Method, apparatus, medium and device for predicting parking space vacancy rate

Abstract

Provided are a method, an apparatus, a medium, and a device for predicting the vacancy rate of a parking space. The vacancy rate of the parking space of each parking lot within the area to be predicted at a plurality of times before the time to be predicted is input into the feature extraction network of the parking space vacancy rate prediction model to obtain a first feature for characterizing the time-dependent relationship of the vacancy rate of the parking space of each parking lot. Also, the spatial relationship graph between each parking lot within the area to be predicted and the first feature are input into the graph fusion network to obtain a fused feature. Then, the fused feature is input into the result prediction network to obtain the vacancy rate of the parking space of each parking lot within the area to be predicted at the time to be predicted. Based on the relationship between the vacancy rate of the parking space and time, the relationship between the vacancy rate of the parking space and space, and the potential correlation between the vacancy rate of the parking space and time and space, by simultaneously obtaining the prediction results of the vacancy rates of the parking spaces of a plurality of parking lots within the area to be predicted, the prediction efficiency of the vacancy rate of the parking space and the accuracy of the prediction result can be improved.

Description

[Technical Field]

[0001] The present disclosure relates to the field of computer technology, and more particularly to a method, apparatus, medium and device for predicting parking space vacancy rate. [Background technology]

[0002] With the development of science and technology, artificial intelligence is also developing rapidly, and machine learning models are widely used in the field of transportation big data.

[0003] Generally, machine learning models can be used to predict the availability of parking spaces in a certain area, allowing drivers to select a parking space based on the predicted information. However, the process of obtaining the model requires consideration of not only the relationship between the availability of parking spaces in the area and time, but also the spatial correlation between each parking space. Therefore, how to simultaneously determine the future availability of parking spaces in multiple parking lots is a difficult task.

[0004] In consideration of these problems, the present disclosure provides a method for predicting the vacancy rate of parking spaces. Summary of the Invention [Problem to be solved by the invention]

[0005] The present disclosure provides a method, apparatus, medium, and device for predicting parking space vacancy rates. [Means for solving the problem]

[0006] The technical solutions used in this disclosure are as follows:

[0007] The present disclosure provides a method for predicting parking space vacancy rate, the method being applied to a computing device, and a pre-trained parking space vacancy rate prediction model being configured on the computing device, the parking space vacancy rate prediction model including a feature extraction network, a graph fusion network, and an outcome prediction network, and the method includes: determining an area to be predicted and a time to be predicted; acquiring, as a historical vacancy rate for each parking lot, the vacancy rate of a parking space at a plurality of times prior to the time to be predicted for each parking lot within the area to be predicted; inputting the historical vacancy rate of each parking space into the feature extraction network to obtain first features, the first features being used to characterize the time dependency of the historical vacancy rate of each parking space; constructing a spatial relationship graph between parking lots in the area to be predicted; inputting the spatial relationship graph and the first feature into the graph fusion network to obtain a fused feature; and inputting the fused features into the result prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted.

[0008] Optionally, before inputting the spatial relationship graph and the first feature into the graph fusion network, the method further comprises: inputting the historical vacancy rate of each parking space into the feature extraction network to obtain a second feature, the second feature being used to characterize the similarity of the historical vacancy rate of each parking space; inputting the spatial relationship graph and the first feature into the graph fusion network, The method further includes inputting the spatial relationship graph, the first feature, and the second feature into the graph fusion network.

[0009] Optionally, the step of constructing a spatial relationship graph between each parking lot in the area to be predicted specifically includes: The method includes a step of constructing the spatial relationship graph by using each parking lot in the area to be predicted as a node and distances between each parking lot in the area to be predicted as edge weights.

[0010] Optionally, the graph fusion network comprises an attention network and a graph convolutional network; inputting the spatial relationship graph, the first feature, and the second feature into the graph fusion network, inputting the spatial relationship graph, the first feature, and the second feature into the attention network to obtain an attention-weighted output result; and inputting the output results and the historical vacancy rates of each parking lot into the graph convolutional network.

[0011] Optionally, the parking space vacancy rate prediction model further comprises an environmental feature extraction network; Before inputting the fused features into the outcome prediction network, the method further comprises: acquiring environmental information for each parking lot within the area to be predicted before the time to be predicted, the environmental information including at least weather information and holiday information; inputting the environmental information into the environmental feature extraction network to obtain environmental features; The step of inputting the fusion features into the outcome prediction network includes: The method further includes inputting the first feature, the fusion feature, the environmental feature, and the historical vacancy rate of each parking lot into the outcome prediction network.

[0012] Optionally, the parking space vacancy rate prediction model comprises: Obtaining the vacancy rate of parking spaces at each parking lot within the designated area at at least two historical times; a step of labeling the vacancy rate of parking spaces in each parking lot at the latest of the at least two historical times, and sampling the vacancy rate of parking spaces in each parking lot at other historical times; inputting the samples into the feature extraction network to obtain sample features, the sample features being used to characterize the relationship between historical vacancy rate and time for each of the parking spaces; constructing a spatial relationship graph between each parking lot in the designated area as a sample spatial relationship graph; inputting the sample spatial relationship graph and the sample features into the graph fusion network to obtain a sample fusion feature; inputting the sample fusion features into the result prediction network to obtain a prediction result of the vacancy rate of each parking space at the latest historical time in the specified area; and training the parking space vacancy rate prediction model with the optimization goal of minimizing the difference between the predicted vacancy rate of each parking space and the label.

[0013] The present disclosure provides a parking space vacancy rate prediction device, wherein the parking space vacancy rate prediction device is configured with a pre-trained parking space vacancy rate prediction model, the parking space vacancy rate prediction model includes a feature extraction network, a graph fusion network, and an outcome prediction network, and the device comprises: a determination module for determining an area to be predicted and a time to be predicted; an acquisition module for acquiring the parking space vacancy rate of each parking lot within the area to be predicted at a plurality of times prior to the time to be predicted as a historical vacancy rate of each parking lot; a first input module for inputting the historical vacancy rate of each parking lot into the feature extraction network to obtain first features, the first features being used to characterize the time dependency of the historical vacancy rate of each parking lot; a construction module for constructing a spatial relationship graph between parking lots in the area to be predicted; a second input module for inputting the spatial relationship graph and the first features into the graph fusion network to obtain a fused feature; and an output module for inputting the fused features into the outcome prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted.

[0014] Optionally, the first input module specifically: inputting the historical vacancy rate of each of the parking spaces into the feature extraction network to obtain second features, the second features being used to characterize the similarity of the historical vacancy rates of each of the parking spaces; The second input module specifically includes: The spatial relationship graph, the first feature, and the second feature are used to input into the graph fusion network.

[0015] Optionally, the construction module specifically: The spatial relationship graph is constructed by using each parking lot in the area to be predicted as a node and the distance between each parking lot in the area to be predicted as an edge weight.

[0016] Optionally, the graph fusion network comprises an attention network and a graph convolutional network; The second input module specifically includes: inputting the spatial relationship graph, the first feature, and the second feature into the attention network and obtaining an attention-weighted output result; The output results and the historical vacancy rate of each parking lot are used to input into the graph convolutional network.

[0017] Optionally, the parking space vacancy rate prediction model further comprises an environmental feature extraction network; The device further includes an environmental information input module; The environmental information input module specifically includes: acquiring environmental information for each parking lot within the area to be predicted before the time to be predicted, the environmental information including at least weather information and holiday information, and inputting the environmental information into the environmental feature extraction network to obtain environmental features; The output module specifically includes: The first features, the fusion features, the environmental features, and the historical vacancy rate of each parking lot are used to input into the outcome prediction network.

[0018] Optionally, the apparatus further comprises a training module; The training module specifically includes: Obtain the parking space vacancy rate for at least two historical times for each parking lot within the specified area; The vacancy rate of the parking spaces of each parking lot at the latest historical time among the at least two historical times is used as a label, and the vacancy rate of the parking spaces of each parking lot at the other historical times is used as a sample; inputting the samples into the feature extraction network to obtain sample features, the sample features being used to characterize the relationship between historical vacancy rate and time for each of the parking spaces; constructing a spatial relationship graph between each parking lot in the specified area as a sample spatial relationship graph; inputting the sample spatial relationship graph and the sample features into the graph fusion network to obtain sample fusion features; inputting the sample fusion features into the result prediction network to obtain a prediction result of the vacancy rate of each parking space at the latest historical time in the specified area; The optimization goal is to train the parking space vacancy rate prediction model by minimizing the difference between the predicted vacancy rate of each parking space and the label.

[0019] The present disclosure provides a computer-readable storage medium that stores a computer program, which, when executed by a processor, performs the method for predicting parking space vacancy rates.

[0020] The present disclosure provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the electronic device implements the method for predicting the vacancy rate of a parking space. [Effects of the Invention]

[0021] At least one of the above technical solutions used in the present disclosure can achieve the following beneficial effects:

[0022] In the method for predicting parking space vacancy rates provided by the present disclosure, the parking space vacancy rates of each parking lot within an area to be predicted prior to the time to be predicted are input into a feature extraction network of a parking space vacancy rate prediction model to obtain a first feature characterizing the relationship between the parking space vacancy rate of each parking lot and time. Furthermore, a spatial relationship graph between each parking lot within the area to be predicted and the first feature are input into a graph fusion network to obtain a fusion feature characterizing the relationship between the parking space vacancy rate of each parking lot and time and space. The fusion feature is then input into a result prediction network of the parking space vacancy rate prediction model to obtain the parking space vacancy rate of each parking lot within the area to be predicted at the time to be predicted. By simultaneously obtaining prediction results for the parking space vacancy rates of multiple parking lots within the area to be predicted based on the relationship between the parking space vacancy rate and time, the relationship between the parking space vacancy rate and space, and the potential interrelationship between the parking space vacancy rate and time and space, the efficiency and accuracy of the parking space vacancy rate prediction can be improved. [Brief explanation of the drawings]

[0023] The accompanying drawings described herein are used to further the understanding of the present disclosure and constitute a part of the present disclosure, and the exemplary embodiments of the present disclosure and the description thereof are used to explain the present disclosure and are not to be construed as unduly limiting the present disclosure.

[0024] [Figure 1] 1 is a schematic diagram showing the flow of a method for predicting parking space vacancy rates provided in the present disclosure. FIG. [Figure 2] FIG. 1 is a schematic diagram illustrating a parking space vacancy rate prediction model provided in the present disclosure. [Figure 3] 1 is a spatial relationship graph for each parking lot provided in the present disclosure. [Figure 4] 1 is a schematic diagram illustrating a parking space vacancy rate prediction device provided in the present disclosure. [Figure 5] 1 is a schematic diagram illustrating an electronic device provided in the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0025] In order to make the objectives, technical solutions and advantages of the present disclosure clearer, the technical solutions of the present disclosure will be clearly and completely described below in conjunction with specific embodiments of the present disclosure and the corresponding accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, but not all of them. All other embodiments that can be obtained by those skilled in the art based on the embodiments of the present disclosure without requiring creative efforts all belong to the protection scope of the present disclosure.

[0026] Below, the technical solutions provided in each embodiment of the present disclosure will be described in detail in conjunction with the accompanying drawings.

[0027] FIG. 1 is a schematic diagram showing the flow of the parking space vacancy rate prediction method provided in the present disclosure, which includes the following steps:

[0028] In S100, the area to be predicted and the time to be predicted are determined.

[0029] In S102, the vacancy rate of parking spaces at a plurality of times prior to the time to be predicted for each parking lot within the area to be predicted is acquired as the historical vacancy rate of each parking lot.

[0030] The prediction method of the present disclosure may be performed by any computing device (e.g., a server or a terminal) having computing capabilities. A pre-trained parking space vacancy prediction model is installed on the computing device. The parking space vacancy prediction model includes a feature extraction network, a graph fusion network, and an outcome prediction network, as shown in FIG. 2.

[0031] The computing device may determine the area to be predicted and the time to be predicted. The area to be predicted and the time to be predicted may be obtained from user input, i.e., the time to be predicted that the user wants to know and the parking space vacancy rate in the area to be predicted. Alternatively, the area to be predicted may be the area where the user is located at the current time, and the time to be predicted may be the current time. In other words, while the user is driving the vehicle, prediction can be performed in real time according to the user's driving route. Alternatively, the area to be predicted and the time to be predicted may be preset, and the area to be predicted and the time to be predicted may be determined based on a preset time length threshold and a geographical area.

[0032] Furthermore, to determine the relationship between the parking space vacancy rate of each parking lot in the area to be predicted and time using the parking space vacancy rate prediction model, the computing device may obtain the parking space vacancy rate of each parking lot in the area to be predicted at a time before the time to be predicted as the historical vacancy rate of each parking lot. For example, if the area to be predicted is Ward B in City A and the time to be predicted is 6:30 PM on April 19, the computing device can obtain the parking space vacancy rate of each parking lot in Ward B in City A before 6:30 PM on April 19. Assuming that there are three parking lots, X, Y, and Z, in Ward B of City A, the computing device can obtain the following: at 18:29 on April 19th, the vacancy rate of parking spaces in Parking Lot X is 20%, the vacancy rate of parking spaces in Parking Lot Y is 30%, and the vacancy rate of parking spaces in Parking Lot Z is 40%; and at 18:24 on April 19th, the vacancy rate of parking spaces in Parking Lot X is 19%, the vacancy rate of parking spaces in Parking Lot Y is 20%, and the vacancy rate of parking spaces in Parking Lot Z is 20%.

[0033] Here, when obtaining the parking space vacancy rate for each parking lot within the area to be predicted at a time before the time to be predicted, the parking space vacancy rate for each parking lot within the area to be predicted at a time before the time to be predicted may be obtained based on a preset time length. That is, a preset time length is determined and sampling is performed, and a sampling time is determined based on the preset time length and set before the time to be predicted, and the parking space vacancy rate for each parking lot at each sampling time is obtained. For example, if the preset time length is 5 minutes and the time to be predicted is 6:50 PM, the parking space vacancy rate for each parking lot at times such as 6:45 PM and 6:40 PM may be obtained.

[0034] The number of times before the time to be predicted for the parking space vacancy rate of each parking lot in the area to be predicted is not particularly limited, and may be acquired as needed. For example, if the input required for the parking space vacancy rate prediction model is the historical vacancy rate of each parking lot at five times before the time to be predicted, the historical vacancy rate of each parking lot at five times before the time to be predicted may be acquired. Furthermore, the vacancy rate of the parking space is determined based on all parking spaces and available (vacant) parking spaces in each parking lot.

[0035] Additionally, the parking space vacancy rate for each parking lot is calculated based on all parking spaces and available (i.e., empty) parking spaces in each parking lot.

[0036] In S104, the historical vacancy rate of each parking lot is input to the feature extraction network to obtain a first feature, and the first feature is the historical vacancy rate of each parking lot. of time dependence Used to characterize relationships.

[0037] The computing device may input the historical vacancy rate of each parking space into a feature extraction network of a parking space vacancy prediction model to obtain a first feature for characterizing the time dependency of the historical vacancy rate of each parking space.

[0038] In S106, a spatial relationship graph between each parking lot in the area to be predicted is constructed.

[0039] In one or more embodiments of the present disclosure, after determining the area to be predicted and the time to be predicted in step S100, a spatial relationship graph may be constructed between each parking lot in the area to be predicted to determine the relationship between the vacancy rate and space of the parking spaces in each parking lot in the area to be predicted.

[0040] Here, when constructing a spatial relationship graph, each parking lot in the area to be predicted may be treated as a node, and the distance between each parking lot in the area to be predicted may be treated as the weight of the edge. Following the example above, if there are three parking lots, X, Y, and Z, in Ward B of City A, and the distance from Parking Lot X to Parking Lot Y is 3 km, the distance from Parking Lot X to Parking Lot Z is 5 km, and the distance from Parking Lot Z to Parking Lot Y is 4 km, then a spatial relationship graph like the one shown in Figure 3 can be constructed.

[0041] Note that when constructing the spatial relationship graph, the distance between each parking lot is not the straight-line distance but the actual road distance. Based on an electronic map of the area to be predicted, the shortest road connecting two parking lots may be determined, and a road topology diagram between each parking lot may be obtained. Based on the road topology diagram, the distance between the two parking lots is determined, and this distance is used as the weight of the corresponding edge between the two parking lots in the spatial relationship graph. Of course, the driving time between the parking lots may also be used as the weight of the edge in the spatial relationship graph. As an example, and not by way of limitation, the weight may be the actual distance, driving time, or a comprehensive consideration of other influencing factors. For example, the actual distance or driving time between two parking lots may be calculated using a map tool, or may be calculated comprehensively by taking into account factors such as traffic congestion.

[0042] In constructing the spatial relationship graph, we may first use an adjacency matrix to represent the relationship between nodes (i.e., parking lots). The elements of the adjacency matrix represent the weight between two nodes.

number

[0043] At S108, the spatial relationship graph and the first feature are input into the graph fusion network to obtain a fused feature.

[0044] After obtaining first features for characterizing the relationship between the vacancy rate of parking spaces in each parking lot in the area to be predicted and time, and a spatial relationship graph for each parking lot in the area to be predicted, the computing device may input the spatial relationship graph and the first features into a graph fusion network to obtain fused features for characterizing the relationship between the vacancy rate of parking spaces in each parking lot and time and space.

[0045] Here, the graph fusion network may include an attention network and a graph convolutional network (GCN). The attention network enables the parking space vacancy prediction model to selectively pay attention to important parts. The graph convolutional network can combine the features of the spatial relationship graph and the first features to output a fusion feature.

[0046] The spatial relationship graph and the first features may be input to an attention network to obtain an attention-weighted output result, and the output result and the historical vacancy rate of each parking space may be input to a graph convolutional network to obtain a fused feature.

[0047] In addition, when the spatial relationship graph and the first feature are input to the attention network, since the spatial relationship graph represents the adjacency relationship using a matrix format and the attention network handles operations between matrices, the first feature is converted into a matrix format similar to the adjacency matrix represented by the spatial relationship graph, for example, by matrix multiplication of the first feature and the transpose of the first feature, and then the calculation result of the matrix multiplication and the subsequently calculated second feature matrix and spatial relationship graph are input to the attention network.

[0048] In S110, the fusion features are input into the result prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted.

[0049] The computing device may input the fusion features and the historical vacancy rate of each parking lot into the outcome prediction network to obtain the vacancy rate of parking spaces in each parking lot in the area to be predicted at the time to be predicted.

[0050] To explain using the example above, if the area to be predicted is Ward B in City A and the time to be predicted is 18:30 on April 19th, the computing device can obtain the parking space vacancy rate for each parking lot in Ward B in City A before 18:30 on April 19th. Assuming that Ward B in City A has three parking lots, X, Y, and Z, the computing device can obtain the following: at 18:29 on April 19th, the parking space vacancy rate for Parking Lot X is 20%, the parking space vacancy rate for Parking Lot Y is 30%, and the parking space vacancy rate for Parking Lot Z is 40%, and at 18:24 on April 19th, the parking space vacancy rate for Parking Lot X is 19%, the parking space vacancy rate for Parking Lot Y is 20%, and the parking space vacancy rate for Parking Lot Z is 20%. Assume that the vacancy rate of parking spaces in each parking lot (i.e., at 18:29 on April 19th, the vacancy rate of parking spaces in parking lot X is 20%, the vacancy rate of parking spaces in parking lot Y is 30%, and the vacancy rate of parking spaces in parking lot Z is 40%; at 18:24 on April 19th, the vacancy rate of parking spaces in parking lot X is 19%, the vacancy rate of parking spaces in parking lot Y is 20%, and the vacancy rate of parking spaces in parking lot Z is 20%) is input into the result prediction network, and the predicted results are as follows: at 18:30 on April 19th, the vacancy rate of parking spaces in parking lot X is 50%, the vacancy rate of parking spaces in parking lot Y is 70%, and the vacancy rate of parking spaces in parking lot Z is 40%.

[0051] In the parking space vacancy rate prediction method described above and provided in the present disclosure and shown in FIG. 1, the parking space vacancy rates of each parking lot in the area to be predicted at multiple times before the time to be predicted are input into a feature extraction network of a parking space vacancy rate prediction model to obtain a first feature for characterizing the time dependency of the parking space vacancy rate of each parking lot. Furthermore, a spatial relationship graph between each parking lot in the area to be predicted and the first feature are input into a graph fusion network to obtain a fusion feature for characterizing the parking space vacancy rate of each parking lot and its relationship with time and space. The fusion feature is then input into a result prediction network of the parking space vacancy rate prediction model to obtain the parking space vacancy rate of each parking lot in the area to be predicted at the time to be predicted. The parking space vacancy rate of a parking lot is not only related to time, but also to the spatial location of each parking lot. In other words, if there are differences in time and spatial location, the parking space vacancy rate of each parking lot will differ. Therefore, by simultaneously obtaining prediction results for the parking space vacancy rates of multiple parking lots within the area to be predicted based on the relationship between the parking space vacancy rate and time, the relationship between the parking space vacancy rate and space, and the potential interrelationship between the parking space vacancy rate and time and space, it is possible to improve the prediction efficiency of the parking space vacancy rate and the accuracy of the prediction results.

[0052] In one or more embodiments of the present disclosure, if the vacancy rates of parking spaces in each parking lot are the same at a certain time, the difference in the vacancy rates of parking spaces in the parking lots at the next time is generally not very large. To improve the accuracy of the prediction results, the similarity of the vacancy rates of parking spaces in each parking lot at the same time before the time to be predicted may be calculated to provide a hint to the model. This allows the model to determine the similarity of the vacancy rates of parking spaces in each parking lot at the time to be predicted based on the similarity of the vacancy rates of parking spaces in each parking lot at the same time before the time to be predicted.

[0053] The feature extraction network may include a first feature extraction network and a second feature extraction network. The first feature extraction network may extract a historical vacancy rate of each parking lot. of time dependence The first feature extraction network is used to extract a first feature for characterizing the relationship. The second feature extraction network is used to extract a second feature for characterizing the similarity of the historical vacancy rate of each parking lot. The historical vacancy rate of each parking lot is input to the first feature extraction network to obtain the first feature, and the historical vacancy rate of each parking lot is input to the second feature extraction network to obtain the second feature.

[0054] In one or more embodiments of the present disclosure, the first feature extraction network may be a Gated Recurrent Unit (GRU) and the second feature extraction network may be a Dynamic Time Warping (DTW).

[0055] GRU has Update Gated t and Reset Gated t Typically, the reset gate is used to determine how much old information is forgotten, while the update gate is used to determine what information needs to be forgotten and what new information needs to be added. t (i.e., the vacancy rate of parking spaces in each parking lot at time t) and the reset gate r t and the hidden state h at the previous time t-1 (t-1) and the candidate hidden state that indicates the information to be updated at time t.

number

number

number

number

number

[0056] In one or more embodiments of the present disclosure, a node represents a parking lot containing multiple parking spaces. Each parking lot has occupancy data at different times. For example, to predict the occupancy rate at time t=3, the DTW distance needs to be calculated using the occupancy rates between different parking lots at times t=1 and t=2. Therefore, in the second feature extraction network, the similarity of the historical vacancy rates between two nodes can be calculated using the following equation (5):

number

[0057] In the above step S108, the first feature, the second feature, and the spatial relationship graph may be input to a graph fusion network to obtain a fused feature. Specifically, the first feature, the second feature, and the spatial relationship graph may be input to an attention network of the graph fusion network to obtain an output result, and the output result may be further input to a graph convolution network to obtain a fused feature.

[0058] When an attention network is used, it is used to selectively focus on specific parts of the input sequence when generating the output. The basic principle of an attention network is to weight each input element according to its importance to the output of a particular step. Thus, the model can selectively focus on the most important parts of the input and ignore less important parts. The attention calculation is expressed as follows:

number

[0059] Next, when using a GCN network, the method of Spectral Graph Convolution is used to first calculate the normalized graph Laplacian matrix based on the fused features, which can be expressed as the following equation (7):

number

number

number

number

[0060] To further simplify the calculation and avoid overfitting, convolution is generally calculated using a first-order approximation of the graph Laplacian, and the final convolution layer operation can be expressed as Equation (9) below.

number

[0061] The attention network and GCN network can effectively capture spatiotemporal dependencies and improve prediction accuracy. Meanwhile, the attention network enables the parking space vacancy prediction model to selectively focus on important parts of the input sequence, thereby improving the generalization ability of the parking space vacancy prediction model.

[0062] Furthermore, the parking space vacancy rate of each parking lot is different under different environments, that is, the parking space vacancy rate of each parking lot is affected by the environment. Therefore, in one or more embodiments of the present disclosure, the parking space vacancy rate prediction model further includes an environmental feature extraction network.

[0063] Before inputting the fusion features into the result prediction network, the computing device may obtain environmental information for each parking lot in the area to be predicted before the time to be predicted, and input the environmental information into an environmental feature extraction network to obtain environmental features, and finally input the first features, the fusion features, the environmental features, and the historical vacancy rates of each parking lot into the result prediction network to obtain the vacancy rates of parking spaces in each parking lot in the area to be predicted at the time to be predicted.

[0064] Here, the environmental information includes at least weather information and holiday information.

[0065] Also, in one or more embodiments of the present disclosure, the environmental feature extraction network may be a GRU, and the outcome prediction network may be a multilayer perceptron (MLP).

[0066] Additionally, one or more embodiments of the present disclosure provide a method for training a parking space vacancy prediction model.

[0067] Specifically, the computing device may obtain the parking space vacancy rate of each parking lot in the specified area at at least two historical times, use the parking space vacancy rate of each parking lot at the latest of the at least two historical times as a label, use the parking space vacancy rate of each parking lot at the other historical times as a sample, input the sample into a feature extraction network to obtain sample features, construct a spatial relationship graph between each parking lot in the specified area as a sample spatial relationship graph, then input the sample spatial relationship graph and the sample features into a graph fusion network to obtain sample fusion features, and input the sample fusion features into a result prediction network to obtain a prediction result of the parking space vacancy rate of each parking lot at the latest historical time in the specified area, and finally train a parking space vacancy prediction model with the optimization goal of minimizing the difference between the prediction result of the parking space vacancy rate of each parking lot and the label, i.e., adjust parameters of the feature extraction network, graph fusion network, and result prediction network.

[0068] Here, the sample features are used to characterize the relationship between the historical vacancy rate and time for each parking lot.

[0069] When training a parking space vacancy rate prediction model, one input sample may be the parking space vacancy rates of each parking lot at multiple historical times.

[0070] In one or more embodiments of the present disclosure, when the feature extraction network includes a first feature extraction network and a second feature extraction network, the graph fusion network includes an attention network and a graph convolution network, and the parking space vacancy rate prediction model further includes an environmental feature extraction network, the sample may be first input into the first feature extraction network to obtain first sample features, the sample may be input into the second feature extraction network to obtain second sample features, and a sample spatial relationship graph between each parking lot in the specified area may be constructed, and then the dot product result of the first sample features and the first sample features, the second sample features, and the sample spatial relationship graph may be input into the attention network to obtain attention-weighted sample output results, and the sample output results may be input into the graph convolution network to obtain sample fusion features.

[0071] Alternatively, environmental information for each parking lot within the specified area may be acquired as sample environmental information, and the sample environmental information may be input to an environmental feature extraction network to obtain sample environmental features.

[0072] Finally, the sample fusion features, sample environment features, first sample features, and samples are input into the result prediction network to obtain the prediction result of the parking space vacancy rate of each parking lot at the latest historical time in the specified area.The parking space vacancy rate prediction model is trained with the optimization goal of minimizing the difference between the prediction result of the parking space vacancy rate and the label.

[0073] In other words, the optimization goal is to minimize the difference between the predicted vacancy rate of each parking space and the label, and the parameters of the first feature extraction network, second feature extraction network, attention network, graph convolution network, environmental feature extraction network, and outcome prediction network are adjusted.

[0074] In one or more embodiments of the present disclosure, the vacancy rate of parking spaces in each parking lot within the area to be predicted obtained by the prediction may be distributed to the cloud. A user may obtain the prediction result through communication between the cloud and a computing device without performing calculations and storage on the computing device. This can improve calculation efficiency and save storage capacity.

[0075] In addition, it is possible to predict the vacancy rate of parking spaces in each parking lot within the area to be predicted at multiple times in the future, allowing users to plan their travel routes in advance based on their travel needs and actual conditions, avoiding waste of time and resources and improving traffic efficiency and quality in the area to be predicted.

[0076] Based on the above-mentioned method for predicting parking space vacancy rate, the embodiment of the present disclosure further provides a corresponding device for predicting parking space vacancy rate, as shown in FIG.

[0077] FIG. 4 is a schematic diagram of a parking space vacancy rate prediction device provided in the present disclosure, in which a pre-trained parking space vacancy rate prediction model is configured, and the parking space vacancy rate prediction model includes a feature extraction network, a graph fusion network, and an outcome prediction network, and the device: a determination module 402 for determining an area to be predicted and a time to be predicted; an acquisition module 406 for acquiring the parking space vacancy rate of each parking lot within the area to be predicted at a plurality of times prior to the time to be predicted as the historical vacancy rate of each parking lot; a first input module 408 for inputting the historical vacancy rate of each parking lot into the feature extraction network to obtain first features, the first features being used to characterize the time dependency of the historical vacancy rate of each parking lot; a construction module 404 for constructing a spatial relationship graph between parking lots in the area to be predicted; a second input module 410 for inputting the spatial relationship graph and the first features into the graph fusion network to obtain a fused feature; and an output module 414 for inputting the fused features into the outcome prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted.

[0078] Optionally, the first input module 408 specifically: inputting the historical vacancy rate of each of the parking spaces into the feature extraction network to obtain second features, the second features being used to characterize the similarity of the historical vacancy rates of each of the parking spaces; The second input module 410 specifically includes: The spatial relationship graph, the first feature, and the second feature are used to input into the graph fusion network.

[0079] Optionally, the construction module 404 specifically: The spatial relationship graph is constructed by using each parking lot in the area to be predicted as a node and the distance between each parking lot in the area to be predicted as an edge weight.

[0080] Optionally, the graph fusion network comprises an attention network and a graph convolutional network; The second input module 410 specifically includes: inputting the spatial relationship graph, the first feature, and the second feature into the attention network and obtaining an attention-weighted output result; The output results and the historical vacancy rate of each parking lot are used to input into the graph convolutional network.

[0081] Optionally, the parking space vacancy rate prediction model further comprises an environmental feature extraction network; The device further includes an environmental information input module 412; The environmental information input module 412 specifically includes: acquiring environmental information for each parking lot within the area to be predicted before the time to be predicted, the environmental information including at least weather information and holiday information, and inputting the environmental information into the environmental feature extraction network to obtain environmental features; The output module 414 specifically includes: The first features, the fusion features, the environmental features, and the historical vacancy rate of each parking lot are used to input into the outcome prediction network.

[0082] Optionally, the apparatus further comprises a training module 400; Specifically, the training module 400: Obtain the parking space vacancy rate for at least two historical times for each parking lot within the specified area; The vacancy rate of the parking spaces of each parking lot at the latest historical time among the at least two historical times is used as a label, and the vacancy rate of the parking spaces of each parking lot at the other historical times is used as a sample; inputting the samples into the feature extraction network to obtain sample features, the sample features being used to characterize the relationship between historical vacancy rate and time for each of the parking spaces; constructing a spatial relationship graph between each parking lot in the specified area as a sample spatial relationship graph; inputting the sample spatial relationship graph and the sample features into the graph fusion network to obtain sample fusion features; inputting the sample fusion features into the result prediction network to obtain a prediction result of the vacancy rate of each parking space at the latest historical time in the specified area; The optimization goal is to train the parking space vacancy rate prediction model by minimizing the difference between the predicted vacancy rate of each parking space and the label.

[0083] The present disclosure further provides a computer-readable storage medium, which stores a computer program, and the computer program is used to perform the method for predicting parking space vacancy rate described above.

[0084] Based on the above-described method for predicting parking space vacancy rate, the present disclosure further provides an electronic device as shown in Fig. 5. As shown in Fig. 5, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, an internal memory, and a non-volatile memory, and may also include hardware required for other operations. The processor loads a corresponding computer program from the non-volatile memory into the internal memory and executes it to implement the above-described method for predicting parking space vacancy rate.

[0085] Of course, in addition to software implementation, this disclosure does not exclude other implementation methods, such as logical devices or a combination of hardware and software. In other words, the entity that executes the following processing process is not limited to each logical unit, but may also be hardware or a logical device.

[0086] In the 1990s, technological improvements could be clearly divided into hardware improvements (such as improvements to circuit structures like diodes, transistors, and switches) and software improvements (such as improvements to methodology). However, with technological advances, many current methodology improvements can be considered direct improvements to hardware circuit structures. Designers often obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it is not necessarily true that methodology improvements cannot be realized using physical hardware modules. For example, programmable logic devices (PLDs) (e.g., field programmable gate arrays, or FPGAs) are such integrated circuits, whose logical functions are determined by user programming of the device. Instead of chip manufacturers designing and manufacturing dedicated integrated circuit chips, designers program and "integrate" digital systems onto a single PLD.Nowadays, instead of handcrafting integrated circuit chips, this programming is often achieved using software called a "logic compiler," which is similar to a software compiler used to write a program. To compile the original code, it must be written in a specific programming language called a Hardware Description Language (HDL). There is not just one HDL; there are many, including ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). The most commonly used HDLs are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. It should be clear to those skilled in the art that a hardware circuit embodying a logical method flow can be easily obtained by simply logically programming the method flow in one of the above hardware description languages ​​and programming it into an integrated circuit.

[0087] The controller may be implemented in any suitable manner. For example, the controller may take the form of a microprocessor or processor, a computer-readable storage medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, microcontrollers such as the ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. A memory controller may also be implemented as part of the memory's control logic. It will be apparent to those skilled in the art that, in addition to implementing the controller purely with computer-readable program code, it is entirely possible to logically program method steps to cause the controller to perform the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, and the like. Therefore, such a controller may be considered a hardware component, and the devices included therein for implementing various functions may also be considered structures within the hardware component. Alternatively, the apparatus for realizing various functions may be considered to be software modules that implement a method, or structures within hardware components.

[0088] The systems, devices, modules, or units described in the above embodiments may be specifically realized by a computer chip, an entity, or a product having some function. A typical realizing device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a mobile phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet, a wearable device, or any combination of these devices.

[0089] For ease of explanation, the above-described device will be described by dividing it into various units according to their functions. Of course, when implementing the present disclosure, the functions of each unit may be realized by the same or multiple pieces of software and / or hardware.

[0090] As will be appreciated by those skilled in the art, embodiments of the present disclosure may be provided as a method, a system, or a computer program product. Thus, the present disclosure may take the form of an embodiment consisting entirely of hardware, an embodiment consisting entirely of software, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product embodied in one or more computer-usable storage media (including, but not limited to, magnetic disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0091] The present disclosure will be described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, whereby the instructions, executed by the processor of the computer or other programmable data processing device, generate an apparatus for implementing the functions specified in one or more flows in the flowcharts and / or one or more blocks in the block diagrams.

[0092] These computer program instructions may be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture that includes an instruction apparatus that implements the functions specified in one or more flows of the flowcharts and / or one or more blocks of the block diagrams.

[0093] These computer program instructions may be loaded into a computer or other programmable data processing device, whereby a series of operational steps are executed on the computer or other programmable device to generate a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows of the flowcharts and / or one or more blocks of the block diagrams.

[0094] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0095] The memory may include volatile memory, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM), among other computer-readable storage media. The memory is one example of a computer-readable storage medium.

[0096] Computer-readable storage media include nonvolatile and volatile media, removable and non-removable media, and may implement information storage by any method or technology. Information may be computer-readable instructions, data structures, program modules, or other data. Computer storage media include, but are not limited to, Phase Change Memory (PRAM), Static Random-Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read-Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), flash memory or other memory technology, Compact Disc Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, magnetic cassette tape, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. As defined herein, computer readable storage media does not include transitory computer readable storage media, such as modulated data signals and carriers.

[0097] Additionally, the terms "comprise," "contain," or any other variation thereof, are intended to include a non-exclusive inclusion, whereby a process, method, article, or device that includes a set of elements not only includes those elements, but also includes other elements not expressly listed, or includes the inherent elements of such process, method, article, or device. Absent more limitations, elements qualified by the phrase "comprise..." do not exclude the presence of further identical elements in a process, method, article, or device that includes said elements.

[0098] As will be appreciated by those skilled in the art, embodiments of the present disclosure may be provided as a method, a system, or a computer program product. Thus, the present disclosure may take the form of an embodiment consisting entirely of hardware, an embodiment consisting entirely of software, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product embodied in one or more computer-usable storage media (including, but not limited to, magnetic disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0099] The present disclosure may be described in the general context of computer-executable instructions being executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present disclosure may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0100] Each embodiment in the present disclosure will be described in a stepwise manner, and the same or similar parts between each embodiment may be referred to, and the description of each embodiment will focus on the differences from other embodiments. In particular, since the system embodiment is basically similar to the method embodiment, it will be briefly described, and the relevant parts may be referred to the description of part of the method embodiment.

[0101] The above is merely an embodiment of the present disclosure and is not used to limit the present disclosure. Those skilled in the art will appreciate that the present disclosure may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made without departing from the spirit and principle of the present disclosure should be included in the scope of the claims of the present disclosure.

Claims

1. A method for predicting parking space vacancy rate applied to a computing device, the computing device being configured with a pre-trained parking space vacancy rate prediction model, the parking space vacancy rate prediction model including a feature extraction network, a graph fusion network, and an outcome prediction network; determining an area to be predicted and a time to be predicted; acquiring, as a historical vacancy rate for each parking lot, the vacancy rate of a parking space at a plurality of times prior to the time to be predicted for each parking lot within the area to be predicted; inputting the historical vacancy rate of each parking lot into the feature extraction network to obtain first features, the first features being used to characterize the time dependency of the historical vacancy rate of each parking lot; inputting the historical vacancy rate of each of the parking spaces into the feature extraction network to obtain second features, the second features being used to characterize the similarity of the historical vacancy rates of each of the parking spaces; constructing a spatial relationship graph between parking lots in the area to be predicted; inputting the spatial relationship graph, the first feature, and the second feature into the graph fusion network to obtain a fused feature; inputting the fusion features into the result prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted; A method for predicting the vacancy rate of a parking space, comprising:

2. The step of constructing a spatial relationship graph between each parking lot in the area to be predicted includes: constructing the spatial relationship graph by using each parking lot in the area to be predicted as a node and distances between each parking lot in the area to be predicted as edge weights, 2. The method of claim 1 .

3. The graph fusion network includes an attention network and a graph convolutional network; inputting the spatial relationship graph, the first feature, and the second feature into the graph fusion network, inputting the spatial relationship graph, the first feature, and the second feature into the attention network to obtain an attention-weighted output result; inputting the output result and the historical vacancy rate of each parking lot into the graph convolutional network; 2. The method of claim 1 .

4. The parking space vacancy rate prediction model further includes an environmental feature extraction network; before inputting the fused features into the outcome prediction network; acquiring environmental information for each parking lot within the area to be predicted before the time to be predicted, the environmental information including at least weather information and holiday information; inputting the environmental information into the environmental feature extraction network to obtain environmental features; The step of inputting the fusion features into the outcome prediction network includes: further comprising inputting the first feature, the fusion feature, the environmental feature, and the historical vacancy rate of each parking lot into the outcome prediction network; 2. The method of claim 1 .

5. The parking space vacancy rate prediction model is Obtaining the parking space vacancy rate for at least two historical times for each parking lot within the designated area; a step of labeling the vacancy rate of parking spaces in each parking lot at the latest of the at least two historical times, and sampling the vacancy rate of parking spaces in each parking lot at the other historical times; inputting the samples into the feature extraction network to obtain sample features, the sample features being used to characterize the relationship between historical vacancy rate and time for each of the parking spaces; constructing a spatial relationship graph between each parking lot in the designated area as a sample spatial relationship graph; inputting the sample spatial relationship graph and the sample features into the graph fusion network to obtain a sample fusion feature; inputting the sample fusion features into the result prediction network to obtain a prediction result of the vacancy rate of each parking space at the latest historical time in the specified area; and training the parking space vacancy rate prediction model with an optimization goal of minimizing the difference between the predicted vacancy rate of each parking space and the label.

2. The method of claim 1 .

6. A parking space vacancy rate prediction device, wherein a pre-trained parking space vacancy rate prediction model is disposed in the parking space vacancy rate prediction device, and the parking space vacancy rate prediction model includes a feature extraction network, a graph fusion network, and an outcome prediction network; a determination module for determining an area to be predicted and a time to be predicted; an acquisition module for acquiring the parking space vacancy rate of each parking lot within the area to be predicted at a plurality of times prior to the time to be predicted as a historical vacancy rate of each parking lot; a first input module for inputting the historical vacancy rate of each parking lot into the feature extraction network to obtain a first feature, and for inputting the historical vacancy rate of each parking lot into the feature extraction network to obtain a second feature, wherein the first feature is used to characterize a time dependency of the historical vacancy rate of each parking lot, and the second feature is used to characterize a similarity of the historical vacancy rate of each parking lot; a construction module for constructing a spatial relationship graph between parking lots in the area to be predicted; a second input module for inputting the spatial relationship graph, the first feature, and the second feature into the graph fusion network to obtain a fused feature; an output module for inputting the fusion features into the outcome prediction network to obtain the vacancy rate of parking spaces in each parking lot within the area to be predicted at the time to be predicted; A device for predicting parking space vacancy rates.

7. A computer-readable storage medium storing a computer program, which, when executed by a processor, performs the method of any one of claims 1 to 5. A computer-readable storage medium comprising:

8. An electronic device comprising a memory, a processor, and a computer program stored in said memory and executable on said processor, said computer program being adapted to perform the method of any one of claims 1 to 5 when executed by said processor. An electronic device characterized by:

Images