A parking scheduling method, electronic device, medium and product

Abstract

The application discloses a parking scheduling method, an electronic device, a medium and a product. The method comprises the following steps: in response to a parking request, obtaining vehicle information of a target vehicle; determining a target buffer parking space from each buffer parking space based on the vehicle information; determining an allocation priority of a candidate stacking parking space based on multi-dimensional stacking parking space allocation factors, wherein the candidate stacking parking space is a stacking parking space in a target parking area, the target parking area is a parking area to which the target buffer parking space belongs, and the multi-dimensional stacking parking space allocation factors comprise an inside-outside parking space factor, which is used for adjusting the allocation priority of the candidate stacking parking space based on a time relationship between a first estimated departure time of the target vehicle and a second estimated departure time of a vehicle parked in an inside parking space; and determining a target stacking parking space from the candidate stacking parking spaces based on the allocation priority, wherein the target stacking parking space is used for parking the target vehicle. The application can improve the efficiency of garage scheduling.

Description

Technical Field

[0001] This application relates to the field of parking scheduling technology, specifically to parking scheduling methods, electronic equipment, media, and products for multi-level parking garages. Background Technology

[0002] Currently, automated parking systems, as a core solution for urban parking, are widely used in hospitals, corporate parks, commercial complexes, and residential communities. Their core scheduling system is responsible for making decisions regarding vehicle allocation and retrieval within the limited space of the automated parking system. Existing automated parking system scheduling technologies primarily employ fixed rules for parking space allocation, such as "allocation based on proximity" or "first-come, first-served." However, if the allocated stacked parking spaces are not properly configured, it will require multiple relocations of vehicles already parked, impacting the efficiency of the parking system.

[0003] Therefore, how to improve the efficiency of garage scheduling is a technical problem that needs to be solved by those skilled in the art. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide a parking scheduling method, electronic device, medium, and product that can improve parking garage scheduling efficiency. The specific solution is as follows: Firstly, this application provides a parking dispatching method, including: In response to a parking request, obtain the vehicle information of the target vehicle; Based on the vehicle information, the target buffer parking space is determined from each buffer parking space; The allocation priority of candidate stacking parking spaces is determined based on multi-dimensional stacking parking space allocation factors. The candidate stacking parking spaces are stacking parking spaces within the target parking area, and the target parking area is the parking area to which the target buffer parking space belongs. The multi-dimensional stacking parking space allocation factors include inner and outer parking space factors, which are used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first expected departure time of the target vehicle and the second expected departure time of the vehicle already parked in the inner parking space. A target stacking parking space is determined from the candidate stacking parking spaces based on the allocation priority, wherein the target stacking parking space is used to park the target vehicle.

[0005] Optionally, the allocation priority of candidate stacking parking spaces can be determined based on multi-dimensional stacking parking space allocation factors, including: Based on the current parking scenario attributes, determine the target stacking parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacking parking space allocation factors; Under the target allocation factors, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters.

[0006] Optionally, when the target allocation factor is the inner and outer parking space factor, the target stacking parking space allocation parameter is used to adjust the allocation priority between the first type of candidate stacking parking space and the second type of candidate stacking parking space based on the first expected departure time of the target vehicle. The first type of candidate stacking parking space is the inner parking space when both the inner and outer parking spaces are vacant; the second type of candidate stacking parking space is the outer parking space when there is a parked vehicle in the inner parking space and the outer parking space is vacant.

[0007] Optionally, the target stacking parking space allocation parameters include: a first basic value corresponding to a first type of candidate stacking parking space, a second basic value corresponding to a second type of candidate stacking parking space, a third basic value corresponding to a third type of candidate stacking parking space, a preset outer parking weight coefficient, and a time tolerance; the third type of candidate stacking parking space is an inner parking space when the inner parking space is empty and the outer parking space is occupied by a parked vehicle. Under the target allocation factor, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters, including: When the second type of candidate stacking parking space meets the preset time tolerance condition, the weighted value of the second type of candidate stacking parking space is determined based on the second basic value and the preset outer parking weight coefficient; wherein, the preset time tolerance condition is that the first expected departure time of the target vehicle is earlier than the second expected departure time of the vehicle already parked in the inner parking space, and the time difference between the first expected departure time and the second expected departure time is greater than the time tolerance. The allocation priority of candidate stacking parking spaces is determined based on the first basic value, the weighted value, and the third basic value.

[0008] Optionally, the method further includes: The departure time of the target vehicle is predicted based on multiple departure time prediction models; The first estimated departure time of the target vehicle is obtained by weighting and calculating multiple departure times.

[0009] Optionally, when the target allocation factor is a distance factor, the target stacking parking space allocation parameters include: coefficients of each dimension coordinate; Under the target allocation factor, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters, including: The distance value is obtained by weighting the coefficients of each dimension of the coordinates and the differences between the coordinates of each dimension of the candidate stacking parking space and the coordinates of each dimension of the target buffer parking space. The allocation priority of candidate stacking parking spaces is determined based on the distance value.

[0010] Optionally, the method further includes: When the parameter optimization conditions are met, parking scheduling is simulated in the simulation environment based on the garage's operating data and the current target stacking parking space allocation parameters, and cost data is obtained. Based on the cost data and using a large model, the target stacking parking space allocation parameters corresponding to each parking scenario attribute are optimized to obtain the optimized parameters; The optimized parameters are determined as the current target stacking parking space allocation parameters, and the steps of triggering parking scheduling based on garage operation data and using the current target stacking parking space allocation parameters in the simulation environment, and obtaining cost data are initiated. When the optimization termination condition is met, the current optimized parameters are cached. Accordingly, based on the current parking scenario attributes, the target stacking parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacking parking space allocation factors are determined, including: The target stacked parking space allocation parameters are determined from the optimized parameters in the cache, which are the target allocation factors among the multi-dimensional stacked parking space allocation factors under the current parking scenario attributes.

[0011] Optionally, the cost data includes energy costs, equipment costs, and time costs, and the method further includes: The comprehensive cost is determined based on the energy consumption cost, the equipment cost, and the time cost. When the overall cost converges, the optimization termination condition is determined to be met.

[0012] Optionally, the allocation priority of candidate stacking parking spaces can be determined based on multi-dimensional stacking parking space allocation factors, including: Based on the multidimensional stacking parking space allocation factors, the candidate stacking parking spaces are sorted sequentially according to a preset dimension order. The subsequent stacking allocation factors sort the candidate stacking parking spaces with the same order in the sorting results of the previous stacking allocation factors to obtain the allocation priority of the candidate stacking parking spaces.

[0013] Optionally, determining a target buffer parking space from each buffer parking space based on the vehicle information includes: determining a set of parking areas in the garage based on the vehicle information, wherein each parking area includes buffer parking spaces and stacking parking spaces; and determining a target buffer parking space from each buffer parking space in the set of parking areas.

[0014] Optionally, determining the target buffer parking space from each buffer parking space in the set of available parking areas includes: The target buffer parking space is determined from each buffer parking space in the set of available parking areas based on the buffer parking space selection factors. The buffer parking space selection factors include one or more of the following: buffer parking space selection parameters, area selection parameters, area adaptation factors, area location factors, buffer parking space status factors, and buffer parking space location factors.

[0015] Secondly, this application provides a parking dispatching device, comprising: The vehicle information acquisition module is used to acquire the vehicle information of the target vehicle in response to a parking request. A buffer parking space determination module is used to determine a target buffer parking space from each buffer parking space based on the vehicle information; The priority determination module is used to determine the allocation priority of candidate stacking parking spaces based on multi-dimensional stacking parking space allocation factors. The candidate stacking parking spaces are stacking parking spaces in the target parking area, and the target parking area is the parking area to which the target buffer parking space belongs. The multi-dimensional stacking parking space allocation factors include inner and outer parking space factors. The module is used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first expected departure time of the target vehicle and the second expected departure time of the vehicle already parked in the inner parking space. The target vehicle determination module is used to determine a target stacking parking space from the candidate stacking parking spaces based on the allocation priority, wherein the target stacking parking space is used to park the target vehicle.

[0016] Thirdly, this application provides an electronic device, including a memory and a processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program to implement the aforementioned parking scheduling method.

[0017] Fourthly, this application provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned parking scheduling method.

[0018] Fifthly, this application provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the aforementioned parking scheduling method.

[0019] As can be seen from the above scheme, this application provides a parking scheduling method, including: in response to a parking request, obtaining vehicle information of a target vehicle; determining a target buffer parking space from each buffer parking space based on the vehicle information; determining the allocation priority of candidate stacked parking spaces based on multi-dimensional stacked parking space allocation factors, wherein the candidate stacked parking spaces are stacked parking spaces within a target parking area, the target parking area is the parking area to which the target buffer parking space belongs, and the multi-dimensional stacked parking space allocation factors include inner and outer parking space factors, used to adjust the allocation priority of candidate stacked parking spaces based on the time relationship between the first estimated departure time of the target vehicle and the second estimated departure time of vehicles already parked in the inner parking spaces; determining a target stacked parking space from the candidate stacked parking spaces based on the allocation priority, wherein the target stacked parking space is used to park the target vehicle.

[0020] As can be seen, the beneficial effects of this application are as follows: After obtaining the vehicle information of the target vehicle, the target buffer parking space is determined from each buffer parking space based on the vehicle information. Then, based on multi-dimensional stacking parking space allocation factors, the allocation priority of candidate stacking parking spaces is determined within the target parking area. Based on the allocation priority, the target stacking parking space is determined from the candidate stacking parking spaces. The target parking area is the parking area to which the target buffer parking space belongs. The multi-dimensional stacking parking space allocation factors include inner and outer parking space factors, which are used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first estimated departure time of the target vehicle and the second estimated departure time of vehicles already parked in the inner parking space. In this way, it is ensured that short-parking vehicles are preferentially allocated to outer parking spaces, and long-parking vehicles are preferentially allocated to inner parking spaces. This avoids the extra handling operation required to move long-parking vehicles on the outer side because short-parking vehicles are blocked on the inner side, thereby effectively reducing the number of vehicle handling operations, shortening the waiting time for retrieving the vehicle, and improving the efficiency of garage scheduling.

[0021] Accordingly, the parking dispatching device, equipment, readable storage medium and product provided in this application also have the above-mentioned technical effects. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0023] Figure 1 A flowchart of a parking scheduling method provided in an embodiment of this application; Figure 2 A flowchart illustrating a specific parking scheduling method provided in this application embodiment; Figure 3 A parameter optimization flowchart provided for an embodiment of this application; Figure 4 A partial schematic diagram of a garage provided as an example of this application; Figure 5 An application architecture diagram of a parking dispatching scheme provided in this application example; Figure 6 This is a schematic diagram of a parking dispatching device provided in an embodiment of this application; Figure 7 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0024] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] Multi-level parking garages (mechanical parking systems), as a core solution to urban parking problems, have been widely applied in hospitals, corporate parks, commercial complexes, residential communities, and other scenarios. Their core scheduling system is responsible for making decisions regarding vehicle allocation, retrieval scheduling, and parking space organization (vehicle relocation) within a limited three-dimensional space, directly impacting garage operating efficiency, energy consumption, and user experience. This application provides a parking scheduling solution that can accurately and efficiently dispatch target vehicles to suitable multi-level parking spaces, reducing scheduling costs and improving user experience.

[0026] See Figure 1 As shown, this application provides a parking scheduling method, including: in response to a parking request, obtaining vehicle information of a target vehicle; determining a target buffer parking space from various buffer parking spaces based on the vehicle information; determining the allocation priority of candidate stacked parking spaces based on multi-dimensional stacked parking space allocation factors, wherein the candidate stacked parking spaces are stacked parking spaces within a target parking area, the target parking area is the parking area to which the target buffer parking space belongs, and the multi-dimensional stacked parking space allocation factors include inner and outer parking space factors, used to adjust the allocation priority of candidate stacked parking spaces based on the time relationship between the first estimated departure time of the target vehicle and the second estimated departure time of vehicles already parked in the inner parking spaces; determining a target stacked parking space from the candidate stacked parking spaces based on the allocation priority, wherein the target stacked parking space is used to park the target vehicle.

[0027] Specifically, step S11: In response to a parking request, obtain the vehicle information of the target vehicle. This vehicle information may include vehicle identification, vehicle model, vehicle dimensions, etc. The vehicle identification may be the vehicle's VIN code, license plate number, etc. Vehicle dimensions can be obtained by capturing images with a camera and recognizing them, or by measuring them with a laser rangefinder, or by obtaining pre-stored vehicle dimensions based on the vehicle identification. In an optional embodiment, permission verification can be performed on the target vehicle based on the vehicle identification. If the permission verification fails, a prompt message is generated to remind the user to obtain the corresponding permissions. If the permission verification passes, subsequent steps are executed. The permission verification method can be a whitelist approach, whereby users can register a vehicle identification to be added to the whitelist and obtain parking permissions in the garage.

[0028] Regarding step S12: Based on the vehicle information, a target buffer parking space is determined from each buffer parking space. In an optional embodiment, a set of available parking areas in the garage can be determined based on the vehicle information, wherein each available parking area includes buffer parking spaces and stacking parking spaces; the target buffer parking space is then determined from each buffer parking space in the set of available parking areas. See also Figure 2 As shown, Figure 2 This is a flowchart illustrating a specific parking scheduling method provided in an embodiment of this application. By first filtering available parking areas based on vehicle information, transforming a global area traversal into a local directional search, and then determining the target available parking area using the target buffer parking space, the selection range of stacked parking spaces is precisely limited, thus improving the efficiency of parking space search.

[0029] The parking garage can be a multi-level parking system, comprising multiple parking areas. Each parking area includes buffer parking spaces and stacking parking spaces. Vehicles can first park in the buffer parking spaces through the entrance gate, and then be moved from the buffer parking spaces to the stacking parking spaces by a transport robot. The stacking parking spaces are the spaces on the multi-level vehicle frame, which can have multiple layers, each layer including multiple stacking parking spaces. Optional charging equipment can be installed in the stacking parking spaces, such as bottom-mounted charging devices. The parking area refers to the parking area that meets the requirements of the vehicle information.

[0030] This application embodiment can determine a set of available parking areas based on vehicle information and parking area information. The target vehicle's attribute information can be determined based on its identifier, including whether it is a VIP user, whether it has a dedicated parking space, and the identifier of that dedicated parking space. Area information can include whether the area is disabled, whether there are vacant stacked parking spaces within the area, the number of vacant stacked parking spaces, and whether there are charging facilities in the vacant stacked parking spaces. Additionally, the target vehicle's charging needs can be obtained. Based on the vehicle information, charging needs, and area information, suitable parking areas are determined, resulting in a set of available parking areas.

[0031] In optional embodiments, the area information may further include whether it is used for traffic diversion, and may also include the time period for traffic diversion. Parking areas can be marked as diversion areas. For example, different parking areas in a garage may have different numbers of stacked parking spaces. Areas with fewer stacked parking spaces can be configured as diversion areas. When the parking flow is low, they are not directly opened to the public. When the traffic flow is high, they are opened according to a strategy, such as releasing one diversion area when the number of consecutive queues at the entrance gate exceeds N, where N is a positive integer. During the diversion period, such as 07:50-09:10, all diversion areas are released. The diversion period can be a preset peak period, an identified peak period, or when the number of consecutive queues at the gate exceeds N during the diversion period, all diversion areas are released. Since in actual use, areas with fewer parking spaces may fill up first, reducing the number of available buffer parking spaces and ultimately affecting overall efficiency, this application embodiment ensures scheduling efficiency by restricting the use of certain areas.

[0032] In an optional embodiment, prohibited areas, debugging areas, vehicle model mismatch areas, areas where charging demand is not met, and areas that are full can be filtered out first. If there is a designated parking area or a designated buffer parking space, the corresponding parking area is determined as a parking area. If the number of consecutive queues at the gate is less than or equal to N, the diversion area is filtered out, or the diversion area is filtered out during periods when there is no diversion.

[0033] Understandably, each parking area includes buffer parking spaces, and once the target buffer parking space is determined, the target parking area can be identified.

[0034] In an optional embodiment, determining a target buffer parking space from each buffer parking space in the set of available parking areas may include: determining a target buffer parking space from each buffer parking space in the set of available parking areas based on buffer parking space selection factors; wherein, the buffer parking space selection factors include one or more of the following: buffer parking space selection parameters, area selection parameters, area adaptation factors, area location factors, buffer parking space status factors, and buffer parking space location factors.

[0035] The buffer parking space selection parameter is the identification information of the specified target buffer parking space. In response to the identification information of the target buffer parking space, the buffer parking space is preferentially determined as the target buffer parking space. The buffer parking space selection parameter can be carried in the storage request, specified by the user, or determined based on parking scenario attributes, etc. The area selection parameter is the identification information of the specified parking area. In response to the identification information of the specified parking area, the buffer parking spaces belonging to that area are determined as target buffer parking spaces, or are given priority over buffer parking spaces in other areas. This parameter can be carried in the storage request, specified by the user, or determined based on parking scenario attributes, etc. The area adaptation factor can refer to the degree of matching between the target vehicle and the area where the candidate buffer parking space is located. This includes preferentially selecting areas with charging stacking parking spaces when the target vehicle needs to be charged, preferentially selecting areas with a high proportion of ordinary stacking parking spaces when the target vehicle does not need to be charged to avoid occupying charging resources, preferentially selecting areas with a large number of remaining stacking parking spaces to balance the load of each area, and preferentially selecting areas where the real-time operating status of the equipment is normal. The equipment may include handling robots, etc. The location factor refers to the spatial relationship between the current entrance location marker of the target vehicle and the entrance location marker of the buffer parking space area. Buffer parking spaces in areas with the closest proximity or located on the same side are prioritized. For example, if a garage has a south and north side, entering from the north entrance makes it closer to the north area, thus shortening the path length from the entrance to the buffer parking space. The buffer parking space status factor refers to the current working status of the buffer parking space. Priority is given to idle status over attitude-adjusted status, attitude-adjusted status over driven-in status, and driven-in status over newly created status, to select buffer parking spaces that are available more quickly. "Newly created" refers to the status of creating a new task to enter the target buffer parking space; "driven-in" indicates the vehicle has entered the buffer parking space, after which attitude adjustment is performed. The buffer parking space location factor can also refer to the location type of the buffer parking space, prioritized in the order of safe zone over critical zone. A safe zone is an area with no collision risk for transport equipment, while a critical zone is the boundary area between adjacent parking areas where collisions with transport equipment may occur. Buffer parking space selection factors can affect scheduling costs.

[0036] In one implementation, candidate buffer parking spaces can be sorted sequentially according to a preset dimensional order based on multidimensional buffer parking space selection factors. Subsequent factors then sort candidate buffer parking spaces with the same ranking among those selected by earlier factors to obtain the optimal buffer parking space as the target buffer parking space. For example, the target buffer parking space can be determined step-by-step according to the following order: buffer parking space selection parameters, area selection parameters, area adaptation factors, area location factors, buffer parking space status factors, and buffer parking space location factors. If multiple highest-priority buffer parking spaces still exist after considering all factors (i.e., multiple highest-priority buffer parking spaces have the same ranking result), then a random selection is made from these multiple highest-priority buffer parking spaces. Random selection can spread wear and tear and prevent a single buffer parking space from being continuously hit.

[0037] In another implementation, the representation values ​​of each buffer parking space selection factor can be weighted based on weighting coefficients to obtain a weighted buffer parking space selection result; the target buffer parking space is then determined based on the weighted buffer parking space selection result. Further, the weighting coefficients corresponding to each buffer parking space selection factor can be determined based on the current parking scenario attributes; the representation values ​​of each buffer parking space selection factor can be weighted based on the weighting coefficients to obtain a weighted buffer parking space selection result; the target buffer parking space is then determined based on the weighted buffer parking space selection result. Parking scenario attributes can include a combination of one or more of the following: garage location, garage equipment type, and current time period type. Garage locations include hospitals, shopping malls, industrial parks, and unmanned vehicle parking lots. Time period types can include peak hours, off-peak hours, and low-peak hours. Equipment type can be specific to a particular manufacturer, customized, or universal. For a parking scenario attribute of "Enterprise - Morning Peak": the weight of the regional location factor is increased, and the weight of the buffer parking space location factor is decreased to achieve the fastest efficiency.

[0038] In an optional embodiment, before determining the target buffer parking space from each buffer parking space in the set of available parking areas based on buffer parking space selection factors, buffer parking spaces that cannot be used for parking, such as those that are occupied, have offline or faulty equipment corresponding to the buffer parking space, or are for emergency stops, can be filtered out.

[0039] This embodiment can determine the current parking scene attributes based on preset scene data, garage operation data, and time period identification data, and by calling a scene recognition model. The scene recognition model can be an artificial intelligence model. Preset scene data can be the garage location, garage operation data can include parking records, equipment information, etc., and time period identification data can be the current time or the determined time period to which the current time belongs, such as 8:00 to 9:00. Alternatively, on the same day, an artificial intelligence model can be omitted. Instead, historical data can be used to divide the time periods corresponding to various time periods, determine the time period corresponding to the current time, and thus obtain the time period type, such as 8:00 to 9:00 being a peak period. Equipment type can be determined based on equipment identifiers, and garage location can be determined based on preset scene data.

[0040] This application embodiment utilizes one or more of the following factors: buffer parking space selection parameters, area selection parameters, area adaptation factors, area location factors, buffer parking space status factors, and buffer parking space location factors. This enables the rapid selection of buffer parking spaces that are suitable for the user or the most efficient buffer parking spaces, thereby improving the efficiency of vehicles parking in buffer parking spaces and ensuring the accurate and efficient allocation of stacking parking spaces subsequently.

[0041] Regarding step S13: Determining the allocation priority of candidate stacking parking spaces based on multi-dimensional stacking parking space allocation factors, wherein the candidate stacking parking spaces are stacking parking spaces within the target parking area, the target parking area is the parking area to which the target buffer parking space belongs, and the multi-dimensional stacking parking space allocation factors include inner and outer parking space factors, used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first estimated departure time of the target vehicle and the second estimated departure time of vehicles already parked in the inner parking spaces. For example, when the first estimated departure time is earlier than the second estimated departure time, the allocation priority of the outer vacant parking spaces is increased.

[0042] In this embodiment, all available stacking spaces within the target parking area can be identified as candidate stacking spaces, and the allocation priority of the candidate stacking spaces can be determined.

[0043] In an optional embodiment, determining the allocation priority of candidate stacking parking spaces based on multi-dimensional stacking parking space allocation factors may include: sorting candidate stacking parking spaces sequentially according to a preset dimension order based on the multi-dimensional stacking parking space allocation factors, and then sorting the candidate stacking parking spaces with the same order in the sorting results of the previous stacking allocation factors to obtain the allocation priority of the candidate stacking parking spaces.

[0044] The multi-dimensional stacking parking space allocation factors may include one or more of the following: Collaborative coverage factor, used to determine whether candidate stacking parking spaces are covered by the working range of handling equipment in different areas, prioritizing the allocation of parking spaces that can be collaboratively covered to avoid additional handling time and resource conflicts caused by cross-regional scheduling. Reserved occupancy factor, used to determine whether candidate stacking parking spaces are already reserved for the target vehicle, prioritizing the allocation of reserved parking spaces to avoid meaningless parking space changes for the same vehicle and reduce system scheduling overhead. Charging compatibility factor, used to determine priority based on the matching relationship between the target vehicle's bottom charging attribute and the charging function of the candidate stacking parking spaces, prioritizing the allocation of vehicles supporting bottom charging to enabled bottom charging parking spaces, and prioritizing the allocation of vehicles that do not support bottom charging to non-bottom charging parking spaces to avoid charging resource conflicts. Vehicle model compatibility factor, used to determine priority based on the degree of matching between the vehicle models allowed to be parked in the candidate stacking parking spaces and the target vehicle model, prioritizing the selection of stacking parking spaces that perfectly match the target vehicle model, and ensuring that larger vehicles do not occupy smaller parking spaces. The vehicle relocation conflict avoidance factor is used to determine whether there are ongoing storage / retrieval tasks or pending queue tasks in the paired inner parking space when the candidate stacking parking space is an outer parking space. It prioritizes parking spaces that will not cause conflict during relocation to reduce internal handling interference. "Inner" and "outer" are relative to the garage aisle; the adjacent aisle is considered outer, otherwise inner. The garage aisle is the area between buffer parking spaces and stacking parking spaces, i.e., the area where the vehicle handling robot operates. The inner / outer parking space factor is used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the predicted departure time of the target vehicle and the predicted departure time of vehicles already parked in the paired inner parking space. This achieves short-term parking on the outside and long-term parking on the inside to avoid short-term parked vehicles blocking the retrieval passage for long-term parked vehicles. The location safety factor is used to determine the priority based on the location zone type of the candidate stacking parking space. It prioritizes parking spaces located in the safety zone where there is no collision risk during the operation of the handling equipment, and secondly, parking spaces located in the critical zone of adjacent area boundaries. Distance factor is used to prioritize candidate stacking parking spaces based on the distance between them and buffer parking spaces, prioritizing the parking space with the shortest handling distance to reduce mechanical handling time. Preset dimension orders can include: collaborative coverage factor, reserved occupancy factor, charging compatibility factor, vehicle model compatibility factor, vehicle relocation conflict avoidance factor, inner / outer parking space factor, location safety factor, and distance factor. Any number of dimensions can be selected for factor sorting, and the order of factors can also be adjusted. Multi-dimensional stacking parking space allocation factors can affect scheduling costs.

[0045] According to the preset dimension order, the subsequent stacking allocation factors sort the candidate stacking spaces with the same order in the sorting results of the previous stacking allocation factors to obtain the allocation priority of the candidate stacking spaces. That is, there are multiple optimal spaces corresponding to the previous factors. If the priorities are the same, further screening is carried out based on the subsequent factors.

[0046] In an optional embodiment, the representation values ​​of each allocation factor can be weighted based on the weight coefficients corresponding to the stacking parking space allocation factors to obtain a weighted result for the stacking parking spaces; the target stacking parking space can then be determined based on the weighted result. Furthermore, the weight coefficients corresponding to the stacking allocation factors can be determined based on the attributes of the current parking scenario. The mapping relationship between scenario attributes and weight coefficients can be predetermined.

[0047] This application embodiment comprehensively considers various stacking parking space allocation factors to ensure accurate and efficient subsequent scheduling while meeting user needs and reducing operating costs.

[0048] In an optional embodiment, determining the allocation priority of candidate stacking parking spaces based on multi-dimensional stacking parking space allocation factors may include: determining the target stacking parking space allocation parameter corresponding to the target allocation factor among the multi-dimensional stacking parking space allocation factors based on the current parking scenario attributes; and determining the allocation priority of candidate stacking parking spaces based on the target stacking parking space allocation parameter under the target allocation factor.

[0049] This embodiment, based on the attributes of the current parking scenario, determines the allocation factors to be processed as target allocation factors from the multi-dimensional stacked parking space allocation factors, and determines the corresponding target stacked parking space allocation parameters based on the target allocation factors. This establishes a mapping relationship between parking scenario attributes and target stacked parking space allocation parameters. In the corresponding allocation factor dimension, the allocation priority of candidate stacked parking spaces is determined based on the target stacked parking space allocation parameters. In this way, priority control based on the scenario can be achieved, thereby meeting the scheduling requirements of the corresponding scenario.

[0050] In an optional embodiment, the stacked parking spaces in the garage may include inner parking spaces and outer parking spaces. When the target allocation factor is an inner / outer parking space factor, the target stacked parking space allocation parameter is used to adjust the allocation priority between a first type of candidate stacked parking spaces and a second type of candidate stacked parking spaces based on the first expected departure time of the target vehicle. The first type of candidate stacked parking space is an inner parking space when both inner and outer parking spaces are vacant. The second type of candidate stacked parking space is an outer parking space when there is a parked vehicle in the inner parking space and the outer parking space is vacant.

[0051] In this way, by adjusting the allocation priority of candidate stacked parking spaces based on the time relationship between the predicted departure time of the target vehicle and the predicted departure time of the vehicles already parked in the paired inner parking spaces, short-term parking can be placed externally and long-term parking can be placed internally in the corresponding scenarios to avoid short-term parked vehicles blocking the retrieval passage of long-term parked vehicles and improve the rationality of scheduling.

[0052] In an optional embodiment, the target stacking parking space allocation parameters include: a first basic value corresponding to a first type of candidate stacking parking space, a second basic value corresponding to a second type of candidate stacking parking space, a third basic value corresponding to a third type of candidate stacking parking space, a preset outer parking weight coefficient, and a time tolerance; the third type of candidate stacking parking space is an inner parking space when the inner parking space is empty and the outer parking space is occupied by a parked vehicle; under the target allocation factor, determining the allocation priority of candidate stacking parking spaces based on the target stacking parking space allocation parameters may include: when the second type of candidate stacking parking space meets the preset time tolerance condition, determining the weighted value of the second type of candidate stacking parking space based on the second basic value and the preset outer parking weight coefficient; wherein, the preset time tolerance condition is that the first expected departure time of the target vehicle is earlier than the second expected departure time of the parked vehicle in the inner parking space, and the time difference between the first expected departure time and the second expected departure time is greater than the time tolerance; determining the allocation priority of candidate stacking parking spaces based on the first basic value, the weighted value, and the third basic value.

[0053] In practice, the first basic value can be less than the second basic value and less than the third basic value. After weighting, the weighted value can be less than the first basic value or still greater than the first basic value. For example, in the peak scenario of a hospital, the preset weight coefficient for outer parking is smaller to enhance the effectiveness of short-term parking outside; in the off-peak scenario of an enterprise, it is kept larger, and long-term parking is mainly built-in.

[0054] In an optional embodiment, the target stacking parking space allocation parameters can be disregarded. If a second type of candidate stacking parking space exists, and the second type of candidate stacking parking space meets a preset time tolerance condition, a first ranking value is assigned to the second type of candidate stacking parking space; otherwise, a second ranking value corresponding to the outer parking space is assigned. A third ranking value is assigned to the first type of candidate stacking parking space, with the allocation priority from high to low being the first ranking value, the third ranking value, and the second ranking value. For example, if the tolerance time is 30 minutes, the first ranking value is 0, the second ranking value is 2, and the third ranking value is 1. This achieves short-term parking external and long-term parking internal, avoiding the blockage of long-term parking retrieval passages by short-term parked vehicles.

[0055] In an optional embodiment, the departure time of the target vehicle can be predicted based on multiple departure time prediction models; a weighted calculation is performed on the multiple departure times to obtain the first estimated departure time of the target vehicle.

[0056] This embodiment uses multiple departure time prediction models for prediction and then performs weighted calculations, which can improve the accuracy of the predicted departure time calculation.

[0057] In optional embodiments, before predicting based on multiple departure time prediction models, if an input estimated departure time parameter exists, it can be used directly. This parameter can be input by the target vehicle's controller, such as the driver or the dispatcher of the autonomous vehicle, or by the intelligent driving platform. Alternatively, if the target vehicle has charging needs and the vehicle information includes battery capacity parameters, current state of charge parameters, and target state of charge parameters, the estimated departure time of the target vehicle can be determined based on the charging power, which can be used as the first candidate estimated departure time. For example, the charging time required for the target vehicle to charge from the current state of charge to the target state of charge can be calculated. If an input estimated departure time parameter exists as a second candidate estimated departure time, the smaller value can be taken as the estimated departure time of the target vehicle. If no input estimated departure time parameter, battery capacity parameter, current state of charge parameter, or target state of charge parameter exists, then prediction is based on multiple departure time prediction models. That is, the estimated departure time adopts a multi-level approach, divided into three priorities according to the reliability of the information source from high to low: the first priority is the prediction of the input parameter, which comes from proactive provision and has the highest credibility and accuracy. The second priority is charging demand time estimation. If the target vehicle has a charging need and the vehicle information includes battery capacity parameters, current state of charge parameters, and target state of charge parameters, the earlier of the first and second priorities is used. If charging demand time estimation is not possible, the first priority is used. The third priority is prediction based on historical visit patterns. If neither the first nor the second priority yields effective results, a comprehensive prediction is made based on multiple departure time prediction models.

[0058] Specifically, the departure time of the target vehicle can be predicted based on its parking records over a preset number of days and using various departure time prediction models.

[0059] The first departure time prediction model can predict based on historical patterns of weekday dates (i.e., a specific day of the week, such as Wednesday). It extracts all parking records from the past preset number of days that match the target prediction date. If the number of records is less than a first preset number, it returns null. If the number of records is greater than or equal to the first preset number, it calculates the average departure time of all parking records to obtain the first departure time. The confidence level of this model is equal to the smaller of the product of the number of records and a first preset coefficient and a first preset upper limit. The first preset coefficient represents the increase in confidence for each additional record. For example, if the prediction is based on the vehicle's access patterns over the past 30 days, and less than 2 records are extracted, it returns null and ignores the model's result. Otherwise, it calculates the average departure time of these records, with confidence level = min(0.9, number of records × 0.1), meaning the confidence level increases by 0.1 for each additional record, with an upper limit of 0.9.

[0060] The second departure time prediction model can be based on the trend of the most recent preset number of days. It extracts all historical parking records within the most recent preset number of days from the current time, calculates the average change in departure time between two adjacent days to obtain the average trend value, and adds the departure time corresponding to the most recent historical record to the average trend value to obtain the second departure time. This second departure time is then limited to a preset reasonable time range. The confidence level of the second departure time is a second preset confidence level. For example, based on the trend of the most recent 7 days, the records of the most recent 7 days are taken, and the time change trend of adjacent two days is calculated (the average change value is obtained). The last departure time + the average change value = the predicted time, limited to a reasonable range of 7:00-24:00, with a fixed confidence level of 0.7.

[0061] The third departure time prediction model can be a time-decay weighted prediction. All historical parking records within a preset number of days are extracted, and the third departure time is calculated using a time-decay weighted method. A decay coefficient can be set so that historical records closer to the current time have a higher weight, and historical records farther away have a lower weight. For example, for the i-th historical record, its corresponding weight is equal to the decay coefficient raised to the power of (total number of records minus i minus one). Historical records closer to the current time have a higher weight, and historical records farther away have a lower weight. The departure time corresponding to each historical record is multiplied by its corresponding weight, and the sum is then divided by the sum of the weights of all historical records for normalization, yielding the third departure time. The confidence level of the third predicted departure time is the third preset confidence level. For example, with a decay coefficient of 0.9: the weight of the i-th record = 0.9^(total number of records - i - 1), the closer the record is to the current time, the greater the weight. The weight of a record 30 days ago = 0.9^29 ≈0.05, which is almost negligible; the weight of the most recent record = 0.9^0 = 1.0. After weighted summation and normalization, the predicted time is obtained, and the confidence level is fixed at 0.8.

[0062] The fourth departure time prediction model can be based on similarity pattern matching. It obtains the departure time of a reference day (the day before the target prediction date); retrieves records from historical data that meet similarity criteria: the difference between the departure time of a certain first day in the historical records and the departure time of the reference day is less than a preset time deviation threshold, and the day of the week of the second day adjacent to that first day is the same as the day of the week of the target prediction date; if the number of sample records meeting the similarity criteria is less than a preset sample threshold, a null value is returned; otherwise, the average departure time of the second day in each sample record is calculated to obtain the fourth departure time.

[0063] The prediction considers the possibility of correlation between vehicle parking behavior on two adjacent days. For example, it finds samples in history whose departure time is similar to today's (difference < 60 minutes) and tomorrow is the same day of the week. When there are less than 2 similar patterns, it returns null and calculates the average of the departure times of these similar patterns.

[0064] The first predicted departure time can be calculated using the following weighted fusion formula: The first predicted departure time equals the sum of the products of the departure time predicted by each model and its corresponding fusion weight. If the sum of the fusion weights is not 1, it can be divided by the sum of the fusion weights. The fusion weights can be their respective confidence levels, or be positively correlated with their respective confidence levels. For example, if the confidence level of the first model is 0.9, the confidence level of the second model is 0.7, the confidence level of the third model is 0.8, and the confidence level of the fourth model is 0.6, then the first predicted departure time = (0.9×T1 + 0.7×T2 + 0.8×T3 + 0.6×T4) / (0.9+0.7+0.8+0.6).

[0065] The weights can be optimized using artificial intelligence models. For example, in peak enterprise scenarios, the optimized parameters could be a weight of 0.5 for the first model, 0.3 for the fourth model, and a decay coefficient of 0.9. Presets can also be made based on the specific application scenario. The artificial intelligence model can be dynamically optimized based on historical prediction accuracy.

[0066] Furthermore, the departure time prediction model can also employ a large model enhancement prediction method (AI_LLM_Enhanced): This method uses scene attributes and historical vehicle parking data as basic inputs, while also incorporating related feature data including weather information, public event information, holiday information, and real-time regional load status. This data is then input into the AI ​​large model for scene-aware inference, dynamically generating prediction results and corresponding confidence levels. This is combined with the aforementioned four models to determine the first estimated departure time. The AI ​​large model can understand scene associations such as increased hospital outpatient volume on rainy days and delayed departure times, and integrates external factors such as weather, holidays, and special events to improve prediction accuracy.

[0067] This embodiment can perform anomaly detection, that is, identify and automatically remove outliers in historical data, thereby improving prediction robustness. Furthermore, the fusion weights can be dynamically adjusted based on the historical accuracy of each model in the current corresponding scenario.

[0068] In an optional embodiment, when the target allocation factor is a distance factor, the target stacking parking space allocation parameter may include: coefficients of each dimension coordinate; under the target allocation factor, determining the allocation priority of candidate stacking parking spaces based on the target stacking parking space allocation parameter includes: performing a weighted calculation based on the coefficients of each dimension coordinate and the difference between each dimension coordinate of the candidate stacking parking space and each dimension coordinate of the target buffer parking space to obtain a distance value; determining the allocation priority of candidate stacking parking spaces based on the distance value.

[0069] Each dimension of the coordinate system can include three dimensions, meaning the coordinate system in the garage can be a three-dimensional coordinate system, with the direction perpendicular to the horizontal ground as the height direction, to represent the corresponding chassis layer. Sorting value = c_x |P.Location_X-CPS.Location_X|+c_y |P.Location_Y-CPS.Location_Y|+c_z |P.Location_Z-CPS.Location_Z|, with smaller values ​​taking precedence. c_x, c_y, and c_z are coefficients for the three coordinate dimensions. P.Location represents the location coordinates of the stacked parking space, and CPS.Location represents the location coordinates of the buffer parking space. X, Y, and Z are the three coordinate axes, which can be set according to the garage design. For example, if Y is the coordinate axis corresponding to the height direction, c_x=1, c_y=1, c_z=1, or different coefficients can be used. The coefficients can be set based on experience or optimized using a model. For example, in a corporate park scenario, c_y can be adjusted according to the floor layout (increasing the depth weight); a more refined distance coefficient is required for an autonomous vehicle scenario. This ensures that the distance is more consistent with the scenario, thereby determining the most efficient stacked parking space in the corresponding scenario.

[0070] In one specific implementation, after obtaining the candidate stacking location list, the optimal stacking location is selected through an 8-level sorting process (Level 1 to Level 8). The multi-level sorting is performed sequentially from Level 1 to Level 8, with each level having a sorting key value (smaller values ​​have higher priority). The stacking location with the smallest sorting key value, corresponding to the highest priority, is selected as the target stacking location. First, the sorting key values ​​of Level 1 are compared; smaller stacking locations are generally ranked higher. If multiple sorting key values ​​are equal at the beginning, then Level 2 is compared, and so on, up to Level 8. Of course, the order of different stacking allocation factors can be adjusted, and some of the stacking allocation factors involved in determining the target stacking location can be selected from the aforementioned 8 dimensions. Specifically, in the above 8-level sorting process, each level from Level 1 to Level 8 represents a specific manifestation of a 1-dimensional stacking parking space allocation factor within the multi-dimensional stacking parking space allocation factors. The multi-dimensional stacking parking space allocation factors can include at least one target allocation factor, such as inner / outer parking space factors or distance factors. Target allocation factors have corresponding target stacking parking space allocation parameters. The determination logic for each level is as follows: Level 1: Collaborative Coverage Factors, Collaborative Parking Spaces Prioritized. Decision Logic: If a candidate stacking parking space is a collaborative parking space, the sorting key value = 0; otherwise, the sorting key value = 1. Collaborative parking spaces are those covered by the working range of handling equipment in different areas, and are given priority.

[0071] Level 2: Prioritize reserved parking spaces. The decision logic is as follows: if a candidate parking space is already reserved for the target vehicle, the sorting key value = 0; otherwise, the sorting key value = 1. If a space has already been reserved for the same vehicle, it will be reused during rescheduling to avoid unnecessary relocation.

[0072] Level 3: Charging compatibility factors. The judgment logic is divided into two branches based on whether the vehicle supports bottom charging (IsBottomCharging): Bottom Charging Vehicles: ① If the vehicle intends to charge (IfCharging=true) + IsDemoPsr, the sorting key value is 0, where IsDemoPsr represents a specific demonstration parking space. ② If the target parking space isBottomCharging + ChargeEnableState=enabled + bottom charging location matches, the sorting key value is 1; ③ If the target parking space does not support bottom charging, the sorting key value is 2; ④ If the target parking space isBottomCharging+enabled but the bottom charging location does not match, the sorting key value is 3; ⑤ If the vehicle is not charging + IsDemoPsr, the sorting key value is 4. Non-Bottom Charging Vehicles: ① If the target parking space isBottomCharging+enabled, the sorting key value is 3 (to avoid occupation); ② If the target parking space does not support bottom charging, the sorting key value is 1. Vehicles supporting bottom charging are prioritized for matching to enabled and suitable bottom charging parking spaces; vehicles not supporting bottom charging are prioritized for matching to non-bottom charging parking spaces to avoid resource conflicts.

[0073] Level 4: Vehicle Model Compatibility Factors. Judgment Logic: The sorting key value equals the number of vehicle models that the parking space can hold; the smaller the value, the higher the priority. Larger parking spaces can hold more vehicle models. Of course, the target vehicle's model must be included. First, select parking spaces that perfectly match the vehicle's size, then expand to the next larger size; simultaneously, adhere to the principle that larger vehicles should not occupy smaller spaces.

[0074] Level 5: Factors to avoid parking conflicts and reduce parking space movement. Judgment logic: When a candidate stacking parking space is an outer parking space, find its paired inner parking space P_inner; if P_inner exists and there is an ongoing parking space movement task (or the parking space movement task is in the queue), then the sort key value = 1; otherwise, the sort key value = 0. If the paired inner parking space of an outer parking space is currently undergoing a storage / access operation, then the outer parking space is pushed back to avoid parking conflicts with inner parking space tasks.

[0075] Level 6: Inner and Outer Parking Space Factors, Optional Judgment Logic 1: Corresponding Configuration Items (Dynamically Deployed by AI Model): For example, PredictedInterval = 30: Time tolerance for inward / outer space flipping (minutes); PredictedCoefficient = 0.4: Weighting coefficient of predicted departure time on outward flipping; io_01 = 300, io_10 = 200, io_00 = 100, representing the basic values ​​of different types of candidate stacked parking spaces, satisfying the time tolerance condition. The basic value corresponding to io_10 is multiplied by the weighting coefficient of outward flipping, prioritizing the outer side; io_XY represents the parking space type, X represents whether the inner side is vacant (1 indicates not vacant, occupied; 0 indicates vacant); Y represents whether the outer position is vacant. AI Optimization Direction: In hospital peak scenarios, PredictedCoefficient is smaller, enhancing the outward flipping force for short stops; in enterprise off-peak scenarios, it is kept larger, with long stops primarily built-in. Basic values ​​can also be predicted by the AI ​​model. Optional Decision Logic 2: When the candidate stacked parking space P is an outer parking space and the current vehicle's `currentVehicle.PredictedEndTime` (i.e., the first estimated departure time) is not empty, search for P's paired inner parking space P_inner; if P_inner exists and P_inner.PredictedEndTime_1 (i.e., the second estimated departure time) is not empty, and P_inner.PredictedEndTime_1 - currentVehicle.PredictedEndTime > PredictedInterval (default 30 minutes), then the sorting key value = 0 (short parking is moved to the outer side); for outer parking spaces, the sorting key value is 2, and for inner parking spaces, the sorting key value is 1. When the vehicle's predicted departure time is more than 30 minutes earlier than the predicted departure time of its paired inner parking space, the vehicle is forcibly placed on the outer side; otherwise, the inner side priority is maintained. Short parking spaces are placed on the outer side, and long parking spaces are placed on the inner side to avoid short-parking vehicles blocking the retrieval of long-parking vehicles.

[0076] Level 7: Location safety factors, priority safety zone, then proximity zone, the collaborative zone is the parking space covered by the area that the handling equipment can cross. The collaborative zone may include the safety zone (the area where the handling equipment will not collide) and the critical zone (two adjacent areas with a risk of collision). Judgment logic: safety zone, sort key value is 0, critical zone sort key value is 1.

[0077] Level 8: Distance factor, using weighted Manhattan distance (dynamically adjusted by AI) for judgment logic: sort key value = c_x |P.Location_X-CPS.Location_X|+c_y |P.Location_Y-CPS.Location_Y|+c_z |P.Location_Z-CPS.Location_Z|, the smaller the value, the higher the priority. Assuming the preceding factors (vehicle type, charging base, inner / outer sides, area, etc.) are met, the parking space closest to the buffer parking entrance is selected based on the weighted Manhattan distance to minimize mechanical time. Corresponding configuration items (dynamically assigned by AI): c_x=1, c_y=1, c_z=1, distance weights for the X / Y / Z axes; 1 can be the initial value. For enterprise park scenarios, c_y is adjusted according to the floor layout (increasing the depth weight); for autonomous vehicle scenarios, the distance coefficient is refined.

[0078] In an optional embodiment, when the parameter optimization condition is met, parking scheduling is simulated in a simulation environment based on the garage's operating data and using the current target stacked parking space allocation parameters, and cost data is obtained; based on the cost data, the target stacked parking space allocation parameters corresponding to each parking scenario attribute are optimized using a large model to obtain optimized parameters; the optimized parameters are determined as the current target stacked parking space allocation parameters, and the step of simulating parking scheduling in a simulation environment based on the garage's operating data and using the current target stacked parking space allocation parameters to obtain cost data is triggered; when the optimization termination condition is met, the current optimized parameters are cached; correspondingly, the target stacked parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacked parking space allocation factors are determined based on the current parking scenario attribute, including: determining the target stacked parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacked parking space allocation factors under the current parking scenario attribute from the cached optimized parameters. Simultaneously, the fusion weights of multiple departure time prediction models can also be optimized.

[0079] The parameter optimization conditions can be either reaching a set time for periodic optimization or the cost in the actual operating environment exceeding a preset cost threshold. The threshold can be set based on experience, and the cost can be a comprehensive cost determined based on energy consumption, equipment costs, and time costs in the actual operating environment. The operating data for different rounds of input models can be the same or different. The large model is an AI model with a higher number of parameters than a certain specification. The large model runs only in the background computing layer, producing optimal parameter configurations. The real-time scheduling layer reads parameters from the cache and executes the scheduling. The two are decoupled through the configuration cache layer, leveraging the optimization capabilities of the large model while meeting millisecond-level real-time response requirements. Parameter optimization based on operating data and the large model ensures parameter rationality and reduces costs. Of course, in optional embodiments, the target stacking space allocation parameters can also be set based on experience. The optimization termination condition can be determined based on the cost curve corresponding to the comprehensive cost. For example, by plotting the cost curve of the comprehensive cost changing with the number of iterations, if the latest value of the comprehensive cost is less than or equal to the preset target cost threshold, the cost curve has decreased to an acceptable cost level. Furthermore, if the change in the cost curve within the most recent preset number of iterations is less than the preset stable threshold, the cost curve has entered a stable state and no longer decreases significantly.

[0080] In an optional embodiment, the cost data may include energy consumption cost, equipment cost, and time cost; the comprehensive cost may be determined based on the energy consumption cost, the equipment cost, and the time cost; when the comprehensive cost converges, it is determined that the optimization termination condition is met.

[0081] The energy cost can include the electricity consumption for scheduling and the electricity cost determined by the electricity price. Equipment cost can include equipment wear and tear, such as the wear and tear caused by handling equipment, which can be determined based on the number of times the action is performed. Time cost can include the time required to complete scheduling and park the item in the stacking bay. The comprehensive cost can be determined by weighting the energy cost, equipment cost, time cost, and the coefficients corresponding to each cost. These coefficients can be set based on empirical values ​​or optimized using artificial intelligence models. For example, the energy cost, equipment cost, and time cost could be 0.3, 0.4, and 0.3, respectively.

[0082] See Figure 3 As shown, Figure 3 This application provides a parameter optimization flowchart. When the parameter optimization conditions are met, the system simulates parking scheduling, obtains cost data, optimizes parameters using a large model, and iteratively updates the parameters. When the overall cost converges, the optimized parameters are cached.

[0083] In this way, by taking into account various costs in the scheduling process, it is possible to reduce scheduling costs when using optimized parameters for scheduling.

[0084] In an optional embodiment, the optimized parameters can be released in a phased manner, and rolled back if a monitoring alarm is triggered. If the overall cost in the real environment exceeds the preset alarm threshold, the threshold can be set based on experience.

[0085] Regarding step S14: Based on the allocation priority, a target stacking parking space is determined from the candidate stacking parking spaces, wherein the target stacking parking space is used to park the target vehicle. Specifically, the candidate stacking parking space with the highest allocation priority is designated as the target stacking parking space, so that the handling equipment is controlled to move the target vehicle from the buffer parking space to the target stacking parking space. If there is a charging requirement, the charging equipment is controlled to charge the target vehicle. In addition, it can also be checked whether there are any conflicts such as reserved occupancy or special vehicle reservation in the target stacking parking space. After confirming that there are no abnormalities, the final parking space scheduling result is generated and sent to terminals such as the barrier gate and the in-yard guidance screen for execution.

[0086] As can be seen, the embodiments of this application first filter parking areas based on vehicle information, transforming global area traversal into local targeted search, and then determine the target parking area based on the target buffer parking space, thus precisely limiting the selection range of stacked parking spaces and improving the efficiency of parking space search. Furthermore, through comprehensive evaluation of multi-dimensional stacked parking space allocation factors, the allocation priority is determined, achieving accurate matching between vehicles and parking spaces, improving the overall resource utilization efficiency of the garage, shortening user waiting time, and accurately and efficiently dispatching target vehicles to suitable stacked parking spaces, reducing dispatching costs and improving user experience.

[0087] This application's embodiments can differentiate between different scenarios, considering the varying characteristics of hospitals, corporate parks, commercial complexes, residential communities, unmanned vehicle parking lots, different time periods, and different equipment. For example, in corporate parks, peak commuting hours (8-9 am, 5-6 pm) are predominantly long-term (8-10 hours), exhibiting strong regularity, requiring priority parking and charging management in scheduling. Unmanned vehicle parking lots operate automatically 24 / 7, with no user waiting concept, requiring priority energy consumption, balanced equipment utilization, and optimized idle load in scheduling. Hospitals have concentrated peak hours (8-10 am, 4-6 pm), predominantly short-term (1-4 hours), with frequent parking and retrieval, requiring priority retrieval, allocation of short-term parked vehicles to the outer areas, and sensitivity to predicted departure times, necessitating differentiated scheduling. Parameter optimization, through artificial intelligence models, can consider the coupling relationships between parameters and address time-related changes. The optimal parameters for the same scenario differ significantly during peak / off-peak / low-peak periods. Simulations are used to verify the effects of parameter adjustments without a systematic approach, and the optimization results can be measured through cost and updated promptly. This approach constructs a multi-scenario classification system combining location, device, and time period, enabling adaptive parameter management for scenario awareness. The AI ​​large-scale model periodically calculates optimal parameters in the background, which are then injected into the real-time scheduling layer through a configuration caching layer, forming a closed loop of offline optimization and online execution. Furthermore, multi-method prediction combined with AI-enhanced prediction ensures the accuracy of the prediction results. A three-dimensional quantitative cost model (equipment / energy consumption / efficiency) is established, and the parameter effects are verified through playback simulations. A four-layer decoupled architecture is provided: an AI background parameter optimization layer, a parameter caching layer, a real-time scheduling layer, and a monitoring and evaluation layer. The AI ​​large-scale model runs only in the background computation layer, producing optimal parameter configurations; the real-time scheduling layer reads parameters from the cache and executes scheduling. These two layers are decoupled through the configuration caching layer, leveraging the optimization capabilities of AI while ensuring millisecond-level real-time response.

[0088] In an optional embodiment, see Figure 4 As shown, Figure 4 This application provides a partial schematic diagram of a garage, taking the partial area as an example of being divided into two parking areas, including area one and area two. Each parking area includes stacking parking spaces, buffer parking spaces, and vehicle transfer devices. As shown in the figure, each parking area includes one vehicle transfer device, namely vehicle transfer device one and vehicle transfer device two. The vehicle transfer device can be a handling robot. Each area can contain four buffer parking spaces.

[0089] In one alternative embodiment, the garage may include, for example: Figure 4 The three local areas shown have lanes in the other two local areas respectively. Figure 4 The lane shown is perpendicular to and is parallel to Figure 4 The lanes shown are parallel, and the three local areas form a U-shape, with stacked parking spaces allowed on the outer edge.

[0090] Specifically, see Figure 5As shown, Figure 5 This document presents an application architecture diagram for a parking scheduling scheme provided in this application example. It implements a four-layer decoupled architecture: an AI-based backend parameter optimization layer (operating offline, minute-level), which serves as the offline parameter optimization module; it acquires historically accumulated and / or real-time collected garage operation data as optimization input; it simulates the parking scheduling process in a simulation environment; and iteratively calculates the simulation results using a large AI model to obtain optimized parameters. The optimization results are output to the parameter cache layer. The parameter cache layer stores parameters, enabling millisecond-level retrieval. This layer is the parameter storage module, used to cache optimized parameters categorized by parking scenario, such as scenario A parameters, scenario B parameters, scenario C parameters, etc., for scenario N parameters, where A, B, and C are the identifiers of the corresponding scenarios. The cache layer provides parameter retrieval services to the real-time scheduling layer with millisecond-level response speed. The real-time scheduling layer is the online scheduling execution module, receiving parking requests from users or upper-layer systems; based on the current parking scenario attributes, it reads the optimal matching parameters from the parameter cache layer in milliseconds; and performs scheduling calculations based on the read optimal parameters to complete parking space allocation. The scheduling results can be output to the monitoring and evaluation layer. The monitoring and evaluation layer can include real-time monitoring and offline analysis. This layer is the system monitoring and evaluation module, collecting scheduling effect (actual operation) data. It collects real-time scheduling effect data from the real-time scheduling layer during actual operation. Based on the collected operational data, it calculates the comprehensive cost from three dimensions: energy cost, equipment cost, and time cost. If the cost exceeds a threshold, an alarm rollback is triggered. Specifically, it determines whether the comprehensive cost exceeds a preset threshold; if so, an alarm is triggered, and parameter rollback is performed. A feedback mechanism is used, and the monitoring and evaluation layer forms two feedback paths based on the judgment result: Normal feedback: If the comprehensive cost does not exceed the preset threshold, the collected garage operation data is fed back to the AI ​​backend parameter optimization layer as input data for the next round of parameter optimization; Abnormal feedback: When the comprehensive cost exceeds the preset threshold, the previous version of the parameters is written back to the parameter cache layer, overwriting the parameters in the current cache. Simultaneously, it can trigger the AI ​​backend parameter optimization layer to start a new round of parameter optimization. See also Figure 6 As shown in the figure, this application provides a parking dispatching device, including: The vehicle information acquisition module 11 is used to acquire the vehicle information of the target vehicle in response to a parking request. The buffer parking space determination module 12 is used to determine the target buffer parking space from each buffer parking space based on the vehicle information; The priority determination module 13 is used to determine the allocation priority of candidate stacking parking spaces based on multi-dimensional stacking parking space allocation factors. The candidate stacking parking spaces are stacking parking spaces in the target parking area, the target parking area is the parking area to which the target buffer parking space belongs, and the multi-dimensional stacking parking space allocation factors include inner and outer parking space factors. The module is used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first expected departure time of the target vehicle and the second expected departure time of the vehicle already parked in the inner parking space. The target vehicle determination module 14 is used to determine a target stacking parking space from the candidate stacking parking spaces based on the allocation priority, wherein the target stacking parking space is used to park the target vehicle.

[0091] As can be seen, after obtaining the vehicle information of the target vehicle, the target buffer parking space is determined from each buffer parking space based on the vehicle information. Then, based on multi-dimensional stacking parking space allocation factors, the allocation priority of candidate stacking parking spaces is determined within the target parking area. Based on the allocation priority, the target stacking parking space is determined from the candidate stacking parking spaces. The target parking area is the parking area to which the target buffer parking space belongs. The multi-dimensional stacking parking space allocation factors include inner and outer parking space factors, which are used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first estimated departure time of the target vehicle and the second estimated departure time of vehicles already parked in inner parking spaces. In this way, short-term parking vehicles are prioritized to be allocated to outer parking spaces, and long-term parking vehicles are prioritized to be allocated to inner parking spaces. This avoids the extra handling operation required to move long-term parking vehicles on the outer side because short-term parking vehicles are blocked on the inner side, thereby effectively reducing the number of vehicle handling operations, shortening the waiting time for retrieving the vehicle, and improving the efficiency of garage scheduling.

[0092] The specific implementation method of the above-mentioned parking dispatching device can be referred to the description of the aforementioned parking dispatching scheme, and will not be repeated here.

[0093] See Figure 7 As shown in the figure, this application discloses an electronic device 20, including a processor 21 and a memory 22; wherein, the memory 22 is used to store a computer program; the processor 21 is used to execute the computer program, the parking scheduling method disclosed in the foregoing embodiments.

[0094] The specific process of the above parking scheduling method can be found in the relevant content disclosed in the foregoing embodiments, and will not be repeated here.

[0095] Furthermore, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk, or optical disk, and the storage method can be temporary storage or permanent storage.

[0096] In addition, the electronic device 20 also includes a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26; wherein, the power supply 23 is used to provide operating voltage for the various hardware devices on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0097] Furthermore, embodiments of this application also disclose a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the parking scheduling method disclosed in the foregoing embodiments.

[0098] The specific process of the above parking scheduling method can be found in the relevant content disclosed in the foregoing embodiments, and will not be repeated here.

[0099] Furthermore, this application also discloses a computer program product, including a computer program / instructions, which, when executed by a processor, implements the parking scheduling method disclosed in the foregoing embodiments.

[0100] The specific process of the above parking scheduling method can be found in the relevant content disclosed in the foregoing embodiments, and will not be repeated here.

[0101] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0102] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0103] The above provides a detailed description of the parking scheduling method, electronic device, medium, and product provided by this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A parking dispatching method, characterized in that, include: In response to a parking request, obtain the vehicle information of the target vehicle; Based on the vehicle information, the target buffer parking space is determined from each buffer parking space; The allocation priority of candidate stacking parking spaces is determined based on multi-dimensional stacking parking space allocation factors. The candidate stacking parking spaces are stacking parking spaces within the target parking area, and the target parking area is the parking area to which the target buffer parking space belongs. The multi-dimensional stacking parking space allocation factors include inner and outer parking space factors, which are used to adjust the allocation priority of candidate stacking parking spaces based on the time relationship between the first expected departure time of the target vehicle and the second expected departure time of the vehicle already parked in the inner parking space. A target stacking parking space is determined from the candidate stacking parking spaces based on the allocation priority, wherein the target stacking parking space is used to park the target vehicle.

2. The parking dispatching method according to claim 1, characterized in that, The allocation priority of candidate stacking parking spaces is determined based on multi-dimensional stacking parking space allocation factors, including: Based on the current parking scenario attributes, determine the target stacking parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacking parking space allocation factors; Under the target allocation factors, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters.

3. The parking dispatching method according to claim 2, characterized in that, When the target allocation factor is the inner and outer parking space factor, the target stacking parking space allocation parameter is used to adjust the allocation priority between the first type of candidate stacking parking space and the second type of candidate stacking parking space based on the first expected departure time of the target vehicle. The first type of candidate stacking parking space is the inner parking space when both the inner and outer parking spaces are vacant; the second type of candidate stacking parking space is the outer parking space when there is a parked vehicle in the inner parking space and the outer parking space is vacant.

4. The parking dispatching method according to claim 3, characterized in that, The target stacking parking space allocation parameters include: a first basic value corresponding to a first type of candidate stacking parking space, a second basic value corresponding to a second type of candidate stacking parking space, a third basic value corresponding to a third type of candidate stacking parking space, a preset outer parking weight coefficient, and a time tolerance; the third type of candidate stacking parking space is an inner parking space when the inner parking space is empty and the outer parking space is occupied by a parked vehicle. Under the target allocation factor, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters, including: When the second type of candidate stacking parking space meets the preset time tolerance condition, the weighted value of the second type of candidate stacking parking space is determined based on the second basic value and the preset outer parking weight coefficient; wherein, the preset time tolerance condition is that the first expected departure time of the target vehicle is earlier than the second expected departure time of the vehicle already parked in the inner parking space, and the time difference between the first expected departure time and the second expected departure time is greater than the time tolerance. The allocation priority of candidate stacking parking spaces is determined based on the first basic value, the weighted value, and the third basic value.

5. The parking dispatching method according to claim 3, characterized in that, The method further includes: The departure time of the target vehicle is predicted based on multiple departure time prediction models; The first estimated departure time of the target vehicle is obtained by weighting and calculating multiple departure times.

6. The parking dispatching method according to claim 2, characterized in that, When the target allocation factor is a distance factor, the target stacking parking space allocation parameters include: coefficients of each dimension coordinate; Under the target allocation factor, the allocation priority of candidate stacking spaces is determined based on the target stacking space allocation parameters, including: The distance value is obtained by weighting the coefficients of each dimension of the coordinates and the differences between the coordinates of each dimension of the candidate stacking parking space and the coordinates of each dimension of the target buffer parking space. The allocation priority of candidate stacking parking spaces is determined based on the distance value.

7. The parking dispatching method according to claim 2, characterized in that, The method further includes: When the parameter optimization conditions are met, parking scheduling is simulated in the simulation environment based on the garage's operating data and the current target stacking parking space allocation parameters, and cost data is obtained. Based on the cost data and using a large model, the target stacking parking space allocation parameters corresponding to each parking scenario attribute are optimized to obtain the optimized parameters; The optimized parameters are determined as the current target stacking parking space allocation parameters, and the steps of triggering parking scheduling based on garage operation data and using the current target stacking parking space allocation parameters in the simulation environment, and obtaining cost data are initiated. When the optimization termination condition is met, the current optimized parameters are cached. Accordingly, based on the current parking scenario attributes, the target stacking parking space allocation parameters corresponding to the target allocation factor in the multi-dimensional stacking parking space allocation factors are determined, including: The target stacked parking space allocation parameters are determined from the optimized parameters in the cache, which are the target allocation factors among the multi-dimensional stacked parking space allocation factors under the current parking scenario attributes.

8. The parking dispatching method according to claim 7, characterized in that, The cost data includes energy costs, equipment costs, and time costs. The method further includes: The comprehensive cost is determined based on the energy consumption cost, the equipment cost, and the time cost. When the overall cost converges, the optimization termination condition is determined to be met.

9. The parking dispatching method according to claim 1, characterized in that, The allocation priority of candidate stacking parking spaces is determined based on multi-dimensional stacking parking space allocation factors, including: Based on the multidimensional stacking parking space allocation factors, the candidate stacking parking spaces are sorted sequentially according to a preset dimension order. The subsequent stacking allocation factors sort the candidate stacking parking spaces with the same order in the sorting results of the previous stacking allocation factors to obtain the allocation priority of the candidate stacking parking spaces.

10. The parking scheduling method according to any one of claims 1 to 9, characterized in that, Based on the vehicle information, the target buffer parking space is determined from each buffer parking space, including: Based on the vehicle information, a set of parking areas in the garage is determined, wherein each parking area includes buffer parking spaces and stacking parking spaces; The target buffer parking space is determined from each buffer parking space in the set of available parking areas.

11. The parking scheduling method according to claim 10, characterized in that, Determining the target buffer parking space from each buffer parking space in the set of available parking areas includes: The target buffer parking space is determined from each buffer parking space in the set of available parking areas based on the buffer parking space selection factors. The buffer parking space selection factors include one or more of the following: buffer parking space selection parameters, area selection parameters, area adaptation factors, area location factors, buffer parking space status factors, and buffer parking space location factors.

12. An electronic device, characterized in that, Includes memory and processor, wherein: The memory is used to store computer programs; The processor is configured to execute the computer program to implement the parking scheduling method as described in any one of claims 1 to 11.

13. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the parking scheduling method as described in any one of claims 1 to 11.

14. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the parking scheduling method as described in any one of claims 1 to 11.

Images