target warehouse-AGV warehouse robot combined dynamic model is formed, and the target warehouse-AGV warehouse robot combined dynamic model is displayed in a preset manner.

[0024] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, an AGV warehouse robot tracking model is constructed, specifically:

[0025] An AGV warehouse robot tracking model is constructed based on a deep neural network, an initial static model of each AGV warehouse robot is obtained, and features of the initial static model of the AGV warehouse robot are extracted through a feature pyramid network to obtain morphological feature data of the AGV warehouse robot;

[0026] Introducing a cyclic spatial attention mechanism, inputting the morphological feature data of the AGV warehouse robot into the cyclic spatial attention mechanism, extracting the morphological feature data of the AGV warehouse robot, and obtaining a feature map with spatial attention;

[0027] Performing an inner product operation on the feature map with spatial attention and the morphological feature data of the AGV warehouse robot, focusing attention on the morphological feature data of the AGV warehouse robot, and generating an attention feature map;

[0028] The attention feature map is input into the hidden layer, and the state of the hidden layer is updated to save the model parameters of the AGV warehouse robot tracking model.

[0029] Furthermore, in the handling task coordination method based on the dynamic real-time AGV warehouse robot model, the AGV warehouse robot of the target warehouse is tracked by the AGV warehouse robot tracking model, and the dynamic model of the AGV warehouse robot is subjected to real-time collision detection, specifically:

[0030] Track the AGV warehouse robots in the target warehouse through the AGV warehouse robot tracking model to obtain the target motion trajectory information of each AGV warehouse robot and the motion information of each AGV warehouse robot;

[0031] Based on the target motion trajectory information of each AGV warehouse robot, the motion information of each AGV warehouse robot and the dynamic model of each AGV warehouse robot in the target warehouse, a motion demonstration is performed to obtain a dynamic motion model of each AGV warehouse robot;

[0032] Based on the dynamic motion model of each AGV warehouse robot, detect whether there is an intersection between the target motion trajectory information of each AGV warehouse robot, obtain the arrival time point of the AGV warehouse robot at the intersection, and calculate the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection;

[0033] When the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection is not greater than the deviation threshold, the corresponding time node and position are used as the collision node between the AGV warehouse robots and output as the collision detection result.

[0034] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, it also includes:

[0035] When the deviation threshold between the arrival time points of the AGV storage robot at the intersection is greater than the deviation threshold, the location information of the AGV storage robot at the same time node and the AGV storage robot-heavy object fusion model are obtained;

[0036] Constructing a detection space, representing the location information of the AGV warehouse robot at the same time node in the detection space, obtaining the representing location node, and inputting the AGV warehouse robot-heavy object fusion model into the detection space according to the representing location node;

[0037] Calculate whether there is interference between the AGV warehouse robot-heavy object fusion model in the detection space, and when there is interference, generate a collision node and output it as a collision detection result.

[0038] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, the final handling task coordination planning result is generated according to the collision detection result, and planning is performed according to the final handling task coordination planning result, specifically including:

[0039] An AGV storage robot with a collision detection result is obtained, and a driving route of the AGV storage robot with a collision detection result is replanned through a route planning algorithm to obtain a replanned route;

[0040] Determine whether there is still collision between AGV warehouse robots in the re-planned route. If not, output the re-planned route and generate the final handling task coordination planning result, and plan according to the final handling task coordination planning result;

[0041] If so, continue to replan the route until there is no collision between AGV warehouse robots in the replanned route, output the final replanned route, and use it as the final handling task coordination planning result, and plan according to the final handling task coordination planning result.

[0042] The second aspect of the present invention provides a handling task coordination system based on the dynamic real-time AGV warehouse robot model, the system includes a memory and a processor, the memory includes a handling task coordination method program based on the dynamic real-time AGV warehouse robot model, and when the handling task coordination method program based on the dynamic real-time AGV warehouse robot model is executed by the processor, any step of the handling task coordination method based on the dynamic real-time AGV warehouse robot model is implemented.

[0043] The third aspect of the present invention provides a computer-readable storage medium, which includes a handling task coordination method program based on the dynamic handling of a real-time AGV warehouse robot model. When the handling task coordination method program based on the dynamic handling of a real-time AGV warehouse robot model is executed by a processor, any step of the handling task coordination method based on the dynamic handling of a real-time AGV warehouse robot model is implemented.

[0044] The present invention solves the defects existing in the background technology and has the following beneficial effects:

[0045] The present invention sets a position detection sensor in the AGV warehouse robot, and obtains the load-carrying weight information of the AGV warehouse robot through the position detection sensor, constructs a dynamic model of the AGV warehouse robot according to the load-carrying weight information of the AGV warehouse robot, and then obtains the weight information and spatial structure diagram data in each cargo area in the target warehouse, and constructs the target warehouse model in each timestamp in the target warehouse according to the weight information and spatial structure diagram data in each cargo area in the target warehouse, thereby constructing a target warehouse-AGV warehouse robot combined dynamic model according to the target warehouse model in each timestamp in the target warehouse and the AGV warehouse robot dynamic model, constructs an AGV warehouse robot tracking model, and finally tracks the AGV warehouse robot of the target warehouse through the AGV warehouse robot tracking model, performs real-time collision detection on the AGV warehouse robot dynamic model, and generates a final handling task coordination planning result according to the collision detection result, and plans according to the final handling task coordination planning result. The present invention manages the AGV warehouse robot through three-dimensional model visualization during the execution of tasks, allowing managers to observe the actual working conditions of the AGV warehouse robot in real time. Moreover, through the dynamic analysis of the three-dimensional model, it can more accurately evaluate the execution status of the AGV warehouse robot during the task process and improve the coordination of the handling tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0047] Figure 1 The overall flow chart of the dynamic handling task coordination method based on the real-time AGV warehouse robot model is shown;

[0048] Figure 2A partial method flow chart of a method for dynamic handling task coordination based on a real-time AGV warehouse robot model is shown;

[0049] Figure 3 The system block diagram of the dynamic handling task coordination system based on the real-time AGV warehouse robot model is shown. DETAILED DESCRIPTION

[0050] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0052] like Figure 1 As shown, the first aspect of the present invention provides a dynamic handling task coordination method based on a real-time AGV warehouse robot model, comprising the following steps:

[0053] S102: by setting a position detection sensor in the AGV storage robot, and obtaining the load information of the AGV storage robot through the position detection sensor, a dynamic model of the AGV storage robot is constructed according to the load information of the AGV storage robot;

[0054] S104: Obtain the heavy object information and spatial structure diagram data in each cargo area in the target warehouse, and construct a target warehouse model in each timestamp in the target warehouse according to the heavy object information and spatial structure diagram data in each cargo area in the target warehouse;

[0055] S106: construct a target warehouse-AGV warehouse robot combined dynamic model according to the target warehouse model in each timestamp in the target warehouse and the AGV warehouse robot dynamic model, and construct an AGV warehouse robot tracking model;

[0056] S108: Track the AGV warehouse robot in the target warehouse through the AGV warehouse robot tracking model, perform real-time collision detection on the dynamic model of the AGV warehouse robot, and generate the final handling task coordination planning result based on the collision detection result, and plan according to the final handling task coordination planning result.

[0057] It should be noted that the present invention performs three-dimensional model visualization management on the AGV warehouse robot during the execution of tasks, which enables managers to observe the actual working conditions of the AGV warehouse robot in real time. Moreover, through the dynamic analysis of the three-dimensional model, it is possible to more accurately evaluate the execution status of the AGV warehouse robot during the task process and improve the coordination of the handling tasks.

[0058] Furthermore, in the handling task coordination method based on the real-time AGV warehouse robot model dynamic, the AGV warehouse robot dynamic model is constructed according to the load information of the AGV warehouse robot, specifically including:

[0059] According to the load information of the AGV warehouse robot, the spatial coordinate information of the load on the load information of the AGV warehouse robot is obtained;

[0060] Obtain the appearance feature information of the AGV warehouse robot (contour feature, volume feature, length feature, morphology feature and other data), and build the initial static model of the AGV warehouse robot based on the appearance feature information of the AGV warehouse robot, and build the load model based on the spatial coordinate information of the load on the load information of the AGV warehouse robot;

[0061] The initial static model of the AGV warehouse robot and the load-carrying heavy object model are integrated to construct an AGV warehouse robot-heavy object fusion model, and the AGV warehouse robot-heavy object fusion model at each timestamp is obtained. A dynamic demonstration is performed based on the AGV warehouse robot-heavy object fusion model at each timestamp.

[0062] Through dynamic demonstration, the dynamic model of each AGV storage robot in the target warehouse is obtained, and the dynamic model of each AGV storage robot in the target warehouse is output.

[0063] It should be noted that the load information includes the mass information of the load, the type information of the load, the volume information of the load, etc., wherein the initial static model of the AGV warehouse robot is constructed according to the appearance feature information of the AGV warehouse robot through 3D modeling software (such as SolidWorks, Maya and other 3D modeling software), so as to obtain the dynamic model of each AGV warehouse robot in the target warehouse through dynamic demonstration, and dynamize the AGV warehouse robot. Among them, when there is no cargo or the cargo is unloaded, there is no heavy object in the dynamic model of the AGV warehouse robot, which is equivalent to the initial static model of the AGV warehouse robot in each timestamp, and the dynamic model of the AGV warehouse robot in a time series.

[0064] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, the heavy object information and spatial structure diagram data in each cargo area in the target warehouse are obtained, and the target warehouse model in each timestamp in the target warehouse is constructed according to the heavy object information and spatial structure diagram data in each cargo area in the target warehouse, specifically including:

[0065] Obtain the heavy object information in each cargo area in the target warehouse, calculate the Euclidean distance value between the heavy object information in each cargo area in the target warehouse, and regard the heavy objects whose Euclidean distance value is not greater than a preset Euclidean distance threshold as heavy objects of the same form;

[0066] The heavy objects whose Euclidean distance values ​​are greater than a preset Euclidean distance threshold are regarded as heavy objects of different shapes, and the quantity information of heavy objects of the same shape and different shapes is counted to obtain the volume information of heavy objects of the same shape and different shapes;

[0067] Based on the volume information of the weights of the same shape and the weights of the different shapes, model diagrams of the weights of the same shape and the weights of different shapes are constructed, and the model diagrams of the weights of the same shape are copied, and all the model diagrams of the weights of the same shape and the weights of different shapes are counted to obtain a plurality of model diagrams of the weights of the same shape and the weights of different shapes;

[0068] Obtain the spatial structure diagram data in each cargo area in the target warehouse, the coordinate positions of the heavy objects of the same shape and of non-same shape, build a target warehouse model in the target warehouse based on the spatial structure diagram data in each cargo area in the target warehouse, and obtain the target warehouse model in the target warehouse at each timestamp.

[0069] It should be noted that the spatial structure diagram data includes the plane structure layout diagram of the target warehouse, factory design drawings, factory area data, etc., and the model diagrams of heavy objects with the same shape and non-same shape are constructed based on the volume information of heavy objects with the same shape and non-same shape, and the model diagrams of heavy objects with the same shape are copied, and all the model diagrams of heavy objects with the same shape and non-same shape are counted, and several model diagrams of heavy objects with the same shape and non-same shape are obtained. This can speed up the construction of the model and improve the rate of model construction.

[0070] Furthermore, in the handling task coordination method based on the dynamic real-time AGV warehouse robot model, a target warehouse-AGV warehouse robot combined dynamic model is constructed according to the target warehouse model at each timestamp in the target warehouse and the dynamic model of the AGV warehouse robot, specifically including:

[0071] By sorting the target warehouse models at each timestamp in the target warehouse in chronological order and constructing the target warehouse dynamic model, the dynamic model of the AGV warehouse robot is obtained;

[0072] Obtain the location information of each AGV storage robot in the target storage at each timestamp, and integrate the AGV storage robot dynamic model into the target storage dynamic model based on the location information of each AGV storage robot in the target storage at each timestamp;

[0073] Through fusion, a target warehouse-AGV warehouse robot combined dynamic model is formed, and the target warehouse-AGV warehouse robot combined dynamic model is displayed in a preset manner.

[0074] It should be noted that this method can obtain the target warehouse-AGV warehouse robot combined dynamic model, which is convenient for visual management. Various models of AGV warehouse robots and various models of warehouses can be completed through three-dimensional modeling software.

[0075] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, an AGV warehouse robot tracking model is constructed, specifically:

[0076] An AGV warehouse robot tracking model is constructed based on a deep neural network to obtain the initial static model of each AGV warehouse robot. The initial static model of the AGV warehouse robot is extracted through a feature pyramid network to obtain the morphological feature data of the AGV warehouse robot.

[0077] Introduce the cyclic spatial attention mechanism, input the morphological feature data of the AGV warehouse robot into the cyclic spatial attention mechanism, extract the morphological feature data of the AGV warehouse robot, and obtain a feature map with spatial attention;

[0078] Perform an inner product operation on the feature map with spatial attention and the morphological feature data of the AGV warehouse robot, focus the attention on the morphological feature data of the AGV warehouse robot, and generate an attention feature map;

[0079] The attention feature map is input into the hidden layer, and the state of the hidden layer is updated to save the model parameters of the AGV warehouse robot tracking model.

[0080] It should be noted that by inputting the morphological feature data of the AGV warehouse robot into the recurrent spatial attention mechanism, the morphological feature data of the AGV warehouse robot is extracted to obtain a feature map with spatial attention, so that the attention is focused on the morphological feature data of the AGV warehouse robot, avoiding the influence of multi-scale data on the AGV warehouse robot tracking model, simplifying the data training of the model, and improving the calculation speed of the model.

[0081] like Figure 2 As shown, further, in the handling task coordination method based on the dynamic real-time AGV warehouse robot model, the AGV warehouse robot of the target warehouse is tracked by the AGV warehouse robot tracking model, and the AGV warehouse robot dynamic model is subjected to real-time collision detection, specifically:

[0082] S202: Tracking the AGV warehouse robots in the target warehouse through the AGV warehouse robot tracking model to obtain the target motion trajectory information of each AGV warehouse robot and the motion information of each AGV warehouse robot;

[0083] S204: Performing motion demonstration based on the target motion trajectory information of each AGV warehouse robot, the motion information of each AGV warehouse robot, and the dynamic model of each AGV warehouse robot in the target warehouse to obtain the dynamic motion model of each AGV warehouse robot;

[0084] S206: Detect whether there is an intersection between the target motion trajectory information of each AGV warehouse robot based on the dynamic motion model of each AGV warehouse robot, obtain the arrival time point of the AGV warehouse robot at the intersection, and calculate the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection;

[0085] S208: When the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection is not greater than the deviation threshold, the corresponding time node and position are used as the collision node between the AGV warehouse robots and output as the collision detection result.

[0086] It should be noted that when the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection is not greater than the deviation threshold, the corresponding time node and position will be used as the collision node between the AGV warehouse robots, so that the collision situation of the AGV warehouse robots can be detected and analyzed in time.

[0087] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, it also includes:

[0088] When the deviation threshold between the arrival time points of the AGV storage robot at the intersection is greater than the deviation threshold, the location information of the AGV storage robot at the same time node and the AGV storage robot-heavy object fusion model are obtained;

[0089] Construct a detection space, represent the location information of the AGV warehouse robot at the same time node in the detection space, obtain the represented location node, and input the AGV warehouse robot-heavy object fusion model into the detection space according to the represented location node;

[0090] Calculate whether there is interference between the AGV warehouse robot-heavy object fusion model in the detection space. When there is interference, generate a collision node and output it as a collision detection result.

[0091] It should be noted that by constructing the detection space, the location information of the AGV warehouse robot at the same time node is represented in the detection space, the position node is obtained, and the AGV warehouse robot-heavy object fusion model is input into the detection space according to the position node to calculate whether there is interference between the AGV warehouse robot-heavy object fusion model in the detection space. When there is interference, a collision node is generated, which fully considers the actual situation of the AGV warehouse robot and the heavy object, and improves the accuracy of collision detection compared to the existing technology. The intersection means the position node where the motion trajectory of the AGV warehouse robot overlaps.

[0092] Furthermore, in the dynamic handling task coordination method based on the real-time AGV warehouse robot model, the final handling task coordination planning result is generated according to the collision detection result, and the planning is performed according to the final handling task coordination planning result, specifically including:

[0093] An AGV storage robot with a collision detection result is obtained, and a driving route of the AGV storage robot with a collision detection result is replanned through a route planning algorithm to obtain a replanned route;

[0094] Determine whether there is still collision between AGV warehouse robots in the re-planned route. If not, output the re-planned route and generate the final handling task coordination planning result, and plan according to the final handling task coordination planning result;

[0095] If so, continue to replan the route until there is no collision between AGV warehouse robots in the replanned route, output the final replanned route, and use it as the final handling task coordination planning result, and plan according to the final handling task coordination planning result.

[0096] It should be noted that the route planning algorithm includes optimization algorithms such as A* algorithm, ant colony algorithm, greedy algorithm, particle swarm algorithm, etc. This method can optimize the handling task coordination of AGV warehouse robots.

[0097] In addition, the method further comprises:

[0098] Obtain a dynamic model of each AGV storage robot in the target warehouse, and obtain mass data of the load-bearing objects on the AGV storage robot, and perform a force analysis on the dynamic model of the AGV storage robot based on the mass data of the load-bearing objects on the AGV storage robot;

[0099] Through force analysis, the position of the center of gravity of each timestamp is obtained, and based on the position of the center of gravity of each timestamp, a time series-based center of gravity position state change feature is constructed, and the time series-based center of gravity position state change feature is input into a Markov model;

[0100] Calculate the transition probability value of the center of gravity position at each time stamp in the center of gravity position state change feature based on the time series to another center of gravity position by using the Markov model;

[0101] When the probability value of the transfer of the center of gravity position to another center of gravity position is greater than a preset probability value, the center of gravity position in the corresponding timestamp is updated to another center of gravity position, and the center of gravity position state change feature based on the time series is updated;

[0102] When the transfer probability value of the center of gravity position transferring to another center of gravity position is not greater than the preset probability value, the state of the center of gravity position in the corresponding timestamp is maintained unchanged, and the center of gravity position data of the current timestamp is obtained based on the center of gravity position state change characteristics of the time series, and collision detection is performed on the dynamic model of each AGV warehouse robot in the target warehouse based on the center of gravity position data of the current timestamp.

[0103] It should be noted that the Markov model (Markov Chain) describes a state sequence, each state value of which depends on the previous finite state. Markov chain is a sequence of random variables X1, X2, X3... Xn with Markov properties. The range of these variables, that is, the set of all their possible values, is called the "state space", and the value of Xn is the state at time n. In this method, since the position state of the center of gravity is a variable, that is, the position state variable of the center of gravity within a predetermined time is X1, X2, X3... Xn, where, from the existing technology and the rules of the Markov model, it can be known that when the transfer probability value of the position state variable X1 of the center of gravity to other position state variables Xn is greater than the preset probability value (for example, the set probability value is 95%), it indicates that the state transfer occurs, that is, the transfer from the position state variable X1 of the center of gravity to other position state variables Xn, at this time, the position of the center of gravity of the AGV warehouse robot is updated, otherwise, the position state of the center of gravity of the AGV warehouse robot remains unchanged.

[0104] During the handling process, the center of gravity of the AGV warehouse robot will shift due to the influence of heavy objects, which will cause the AGV warehouse robot to shift during movement, and then cause a collision. By updating the center of gravity position state change characteristics based on the time series through the Markov model, the accuracy of obtaining the center of gravity position can be improved, thereby improving the accuracy of collision detection.

[0105] In addition, performing collision detection on the dynamic model of each AGV storage robot in the target warehouse based on the center of gravity position data of the current timestamp also includes:

[0106] Based on the center of gravity position data of the current timestamp, an offset analysis is performed on the dynamic model of each AGV warehouse robot in the target warehouse, the offset dynamic model of the AGV warehouse robot is obtained, and a dynamic simulation is performed on the dynamic model of the AGV warehouse robot;

[0107] When the AGV warehouse robot dynamic model dumps during the dynamic simulation process, the location information of the corresponding AGV warehouse robot is obtained, and an early warning prompt is given according to the location information of the corresponding AGV warehouse robot, and displayed in a preset manner;

[0108] When the AGV warehouse robot dynamic model does not fall over during the dynamic simulation process, it is determined whether there is interference between the AGV warehouse robot dynamic models after each offset at the same time node;

[0109] When there is interference, a collision node is generated and used as the collision detection result. The final handling task coordination planning result is generated based on the collision detection result, and planning is performed according to the final handling task coordination planning result.

[0110] It should be noted that, in fact, when the position of the center of gravity changes, the dynamic model of the AGV warehouse robot will generally be offset. Through this method, the dynamic model of each AGV warehouse robot in the target warehouse can be offset and updated, further improving the collision detection accuracy.

[0111] like Figure 3 As shown, the second aspect of the present invention provides a handling task coordination system 4 based on the dynamic real-time AGV warehouse robot model. The system 4 includes a memory 41 and a processor 42. The memory 41 includes a handling task coordination method program based on the dynamic real-time AGV warehouse robot model. When the handling task coordination method program based on the dynamic real-time AGV warehouse robot model is executed by the processor 42, any step of the handling task coordination method based on the dynamic real-time AGV warehouse robot model is implemented.

[0112] The third aspect of the present invention provides a computer-readable storage medium, which includes a handling task coordination method program based on the dynamic handling of real-time AGV warehouse robot model. When the handling task coordination method program based on the dynamic handling of real-time AGV warehouse robot model is executed by a processor, any step of the handling task coordination method based on the dynamic handling of real-time AGV warehouse robot model is implemented.

[0113] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0114] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0115] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0116] Those skilled in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), disks or optical disks, and other media that can store program codes.

[0117] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0118] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A dynamic handling task coordination method based on a real-time AGV warehouse robot model, characterized in that: The following steps are involved: By setting a position detection sensor in the AGV warehouse robot, and obtaining the load information of the AGV warehouse robot through the position detection sensor, a dynamic model of the AGV warehouse robot is constructed according to the load information of the AGV warehouse robot; Obtaining heavy object information and spatial structure diagram data in each cargo area in the target warehouse, and constructing a target warehouse model in each timestamp in the target warehouse according to the heavy object information and spatial structure diagram data in each cargo area in the target warehouse; A target warehouse-AGV warehouse robot combined dynamic model is constructed according to the target warehouse model in each timestamp in the target warehouse and the AGV warehouse robot dynamic model, and an AGV warehouse robot tracking model is constructed; Track the AGV warehouse robot in the target warehouse through the AGV warehouse robot tracking model, perform real-time collision detection on the dynamic model of the AGV warehouse robot, generate the final handling task coordination planning result according to the collision detection result, and plan according to the final handling task coordination planning result; Construct an AGV warehouse robot tracking model, specifically: An AGV warehouse robot tracking model is constructed based on a deep neural network, an initial static model of each AGV warehouse robot is obtained, and features of the initial static model of the AGV warehouse robot are extracted through a feature pyramid network to obtain morphological feature data of the AGV warehouse robot; A cyclic spatial attention mechanism is introduced, the morphological feature data of the AGV warehouse robot is input into the cyclic spatial attention mechanism, the morphological feature data of the AGV warehouse robot is processed, and a feature map with spatial attention is obtained; Performing an inner product operation on the feature map with spatial attention and the morphological feature data of the AGV warehouse robot, focusing attention on the morphological feature data of the AGV warehouse robot, and generating an attention feature map; The attention feature map is input into the hidden layer, and the state of the hidden layer is updated to save the model parameters of the AGV warehouse robot tracking model.

2. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 1 is characterized in that: The dynamic model of the AGV warehouse robot is constructed according to the load information of the AGV warehouse robot, specifically including: Obtaining spatial coordinate information of the load-bearing object on the load-bearing object information of the AGV warehouse robot according to the load-bearing object information of the AGV warehouse robot; Acquire the appearance feature information of the AGV warehouse robot, and construct an initial static model of the AGV warehouse robot according to the appearance feature information of the AGV warehouse robot, and construct a load-bearing object model based on the spatial coordinate information of the load-bearing object on the load-bearing object information of the AGV warehouse robot; The initial static model of the AGV warehouse robot and the load-carrying heavy object model are integrated to construct an AGV warehouse robot-heavy object fusion model, and the AGV warehouse robot-heavy object fusion model in each timestamp is obtained, and a dynamic demonstration is performed based on the AGV warehouse robot-heavy object fusion model in each timestamp; Through dynamic demonstration, the dynamic model of each AGV storage robot in the target warehouse is obtained, and the dynamic model of each AGV storage robot in the target warehouse is output.

3. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 1 is characterized in that: Obtaining the heavy object information and the spatial structure diagram data in each cargo area in the target warehouse, and constructing the target warehouse model in each timestamp in the target warehouse according to the heavy object information and the spatial structure diagram data in each cargo area in the target warehouse, specifically including: Obtain the heavy object information in each cargo area in the target warehouse, calculate the Euclidean distance value between the heavy object information in each cargo area in the target warehouse, and regard the heavy objects whose Euclidean distance value is not greater than a preset Euclidean distance threshold as heavy objects of the same form; The heavy objects whose Euclidean distance values ​​are greater than a preset Euclidean distance threshold are regarded as heavy objects of different shapes, and the quantity information of heavy objects of the same shape and different shapes is counted to obtain the volume information of heavy objects of the same shape and different shapes; Based on the volume information of the weights of the same shape and non-same shape, construct weight model images of the same shape and non-same shape, and copy the weight model image of the same shape, count all weight model images of the same shape and non-same shape, and obtain a plurality of weight model images of the same shape and non-same shape; Obtain the spatial structure diagram data in each cargo area in the target warehouse, and the coordinate positions of heavy objects of the same shape and of non-same shape; build a target warehouse model in the target warehouse based on the spatial structure diagram data in each cargo area in the target warehouse, and obtain the target warehouse model in the target warehouse at each timestamp.

4. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 1 is characterized in that: The target warehouse-AGV warehouse robot combined dynamic model is constructed according to the target warehouse model at each timestamp in the target warehouse and the AGV warehouse robot dynamic model, specifically including: By sorting the target warehouse models at each timestamp in the target warehouse in chronological order and constructing the target warehouse dynamic model, the dynamic model of the AGV warehouse robot is obtained; Obtaining the position information of each AGV storage robot in the target storage at each time stamp, and integrating the AGV storage robot dynamic model into the target storage dynamic model based on the position information of each AGV storage robot in the target storage at each time stamp; Through fusion, a target warehouse-AGV warehouse robot combined dynamic model is formed, and the target warehouse-AGV warehouse robot combined dynamic model is displayed in a preset manner.

5. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 1 is characterized in that: The AGV warehouse robot tracking model is used to track the AGV warehouse robot in the target warehouse, and the dynamic model of the AGV warehouse robot is subjected to real-time collision detection, specifically: Track the AGV warehouse robots in the target warehouse through the AGV warehouse robot tracking model to obtain the target motion trajectory information of each AGV warehouse robot and the motion information of each AGV warehouse robot; Based on the target motion trajectory information of each AGV warehouse robot, the motion information of each AGV warehouse robot and the dynamic model of each AGV warehouse robot in the target warehouse, a motion demonstration is performed to obtain a dynamic motion model of each AGV warehouse robot; Based on the dynamic motion model of each AGV warehouse robot, detect whether there is an intersection between the target motion trajectory information of each AGV warehouse robot, obtain the arrival time point of the AGV warehouse robot at the intersection, and calculate the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection; When the deviation threshold between the arrival time points of the AGV warehouse robots at the intersection is not greater than the deviation threshold, the corresponding time node and position are used as the collision node between the AGV warehouse robots and output as the collision detection result.

6. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 5 is characterized in that: Also includes: When the deviation threshold between the arrival time points of the AGV storage robot at the intersection is greater than the deviation threshold, the location information of the AGV storage robot at the same time node and the AGV storage robot-heavy object fusion model are obtained; Constructing a detection space, representing the location information of the AGV warehouse robot at the same time node in the detection space, obtaining the representing location node, and inputting the AGV warehouse robot-heavy object fusion model into the detection space according to the representing location node; Calculate whether there is interference between the AGV warehouse robot-heavy object fusion model in the detection space, and when there is interference, generate a collision node and output it as a collision detection result.

7. The method for coordinating handling tasks based on the dynamic real-time AGV warehouse robot model according to claim 1 is characterized in that: The final handling task coordination planning result is generated according to the collision detection result, and the planning is performed according to the final handling task coordination planning result, specifically including: An AGV storage robot with a collision detection result is obtained, and a driving route of the AGV storage robot with a collision detection result is replanned through a route planning algorithm to obtain a replanned route; Determine whether there is still collision between AGV warehouse robots in the re-planned route. If not, output the re-planned route and generate the final handling task coordination planning result, and plan according to the final handling task coordination planning result; If so, continue to replan the route until there is no collision between AGV warehouse robots in the replanned route, output the final replanned route, and use it as the final handling task coordination planning result, and plan according to the final handling task coordination planning result.

8. A dynamic handling task coordination system based on a real-time AGV warehouse robot model, characterized in that: The system includes a memory and a processor, wherein the memory includes a handling task coordination method program based on a dynamic real-time AGV warehouse robot model. When the handling task coordination method program based on a dynamic real-time AGV warehouse robot model is executed by the processor, the steps of the handling task coordination method based on a dynamic real-time AGV warehouse robot model as described in any one of claims 1-7 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a handling task coordination method program based on a dynamic real-time AGV warehouse robot model. When the handling task coordination method program based on a dynamic real-time AGV warehouse robot model is executed by a processor, the steps of the handling task coordination method based on a dynamic real-time AGV warehouse robot model as described in any one of claims 1-7 are implemented.

Images