Logistics vehicle scheduling method based on visual perception industrial vehicle ADAS system

Abstract

The present invention relates to the technical field of logistics vehicle scheduling, and in particular to a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system. The method obtains the speed of the logistics vehicle, and then obtains the reference time for the target vehicle to reach a designated location at the previous moment; predicts the target position of the non-idle logistics vehicle according to the speed of the non-idle logistics vehicle and the path length of its position in the planned path corresponding to the reference time to reach the designated location; based on the planned direction and position of the target vehicle at the current moment, the target position and speed of the non-idle logistics vehicle, obtains the target prediction time, corrects the cost function of the target vehicle reaching the designated location in the A‑star algorithm, and obtains the corrected cost function to schedule the logistics vehicle. The present invention reduces the congestion and path conflict of logistics vehicles by correcting the cost function in the A‑star algorithm, and improves the accuracy and efficiency of logistics vehicle scheduling.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics vehicle scheduling, and in particular to a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system. Background Art

[0002] With the rapid development of smart logistics, industrial vehicles are increasingly used in logistics and warehousing, especially automated guided vehicles and forklifts, which have become important tools for efficient material handling. With the popularization of industrial vehicle ADAS (Advanced Driver Assistance System) technology, industrial vehicle ADAS systems based on visual perception have gradually acquired the ability to obtain real-time environmental information, effectively improving the safety, operating efficiency and automation of industrial vehicles (such as forklifts, stackers, storage vehicles, etc.), and providing a technical basis for further optimizing vehicle scheduling.

[0003] Existing methods use the cost function in the A-star algorithm to schedule logistics vehicles, mainly taking the path length and logistics vehicle status as the scheduling basis for logistics vehicles, and completing the path planning of logistics vehicles at the same time. However, in actual situations, logistics vehicles may be congested during operation. The A-star algorithm does not consider the dynamic congestion of logistics vehicles on the path, lacks the ability to predict and dynamically adjust possible congestion in the future, making it difficult to optimize the paths of logistics vehicles in a timely manner according to the rapid changes in the logistics environment, limiting the accuracy of multi-logistics vehicle scheduling, and easily leading to congestion of logistics vehicles, which in turn easily causes delays in material handling tasks. Summary of the invention

[0004] In order to solve the technical problem that the A-star algorithm does not consider the dynamic congestion of logistics vehicles on the path, which limits the accuracy of multi-logistics vehicle scheduling and causes delays in material handling tasks, the purpose of the present invention is to provide a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system. The technical solution adopted is as follows:

[0005] An embodiment of the present invention provides a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system, the method comprising the following steps:

[0006] Obtain the speed of each logistics vehicle in the logistics area based on the visual perception industrial vehicle ADAS system;

[0007] Any idle logistics vehicle in the logistics area at the current moment is taken as the target vehicle, and the ratio of the path length from the position of the target vehicle at the previous moment to the designated location to the speed of the target vehicle is taken as the reference time for the target vehicle to reach the designated location;

[0008] According to the reference duration and the speed of each non-idle logistics vehicle in the corresponding time period, and the path length of each non-idle logistics vehicle from each position in the planned path corresponding to the reference duration to the designated location, the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated location is predicted;

[0009] Based on the planned travel direction and position of the target vehicle at the current moment, and the target position and speed of each non-idle logistics vehicle, obtain the target predicted time for the target vehicle to reach the specified location;

[0010] According to the speed of the target vehicle and the target prediction time, the cost function of the target vehicle reaching the specified position in the A-star algorithm is corrected to obtain a corrected cost function;

[0011] Based on the modified cost function, logistics vehicles are scheduled for the visual perception industrial vehicle ADAS system.

[0012] Furthermore, the method for obtaining the path length from the position of the target vehicle at the previous moment to the specified position is:

[0013] The cost function in the A-star algorithm is used to obtain the path length from the target vehicle's position at the previous moment to the specified position.

[0014] Furthermore, the method for predicting the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated location is:

[0015] For any non-idle logistics vehicle in the time period corresponding to the reference duration, obtain the product of the speed of the non-idle logistics vehicle and the reference duration as the reference travel distance of the non-idle logistics vehicle;

[0016] Obtain each location that the non-idle logistics vehicle passes through during the planned path within the time period corresponding to the reference duration, and use them as reference locations;

[0017] Obtain the path length of the non-idle logistics vehicle from each reference location to the designated location, which is used as the reference path length;

[0018] Obtaining a difference between each reference path length and the reference travel distance as a first difference;

[0019] The reference position corresponding to the smallest first difference is predicted as the target position of the non-idle logistics vehicle when the target vehicle arrives at the designated position.

[0020] Furthermore, the method for obtaining the target predicted time for the target vehicle to reach the specified location is:

[0021] According to the planned moving direction and position of the target vehicle at the current moment, the moving direction area of ​​the target vehicle at the current moment is obtained;

[0022] According to the distance between the target position of each non-idle logistics vehicle in the time period corresponding to the reference duration and the travel direction area of ​​the target vehicle at the current moment, the reference non-idle logistics vehicles are selected;

[0023] According to the location distribution and speed of each reference non-idle logistics vehicle, as well as the travel direction area and planned travel direction of the target vehicle at the current moment, the target predicted time for the target vehicle to reach the specified location is obtained.

[0024] Furthermore, the method for obtaining the moving direction area of ​​the target vehicle at the current moment is:

[0025] The previous position that the target vehicle passed before reaching the current position is used as the historical reference position;

[0026] The direction of the target vehicle from the historical reference position to the current position is used as the planned travel direction of the target vehicle at the current moment;

[0027] Obtain the traversable grids in the logistics area, and point the target vehicle's current position to the direction of the center point of each traversable grid, all of which are taken as the first direction;

[0028] Obtain the distance between the current position of the target vehicle and the center point of each traversable grid, all as the first distance;

[0029] The area formed by the walkable grids corresponding to the first distance being less than the preset distance threshold and the angle between the first direction and the planned traveling direction of the target vehicle at the current moment being less than the preset angle threshold is taken as the traveling direction area of ​​the target vehicle at the current moment.

[0030] Furthermore, the method for obtaining the reference non-idle logistics vehicles is:

[0031] Obtain the minimum distance between the target position of each non-idle logistics vehicle and the target vehicle's current traveling direction area within the time period corresponding to the reference duration as the target distance of each non-idle logistics vehicle; wherein, when the target position of the non-idle logistics vehicle is within the target vehicle's current traveling direction area, the corresponding target distance is 0;

[0032] When the number of non-idle logistics vehicles corresponding to the target distance of 0 is greater than or equal to the preset number threshold, all non-idle logistics vehicles corresponding to the target distance of 0 are used as reference non-idle logistics vehicles;

[0033] When the number of non-idle logistics vehicles corresponding to the target distance of 0 is less than the preset number threshold, the target distances are arranged from small to large to obtain a target distance sequence;

[0034] The non-idle logistics vehicles corresponding to the target distances before a preset number threshold in the target distance sequence are all used as reference non-idle logistics vehicles.

[0035] Furthermore, the method for obtaining the target predicted time for the target vehicle to reach the specified location is:

[0036] The result of negatively correlating and normalizing the target distance of each reference non-idle logistics vehicle is used as the reference weight of each reference non-idle logistics vehicle;

[0037] Obtain the occlusion density of the target vehicle at the current moment according to the reference weight of each reference non-idle logistics vehicle and the number of walkable grids it occupies at the target position, as well as the number of walkable grids in the travel direction area of ​​the target vehicle at the current moment;

[0038] The speed of each reference non-idle logistics vehicle, the obstruction density of the target vehicle at the current moment, and the planned travel direction are taken as input, and the target predicted time for the target vehicle to reach the specified location is output through the trained LSTM network.

[0039] Furthermore, the method for obtaining the obstruction density of the target vehicle at the current moment is:

[0040] The product of the reference weight of each reference non-idle logistics vehicle and the number of traversable grids occupied by it at the target location is taken as the reference number of traversable grids occupied by each reference non-idle logistics vehicle;

[0041] The ratio of the cumulative number of reference occupied walkable grids of all reference non-idle logistics vehicles to the number of walkable grids in the traveling direction area of ​​the target vehicle at the current moment is taken as the occlusion density of the target vehicle at the current moment.

[0042] Furthermore, the method for obtaining the trained LSTM network is:

[0043] According to the planned travel direction and position of each logistics vehicle at the congestion moment in the historical time period, the travel direction area of ​​each logistics vehicle at the congestion moment is obtained;

[0044] The total number of traversable grids occupied by non-idle logistics vehicles in the travel direction area of ​​each logistics vehicle at the congested moment is obtained through the end-side large model as the first number of each logistics vehicle at the congested moment;

[0045] The ratio of the first number of each logistics vehicle at the congestion moment to the number of all walkable grids in the area in its travel direction is used as the blocking density of each logistics vehicle at the congestion moment;

[0046] The density of obstructions and the planned travel direction of each logistics vehicle at the time of congestion, as well as the speed of non-idle logistics vehicles in the travel direction area of ​​each logistics vehicle at the time of congestion are taken as input, and the duration of each logistics vehicle in the travel direction area at the time of congestion is used as output to train the LSTM network and obtain a trained LSTM network.

[0047] Furthermore, the method for obtaining the modified cost function is:

[0048] The product of the target vehicle's speed and the target predicted time is taken as the congestion cost for the target vehicle to reach the designated location;

[0049] The cost function for the target vehicle to reach the designated location and the congestion cost are added together to obtain the corrected cost function for the target vehicle to reach the designated location.

[0050] The present invention has the following beneficial effects:

[0051] The present invention uses the ratio of the path length from the position of the target vehicle at the previous moment to the designated position to the speed of the target vehicle as the reference time for the target vehicle to arrive at the designated position, accurately reflecting the time consumed by the target vehicle to arrive at the designated position without considering the congestion of logistics vehicles; in order to analyze the congestion of logistics vehicles in the process of the target vehicle arriving at the designated path and improve the dynamic adaptability and accuracy of path planning, the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated position is predicted based on the reference time and the speed of each non-idle logistics vehicle in the corresponding time period, and the path length from each position in the planned path corresponding to the reference time to the designated position, accurately reflecting the congestion impact that may be caused by each non-idle logistics vehicle in the process of the target vehicle arriving at the designated position; based on the planned travel direction and position of the target vehicle at the current moment, and each non-idle logistics vehicle The target position and speed of idle logistics vehicles are used to accurately obtain the target predicted time for the target vehicle to reach the designated location, which reflects the time it takes for the target vehicle to reach the designated location when considering the congestion caused by non-idle logistics vehicles. In order to accurately plan the path for the target vehicle to reach the designated location, the cost function for the target vehicle to reach the designated location in the A-star algorithm is corrected according to the speed of the target vehicle and the target predicted time, and the corrected cost function is obtained to accurately obtain the cost and the optimal planned path for the target vehicle to reach the designated location. Based on the corrected cost function, logistics vehicles based on the visual perception industrial vehicle ADAS system are dispatched, which effectively improves the intelligence level and efficiency of logistics vehicle scheduling, improves the prediction and dynamic adjustment capabilities of possible congestion in the future, reduces congestion and path conflicts of logistics vehicles, and significantly optimizes the overall operation performance of the industrial logistics system, thereby improving the efficiency of material handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 A schematic flow chart of a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided by an embodiment of the present invention;

[0054] Figure 2 A flow chart of a method for obtaining target prediction duration provided by one embodiment of the present invention;

[0055] Figure 3A structural diagram of a logistics vehicle dispatching system based on a visual perception industrial vehicle ADAS system provided by one embodiment of the present invention;

[0056] Figure 4 A schematic diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the logistics vehicle scheduling method based on the visual perception industrial vehicle ADAS system proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0058] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0059] The specific scheme of the logistics vehicle scheduling method based on the visual perception industrial vehicle ADAS system provided by the present invention is described in detail below with reference to the accompanying drawings.

[0060] Embodiment 1:

[0061] The specific scenario of this embodiment is: in the existing method, the logistics vehicles in the logistics area are scheduled by the A-star algorithm. In actual situations, many logistics vehicles will be congested on the running path. The A-star algorithm does not consider the congestion of logistics vehicles on the path, which easily leads to inaccurate logistics vehicles scheduled by the A-star algorithm, causing congestion, and then causing delays in material handling tasks. In order to accurately schedule logistics vehicles in the logistics area, this embodiment installs various visual devices on the logistics vehicles to analyze the path congestion of the logistics vehicles during the driving process, and then accurately predicts the duration of the logistics vehicles passing through the congestion stage, and further combines the travel changes of multiple non-idle logistics vehicles on the planned paths, and corrects the cost function of the idle logistics vehicles obtained by the A-star algorithm to improve the dynamic adaptability and accuracy of the path planning; and then based on the corrected cost function, the logistics vehicles based on the visual perception industrial vehicle ADAS system are accurately scheduled, which effectively improves the intelligence level and efficiency of logistics vehicle scheduling, reduces congestion and path conflicts, improves material handling efficiency, and significantly optimizes the overall operation performance of the industrial logistics system. Among them, the A-star algorithm is a well-known technology and will not be repeated.

[0062] This invention proposes a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system. Figure 1 , which shows a schematic flow chart of a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided by an embodiment of the present invention, the method comprising the following steps:

[0063] Step S1: Obtain the speed of each logistics vehicle in the logistics area based on the visual perception industrial vehicle ADAS system.

[0064] Specifically, in order to accurately plan the paths of logistics vehicles, before the logistics vehicles are dispatched, this embodiment first uses a laser scanner, camera or other visual sensor equipment to scan the logistics area to accurately obtain a two-dimensional map of the logistics area. The two-dimensional map includes all fixed structures within the logistics area, such as shelves, workstations, and other areas where logistics vehicles cannot pass and areas where logistics vehicles can drive. When planning the paths of logistics vehicles, the A-star algorithm needs to evaluate the costs of multiple paths. If the drivable area is directly represented by continuous two-dimensional coordinates, the state space will be too large and the computational complexity will increase significantly. Therefore, this embodiment divides the logistics area into regular grids of fixed size and discretizes the continuous area, which can reduce the search space and improve the efficiency of the A-star algorithm. This embodiment sets the size of the regular grid to , the implementer can set the size of the regular grid according to the actual situation, and there is no limitation here.

[0065] When the divided grid contains areas that vehicles cannot pass through, such as shelves and workstations, the corresponding grid is recorded as an untravelable grid; considering that the turning ability of logistics vehicles is limited (a larger radius is required to turn), and the movement trajectory of logistics vehicles may not be completely consistent with the planned path due to errors during operation, it is necessary to leave a certain safety margin for the vehicle width. Therefore, when the distance between the remaining grids of the non-untravelable grids in the logistics area and the untravelable grids is less than or equal to the safety margin, the corresponding grid is also recorded as an untravelable grid. Among them, the distance between the grids is the distance between the center points of the corresponding grids. When the boundary of the logistics area is less than one grid, the boundary of the logistics area is directly defaulted to an untravelable area. In this embodiment, the safety margin is set to 30cm, and the implementer can set the size of the safety margin according to the actual situation, which is not limited here. At this time, all non-untravelable grids in the logistics area are used as walkable grids. So far, the grids in the logistics area are divided into two categories, namely, untravelable grids and walkable grids. It should be noted that the subsequent path planning of logistics vehicles is all carried out on the walkable grid, and the position corresponding to each walkable grid is the coordinate of its center point.

[0066] In order to accurately analyze the congestion situation of logistics vehicles during driving, this embodiment equips each logistics vehicle in the logistics area with multiple cameras, including front-view, rear-view and side-view cameras, to provide 360-degree visual coverage for each logistics vehicle. At the same time, combined with sensors such as lidar, each logistics vehicle can accurately perceive the surrounding environment, including obstruction information such as other logistics vehicles and pedestrians.

[0067] In order to accurately analyze the scheduling of logistics vehicles in the future, this embodiment obtains the average speed of each logistics vehicle in the historical time period as the speed of each logistics vehicle. Among them, the speed of each logistics vehicle includes the speed reduction caused by the congestion in the historical time period and the idleness of each logistics vehicle. This embodiment sets the historical time period to 30 days. The implementer can set the size of the historical time period according to the actual situation, which is not limited here.

[0068] Step S2: Take any idle logistics vehicle in the logistics area at the current moment as the target vehicle, and take the ratio of the path length from the position of the target vehicle at the previous moment to the designated location to the speed of the target vehicle as the reference time for the target vehicle to reach the designated location.

[0069] Specifically, for the sake of clear description, this embodiment takes any idle logistics vehicle in the logistics area at the current moment as the target vehicle and analyzes the target vehicle. It is known that the path length from the position of the target vehicle at the current moment to the designated location is obtained through the cost function in the A-star algorithm, and then this embodiment takes the ratio of the path length from the position of the target vehicle at the previous moment to the designated location to the speed of the target vehicle as the reference time for the target vehicle to reach the designated location, which is convenient for subsequent prediction of the position of non-idle logistics vehicles when the target vehicle arrives at the designated location, and is conducive to analyzing the congestion situation of the target vehicle arriving at the designated location.

[0070] Among them, the method of obtaining the path length from the current position of the target vehicle to the specified position through the cost function in the A-star algorithm is: It is known that the formula of the cost function in the A-star algorithm is: ;in, is the actual cost (path length) from the starting position to position x; is the estimated cost from location x to the specified location; is the cost from the starting position to the specified position. Assume that the position of the target vehicle at the current moment is , specify the location as , then the path length from the target vehicle's current position to the specified position is At this point, the path length from the target vehicle's current position to the specified position is obtained.

[0071] It should be noted that if the target vehicle's position at the current moment occupies at least two traversable grids, the average of the center point coordinates of all traversable grids occupied by the target vehicle's position at the current moment is taken as the target vehicle's position at the current moment.

[0072] Step S3: Based on the reference duration and the speed of each non-idle logistics vehicle in the corresponding time period, and the path length of each non-idle logistics vehicle from each position in the planned path corresponding to the reference duration to the designated position, predict the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated position.

[0073] In the existing methods, when a new task appears at a certain location in the logistics area, the task is usually assigned to the idle logistics vehicle closest to the location, without considering the congestion of the planned path of the idle logistics vehicle during the journey, that is, without considering the time it takes for the idle logistics vehicle to pass through different paths. Therefore, this embodiment analyzes the impact of other logistics vehicles on the idle logistics vehicle's arrival at the location where the task needs to be processed, and the impact of the time it takes, and corrects the cost function in the A-star algorithm, so as to accurately schedule the logistics vehicles.

[0074] In order to accurately predict the time it takes for the target vehicle to reach the designated location in the presence of congestion, this embodiment first directly obtains the non-idle logistics vehicles in the time period corresponding to the reference time through the path planned by the control center, and then predicts the position of each non-idle logistics vehicle when the target vehicle arrives at the designated location, and then analyzes the congestion impact of the non-idle logistics vehicles in the process of the target vehicle arriving at the designated location, which is conducive to accurately predicting the time it takes for the target vehicle to reach the designated location in the presence of other non-idle logistics vehicles. Therefore, this embodiment analyzes the travel distance of each non-idle logistics vehicle in the time period corresponding to the reference time according to the speed of each non-idle logistics vehicle in the reference time and its corresponding time period, and then obtains the path length of each non-idle logistics vehicle from each position in the planned path corresponding to the reference time to the designated location. When the path length of a non-idle logistics vehicle from a certain position in the planned path corresponding to the reference time to the designated location is equal to the travel distance of the non-idle logistics vehicle in the time period corresponding to the reference time, it means that when the target vehicle arrives at the designated location, the position of the non-idle logistics vehicle is more likely to be close to the target vehicle. Because this is a prediction, the analysis is performed based on the most likely situation of each non-idle logistics vehicle approaching the designated location during the time period corresponding to the reference duration. Furthermore, this embodiment predicts the target location of each non-idle logistics vehicle when the target vehicle arrives at the designated location based on the speed of each non-idle logistics vehicle during the reference duration and its corresponding time period, and the path length of each non-idle logistics vehicle from each location in the planned path corresponding to the reference duration to the designated location.

[0075] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated position is: for any non-idle logistics vehicle in the time period corresponding to the reference time length, obtain the product of the speed of the non-idle logistics vehicle and the reference time length as the reference travel distance of the non-idle logistics vehicle; the larger the reference travel distance, the larger the distance moved by the non-idle logistics vehicle in the time period corresponding to the reference time length. In order to predict whether the non-idle logistics vehicle will cause congestion to the target vehicle in the process of reaching the designated position, therefore, in this embodiment, each position passed by the planned path of the non-idle logistics vehicle in the time period corresponding to the reference time length is obtained as a reference position; the path length of the non-idle logistics vehicle from each reference position to the designated position is obtained as the reference path length; the absolute value of the difference between each reference path length and the reference travel distance is obtained as the first difference; the smaller the first difference, the closer the reference position corresponding to the reference path length is to the designated position, and then the reference position corresponding to the smallest first difference is predicted as the target position of the non-idle logistics vehicle when the target vehicle arrives at the designated position.

[0076] At this point, the target position of each non-idle logistics vehicle when the target vehicle arrives at the specified location is obtained.

[0077] Step S4: Based on the planned travel direction and position of the target vehicle at the current moment, and the target position and speed of each non-idle logistics vehicle, obtain the target predicted time for the target vehicle to reach the specified location.

[0078] Specifically, in order to accurately analyze the time it takes for a target vehicle to reach a designated location under congestion conditions, which is conducive to the subsequent accurate analysis of the cost of the target vehicle reaching the designated location, the present embodiment first obtains the planned direction and position of the target vehicle at the current moment, determines the direction of travel of the target vehicle to reach the designated location, and then analyzes the target position and speed of each non-idle logistics vehicle to determine the congestion impact of the non-idle logistics vehicles in the process of the target vehicle reaching the designated location. Furthermore, the present embodiment obtains the target predicted time it takes for the target vehicle to reach the designated location based on the planned direction and position of the target vehicle at the current moment, and the target position and speed of each non-idle logistics vehicle.

[0079] Preferably, in one possible implementation of this embodiment, the method for obtaining the target prediction duration is as follows: Figure 2 , which shows a flow chart of a method for obtaining a target prediction duration provided by this embodiment, the method comprising the following steps:

[0080] Step S201: According to the planned traveling direction and position of the target vehicle at the current moment, the traveling direction area of ​​the target vehicle at the current moment is obtained.

[0081] In actual situations, there may be a deviation between the direction of travel of the target vehicle at the current moment to reach the designated location and the planned direction of travel. At the same time, considering that when the target vehicle encounters obstruction from other logistics vehicles, in order for the target vehicle to continue moving toward the designated location, the actual direction of travel of the target vehicle may also deviate from the planned direction of travel. Therefore, this embodiment obtains the travel direction area of ​​the target vehicle at the current moment based on the planned direction of travel and position of the target vehicle at the current moment.

[0082] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the target vehicle's travel direction area at the current moment is: taking the previous position that the target vehicle passes through when reaching the position at the current moment as the historical reference position; taking the direction of the target vehicle from the historical reference position to the position at the current moment as the planned travel direction of the target vehicle at the current moment; obtaining the direction of the target vehicle's position at the current moment pointing to the center point of each walkable grid, all as the first direction; obtaining the distance between the target vehicle's position at the current moment and the center point of each walkable grid, all as the first distance; taking the area composed of the walkable grids corresponding to the first distance being less than the preset distance threshold and the angle between the first direction and the planned travel direction of the target vehicle at the current moment being less than the preset angle threshold as the travel direction area of ​​the target vehicle at the current moment. In this embodiment, the preset distance threshold is set to 500cm and the preset angle threshold is set to 60°. The implementer can set the preset distance threshold and the preset angle threshold according to the actual situation, which is not limited here.

[0083] Step S202: Filter out reference non-idle logistics vehicles according to the distance between the target position of each non-idle logistics vehicle in the time period corresponding to the reference duration and the travel direction area of ​​the target vehicle at the current moment.

[0084] The smaller the distance between the target position of a non-idle logistics vehicle in the time period corresponding to the reference duration and the travel direction area of ​​the target vehicle at the current moment, the more likely the non-idle logistics vehicle will cause congestion in the process of the target vehicle reaching the designated location, affecting the time it takes for the target vehicle to reach the designated location. Therefore, this embodiment selects reference non-idle logistics vehicles based on the distance between the target position of each non-idle logistics vehicle in the time period corresponding to the reference duration and the travel direction area of ​​the target vehicle at the current moment, and subsequently analyzes the congestion situation when the target vehicle arrives at the designated location based on the reference non-idle logistics vehicles.

[0085] Preferably, in a manner that can be implemented in this embodiment, the method for obtaining the reference non-idle logistics vehicles is: obtaining the minimum distance between the target position of each non-idle logistics vehicle in the time period corresponding to the reference duration and the traveling direction area of ​​the target vehicle at the current moment, as the target distance of each non-idle logistics vehicle; wherein, when the target position of the non-idle logistics vehicle is located in the traveling direction area of ​​the target vehicle at the current moment, the corresponding target distance is 0; when the number of non-idle logistics vehicles corresponding to the target distance of 0 is greater than or equal to a preset number threshold, the non-idle logistics vehicles corresponding to the target distance of 0 are all used as reference non-idle logistics vehicles; when the number of non-idle logistics vehicles corresponding to the target distance of 0 is less than the preset number threshold, the target distances are arranged from small to large to obtain a target distance sequence; the non-idle logistics vehicles corresponding to the target distances before the preset number threshold in the target distance sequence are all used as reference non-idle logistics vehicles.

[0086] Step S203: According to the position distribution and speed of each reference non-idle logistics vehicle, as well as the travel direction area and planned travel direction of the target vehicle at the current moment, obtain the target predicted time for the target vehicle to reach the specified location.

[0087] In order to predict the time it takes for a target vehicle to reach a designated location under congestion conditions, this embodiment analyzes the travel conditions of logistics vehicles under congestion conditions in a historical time period, and obtains the travel direction area of ​​each logistics vehicle at the congestion moment through step S201 according to the planned travel direction and position of each logistics vehicle at the congestion moment in the historical time period; the total number of walkable grids occupied by non-idle logistics vehicles in the travel direction area of ​​each logistics vehicle at the congestion moment is obtained through the end-side big model, as the first number of each logistics vehicle at the congestion moment; wherein, the end-side big model is a well-known technology and will not be described in detail. The ratio of the first number of each logistics vehicle at the time of congestion to the number of all traversable grids in its traveling direction area is used as the occlusion density of each logistics vehicle at the time of congestion; the occlusion density and the planned traveling direction of each logistics vehicle at the time of congestion, as well as the speed of non-idle logistics vehicles in the traveling direction area of ​​each logistics vehicle at the time of congestion are used as inputs, and each logistics vehicle uses the duration of the traveling direction area at the time of congestion as output to train the LSTM (Long Short-Term Memory network) network to obtain a trained LSTM network. It should be noted that since the number of non-idle logistics vehicles in the traveling direction area of ​​different logistics vehicles is different, this embodiment uses the padding technology to fill the speed of non-idle logistics vehicles in the traveling direction area of ​​each logistics vehicle at the time of congestion with zero, so that the number of non-idle logistics vehicles in the traveling direction area of ​​each logistics vehicle at the time of congestion is the same, that is, they are all consistent with the largest number of non-idle logistics vehicles in the traveling direction area. Therefore, this embodiment sets the preset number threshold in step S202 to the largest number of non-idle logistics vehicles in the traveling direction area of ​​logistics vehicles at the time of congestion in the historical time period. At the same time, the padding part is ignored through the masking operation when calculating the loss. Among them, the padding technology, the masking operation and the LSTM network are all well-known technologies and will not be described in detail. At this point, the trained LSTM network is obtained to accurately predict the target prediction time for the target vehicle to reach the specified location in the subsequent process.

[0088] In order to obtain the target predicted time for the target vehicle to reach the specified location through the trained LSTM network, this embodiment needs to predict the density of obstructions of the target vehicle at the current moment, and then analyze the position distribution of each reference non-idle logistics vehicle in step S202. When a reference non-idle logistics vehicle is located in the travel direction area of ​​the target vehicle at the current moment, it means that the reference non-idle logistics vehicle is more meaningful in analyzing the density of obstructions of the target vehicle at the current moment; the farther the reference non-idle logistics vehicle is from the travel direction area of ​​the target vehicle at the current moment, the more meaningful the reference non-idle logistics vehicle is in analyzing the density of obstructions of the target vehicle at the current moment. The smaller the reference value of the obstruction density at the previous moment, the smaller the reference value of the target vehicle at the current moment. Therefore, in this embodiment, the target distance of each reference non-idle logistics vehicle is negatively correlated and normalized as the reference weight of each reference non-idle logistics vehicle. The larger the reference weight, the greater the reference value of the corresponding reference non-idle logistics vehicle in predicting the obstruction density of the target vehicle at the current moment. Then, according to the reference weight of each reference non-idle logistics vehicle and the number of walkable grids occupied by each reference non-idle logistics vehicle at its target position, and the number of walkable grids in the travel direction area of ​​the target vehicle at the current moment, the obstruction density of the target vehicle at the current moment is obtained.

[0089] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the occlusion density of the target vehicle at the current moment is: taking the product of the reference weight of each reference non-idle logistics vehicle and the number of walkable grids it occupies at the target position as the reference number of walkable grids occupied by each reference non-idle logistics vehicle; taking the ratio of the cumulative result of the reference number of walkable grids occupied by all reference non-idle logistics vehicles and the number of walkable grids in the traveling direction area of ​​the target vehicle at the current moment as the occlusion density of the target vehicle at the current moment.

[0090] The calculation formula for obtaining the density of obstructions of the target vehicle at the current moment is: ; In the formula, is the occlusion density of the target vehicle at the current moment; R is the number of reference non-idle logistics vehicles corresponding to the target vehicle at the current moment; is the target distance of the rth reference non-idle logistics vehicle; is the number of traversable grids occupied by the rth reference non-idle logistics vehicle; is the reference weight of the rth reference non-idle logistics vehicle; N is the number of grids that the target vehicle can travel in the moving direction area at the current moment; is the number of walkable grids occupied by the rth reference non-idle logistics vehicle; norm is the normalization function.

[0091] The speed of each reference non-idle logistics vehicle, the occlusion density of the target vehicle at the current moment, and the planned travel direction are taken as input into the trained LSTM network, and the target predicted time for the target vehicle to arrive at the specified location is output.

[0092] At this point, the target predicted time for the target vehicle to reach the specified location is obtained.

[0093] Step S5: According to the speed of the target vehicle and the target prediction time, the cost function of the target vehicle reaching the specified position in the A-star algorithm is corrected to obtain a corrected cost function.

[0094] In order to accurately analyze the cost of the target vehicle reaching the designated location, this embodiment first obtains the congestion cost of the target vehicle reaching the designated location according to the speed of the target vehicle and the target prediction time, and then corrects the cost function of the target vehicle reaching the designated location in the A-star algorithm to obtain the corrected cost function of the target vehicle reaching the designated location, and accurately obtains the cost of the target vehicle reaching the designated location. Furthermore, this embodiment corrects the cost function of the target vehicle reaching the designated location in the A-star algorithm according to the speed of the target vehicle and the target prediction time to obtain the corrected cost function.

[0095] Preferably, in a method that can be implemented in this embodiment, the method for obtaining the corrected cost function is: taking the product of the speed of the target vehicle and the target predicted duration as the congestion cost for the target vehicle to arrive at the designated location; and taking the sum of the cost function for the target vehicle to arrive at the designated location and the congestion cost as the corrected cost function for the target vehicle to arrive at the designated location.

[0096] At this point, the corrected cost function for each idle logistics vehicle to reach the designated location at the current moment is obtained.

[0097] Step S6: Based on the modified cost function, the logistics vehicles of the visual perception industrial vehicle ADAS system are scheduled.

[0098] Based on the modified cost function, the cost function size of each idle logistics vehicle reaching the designated location at the current moment is recalculated, and then the improved path length of each idle logistics vehicle reaching the designated location can be obtained, and finally the idle logistics vehicle corresponding to the modified minimum cost function is selected and assigned to the designated location. Since the path is obtained based on the modified cost function, the path of multiple logistics vehicles will avoid congestion as much as possible, which improves the accuracy of logistics vehicle scheduling, reduces congestion and path conflicts, improves the efficiency of material handling, and significantly optimizes the overall operation performance of the industrial logistics system.

[0099] In summary, this embodiment obtains the speed of the logistics vehicle, and then obtains the reference time for the target vehicle to reach the designated location at the previous moment; predicts the target position of the non-idle logistics vehicle according to the speed of the non-idle logistics vehicle and the path length of its position in the planned path corresponding to the reference time to reach the designated location; based on the planned direction and position of the target vehicle at the current moment, the target position and speed of the non-idle logistics vehicle, obtains the target prediction time, corrects the cost function of the target vehicle reaching the designated location in the A-star algorithm, and obtains the corrected cost function to schedule the logistics vehicle. The present invention reduces the congestion and path conflict of logistics vehicles by correcting the cost function in the A-star algorithm, and improves the accuracy and efficiency of logistics vehicle scheduling.

[0100] Embodiment 2:

[0101] The present invention also proposes a logistics vehicle dispatching system based on a visual perception industrial vehicle ADAS system, please refer to Figure 3 , which shows a structural diagram of a logistics vehicle scheduling system based on a visual perception industrial vehicle ADAS system provided by an embodiment of the present invention. The system includes: a data acquisition module 10, a reference duration acquisition module 20, a target position acquisition module 30, a target prediction duration acquisition module 40, a modified cost function acquisition module 50 and a scheduling module 60.

[0102] The data acquisition module 10 is used to acquire the speed of each logistics vehicle in the logistics area based on the visual perception industrial vehicle ADAS system.

[0103] The reference duration acquisition module 20 is used to take any idle logistics vehicle in the logistics area at the current moment as the target vehicle, and take the ratio of the path length from the position of the target vehicle at the previous moment to the designated position to the speed of the target vehicle as the reference duration for the target vehicle to reach the designated position.

[0104] The target position acquisition module 30 is used to predict the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated position based on the speed of each non-idle logistics vehicle in the reference time and its corresponding time period, and the path length of each non-idle logistics vehicle from each position in the planned path corresponding to the reference time to the designated position.

[0105] The target prediction duration acquisition module 40 is used to obtain the target prediction duration for the target vehicle to reach the designated location based on the planned travel direction and position of the target vehicle at the current moment, and the target position and speed of each non-idle logistics vehicle.

[0106] The modified cost function acquisition module 50 is used to modify the cost function of the target vehicle reaching the specified position in the A-star algorithm according to the speed of the target vehicle and the target prediction time, and obtain the modified cost function.

[0107] The scheduling module 60 is used to schedule the logistics vehicles of the visual perception industrial vehicle ADAS system based on the modified cost function.

[0108] It should be noted that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the logistics vehicle scheduling system based on the visual perception industrial vehicle ADAS system and the logistics vehicle scheduling method embodiment based on the visual perception industrial vehicle ADAS system provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0109] Embodiment 3:

[0110] The present invention also proposes a logistics vehicle scheduling device based on a visual perception industrial vehicle ADAS system, the device includes a memory and a processor, wherein the memory stores an executable program code, and the processor is used to call and execute the executable program code to execute a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided in an embodiment of the present application. The device can specifically be a chip, a component or a module, and the chip may include a connected processor and a memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided in the above embodiment.

[0111] In addition, the present application embodiment also protects a computer device, see Figure 4 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402, wherein when the processor 402 executes the computer program 403, the computer device can execute any of the logistics vehicle scheduling methods based on the visual perception industrial vehicle ADAS system introduced above.

[0112] Embodiment 4:

[0113] This embodiment also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes the above-mentioned related method steps to implement a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided in the above embodiment.

[0114] Embodiment 5:

[0115] This embodiment also provides a computer program product. When the computer program product runs on a computer, it enables the computer to execute the above-mentioned related steps to implement a logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system provided by the above embodiment.

[0116] Among them, the device, computer-readable storage medium, computer program product or chip provided in this embodiment is used to execute the corresponding method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here.

[0117] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0118] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system, characterized in that: The method comprises the following steps: Obtain the speed of each logistics vehicle in the logistics area based on the visual perception industrial vehicle ADAS system; Any idle logistics vehicle in the logistics area at the current moment is taken as the target vehicle, and the ratio of the path length from the position of the target vehicle at the previous moment to the designated location to the speed of the target vehicle is taken as the reference time for the target vehicle to reach the designated location; According to the reference duration and the speed of each non-idle logistics vehicle in the corresponding time period, and the path length of each non-idle logistics vehicle from each position in the planned path corresponding to the reference duration to the designated location, the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated location is predicted; Based on the planned travel direction and position of the target vehicle at the current moment, and the target position and speed of each non-idle logistics vehicle, obtain the target predicted time for the target vehicle to reach the specified location; According to the speed of the target vehicle and the target prediction time, the cost function of the target vehicle reaching the specified position in the A-star algorithm is corrected to obtain a corrected cost function; Based on the modified cost function, the logistics vehicles are dispatched for the ADAS system of industrial vehicles based on visual perception; The method for obtaining the target predicted time for the target vehicle to arrive at the specified location is: According to the planned moving direction and position of the target vehicle at the current moment, the moving direction area of ​​the target vehicle at the current moment is obtained; According to the distance between the target position of each non-idle logistics vehicle in the time period corresponding to the reference duration and the travel direction area of ​​the target vehicle at the current moment, the reference non-idle logistics vehicles are selected; According to the location distribution and speed of each reference non-idle logistics vehicle, as well as the travel direction area and planned travel direction of the target vehicle at the current moment, the target predicted time for the target vehicle to reach the specified location is obtained; The method for obtaining the modified cost function is: The product of the target vehicle's speed and the target predicted time is taken as the congestion cost for the target vehicle to reach the designated location; The cost function for the target vehicle to reach the designated location and the congestion cost are added together to obtain the corrected cost function for the target vehicle to reach the designated location.

2. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system as claimed in claim 1, characterized in that: The method for obtaining the path length from the position of the target vehicle at the previous moment to the specified position is: The cost function in the A-star algorithm is used to obtain the path length from the target vehicle's position at the previous moment to the specified position.

3. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system as claimed in claim 1, characterized in that: The method for predicting the target position of each non-idle logistics vehicle when the target vehicle arrives at the designated location is: For any non-idle logistics vehicle in the time period corresponding to the reference duration, obtain the product of the speed of the non-idle logistics vehicle and the reference duration as the reference travel distance of the non-idle logistics vehicle; Obtain each location that the non-idle logistics vehicle passes through during the planned path within the time period corresponding to the reference duration, and use them as reference locations; Obtain the path length of the non-idle logistics vehicle from each reference location to the designated location, which is used as the reference path length; Obtaining a difference between each reference path length and the reference travel distance as a first difference; The reference position corresponding to the smallest first difference is predicted as the target position of the non-idle logistics vehicle when the target vehicle arrives at the designated position.

4. The logistics vehicle scheduling method based on the visual perception industrial vehicle ADAS system according to claim 1, characterized in that: The method for obtaining the moving direction area of ​​the target vehicle at the current moment is: The previous position that the target vehicle passed before reaching the current position is used as the historical reference position; The direction of the target vehicle from the historical reference position to the current position is used as the planned travel direction of the target vehicle at the current moment; Obtain the traversable grids in the logistics area, and point the target vehicle's current position to the direction of the center point of each traversable grid, all of which are taken as the first direction; Obtain the distance between the current position of the target vehicle and the center point of each traversable grid, all as the first distance; The area formed by the walkable grids corresponding to the first distance being less than the preset distance threshold and the angle between the first direction and the planned traveling direction of the target vehicle at the current moment being less than the preset angle threshold is taken as the traveling direction area of ​​the target vehicle at the current moment.

5. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system as claimed in claim 4, characterized in that: The method for obtaining the reference non-idle logistics vehicle is: Obtain the minimum distance between the target position of each non-idle logistics vehicle and the target vehicle's current traveling direction area within the time period corresponding to the reference duration as the target distance of each non-idle logistics vehicle; wherein, when the target position of the non-idle logistics vehicle is within the target vehicle's current traveling direction area, the corresponding target distance is 0; When the number of non-idle logistics vehicles corresponding to the target distance of 0 is greater than or equal to the preset number threshold, all non-idle logistics vehicles corresponding to the target distance of 0 are used as reference non-idle logistics vehicles; When the number of non-idle logistics vehicles corresponding to the target distance of 0 is less than the preset number threshold, the target distances are arranged from small to large to obtain a target distance sequence; The non-idle logistics vehicles corresponding to the target distances before a preset number threshold in the target distance sequence are all used as reference non-idle logistics vehicles.

6. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system as claimed in claim 5, characterized in that: The method for obtaining the target predicted time for the target vehicle to arrive at the specified location is: The result of negatively correlating and normalizing the target distance of each reference non-idle logistics vehicle is used as the reference weight of each reference non-idle logistics vehicle; Obtain the occlusion density of the target vehicle at the current moment according to the reference weight of each reference non-idle logistics vehicle and the number of walkable grids it occupies at the target position, as well as the number of walkable grids in the travel direction area of ​​the target vehicle at the current moment; The speed of each reference non-idle logistics vehicle, the obstruction density of the target vehicle at the current moment, and the planned travel direction are taken as input, and the target predicted time for the target vehicle to reach the specified location is output through the trained LSTM network.

7. A logistics vehicle scheduling method based on a visual perception industrial vehicle ADAS system as claimed in claim 6, characterized in that: The method for obtaining the blocking object density of the target vehicle at the current moment is: The product of the reference weight of each reference non-idle logistics vehicle and the number of traversable grids occupied by it at the target location is taken as the reference number of traversable grids occupied by each reference non-idle logistics vehicle; The ratio of the cumulative number of reference occupied walkable grids of all reference non-idle logistics vehicles to the number of walkable grids in the traveling direction area of ​​the target vehicle at the current moment is taken as the occlusion density of the target vehicle at the current moment.

8. The method for dispatching logistics vehicles based on a visual perception industrial vehicle ADAS system as claimed in claim 6, characterized in that: The method for obtaining the trained LSTM network is: According to the planned travel direction and position of each logistics vehicle at the congestion moment in the historical time period, the travel direction area of ​​each logistics vehicle at the congestion moment is obtained; The total number of traversable grids occupied by non-idle logistics vehicles in the travel direction area of ​​each logistics vehicle at the congested moment is obtained through the end-side large model as the first number of each logistics vehicle at the congested moment; The ratio of the first number of each logistics vehicle at the congestion moment to the number of all walkable grids in the area in its travel direction is used as the blocking density of each logistics vehicle at the congestion moment; The density of obstructions and the planned travel direction of each logistics vehicle at the time of congestion, as well as the speed of non-idle logistics vehicles in the travel direction area of ​​each logistics vehicle at the time of congestion are taken as input, and the duration of each logistics vehicle in the travel direction area at the time of congestion is used as output to train the LSTM network and obtain a trained LSTM network.

Images