On-road parking resource allocation method considering automatic driving vehicle

Abstract

The invention relates to the field of intelligent parking, in particular to a curb parking resource allocation method considering an automatic driving vehicle. Comprising establishing a road network model; generating a travel demand of each time period; planning a shortest path for the target vehicle to reach the road intersection closest to the destination, calculating the total correlation degree of available parking spaces by taking the driving distance, the parking price and the walking distance as attribute indexes, and determining the optimal parking space corresponding to the target vehicle according to the vehicle type and the parking space state; the optimal parking space of the HVs is the idle parking space with the maximum total correlation degree in the common parking spaces; the optimal parking space of the AVs is the parking space with the maximum total correlation degree and the optimal common parking space parking cost and the optimal PUDO parking space empty driving cost; and adjusting the proportion of the number of the PUDO parking spaces to the number of the common parking spaces, and optimizing the parking space configuration of the parking lot on the road by taking minimization of the total travel time as a target. The number of common parking spaces and PUDO parking spaces in the on-road parking area is reasonably configured so as to adapt to the parking behavior characteristics in the AVs application process.

Description

Technical Field

[0001] The present application relates to the field of smart parking, and more specifically, to a method for allocating on-street parking resources taking into account autonomous driving vehicles. Background Art

[0002] Against the backdrop of the rapid development of autonomous driving and smart urban transportation, traditional urban traffic management systems face new requirements for transformation and upgrading. Autonomous vehicles (AVs), with capabilities such as autonomous cruising and remote parking, can independently select a suitable parking location upon arrival. This feature not only reshapes parking models but also drives a surge in demand for "stop-and-go" passenger pickup and drop-off, profoundly impacting the functional layout and management mechanisms of existing on-street parking spaces.

[0003] Existing parking resource planning approaches often focus on meeting parking demand, improving facility utilization, and reducing system travel costs. They utilize multi-objective and bi-level planning methods to construct static supply-demand equilibrium models and achieve optimal spatial allocation of parking resources. These parking facility spatial allocation methods are primarily designed based on static parking demand distribution and vehicle parking habits, but lack the ability to respond to the short-term, high-frequency boarding and alighting behaviors of AVs. The boarding and alighting behaviors of AVs not only alter parking demand characteristics but also significantly impact the utilization of on-street parking resources. Therefore, research is needed to develop on-street parking resource planning methods that consider the boarding and alighting needs of AVs to achieve efficient utilization of on-street parking resources and improve travel efficiency within transportation networks. Summary of the Invention

[0004] In view of this, the present application provides a method for allocating on-street parking resources taking into account autonomous vehicles. By means of simulation optimization, the existing on-street parking space configuration scheme is replanned. The passenger boarding and disembarking processes involved in AVs carrying travelers to complete travel activities are considered, and the number of ordinary parking spaces and PUDO parking spaces in the on-street parking area is reasonably allocated to adapt to the parking behavior characteristics of AVs during application, thereby achieving the sustainable development of on-street parking resources.

[0005] The technical solutions provided in this application are as follows:

[0006] A method for allocating on-street parking resources considering autonomous vehicles, comprising:

[0007] A method for allocating on-street parking resources considering autonomous driving vehicles, characterized by comprising:

[0008] Establish a road network model and configure the total number of parking spaces, parking prices, and parking space categories in on-street parking lots, including regular parking spaces and PUDO parking spaces;

[0009] Divide historical data into multiple time periods and use normal distribution to generate travel demand for each period, including departure point, destination, vehicle type, and activity time;

[0010] The shortest path is planned for the target vehicle to reach the nearest intersection to the destination. The total correlation of available parking spaces is calculated using the driving distance from the target vehicle to the parking space, the parking price of the parking space, and the walking distance of the passenger from the parking space to the destination as attribute indicators. The optimal parking space for the target vehicle is determined based on the vehicle type and parking space status. The optimal parking space for human-driven vehicles (HVs) is the vacant parking space with the highest total correlation among ordinary parking spaces. The optimal parking space for autonomous vehicles (AVs) is the parking space with the highest total correlation and the optimal combination of the parking cost of ordinary parking spaces and the vacant driving cost of PUDO parking spaces.

[0011] The target vehicle's location, cumulative travel distance, vehicle status, and parking space occupancy status are updated in real time. By dynamically calculating the cumulative time spent by AVs / HVs in various states, the ratio of PUDO parking spaces to regular parking spaces is adjusted to optimize the parking space configuration in on-street parking lots with the goal of minimizing total travel time.

[0012] In one possible implementation, a road network model is established to configure the total number of parking spaces, parking prices, and parking space categories in on-street parking lots, including:

[0013] Obtaining scene information corresponding to the on-street parking lot and initializing the road grid and parking space status of the on-street parking lot; wherein the scene information includes the distribution layout of the on-street parking lot, the total number of parking spaces, and the parking status of each parking space;

[0014] Construct a multi-row and multi-column grid road network model, set the distance between adjacent intersections to a fixed distance, and set a destination node at the center of each block;

[0015] On-street parking spaces are distributed evenly or according to preset rules on both sides of the road network. The total number of parking spaces is set and divided into multiple areas. Each area contains continuously distributed parking spaces, and different parking prices are set for the parking spaces in each area.

[0016] In one possible implementation, the total relevance of available parking spaces is calculated using the driving distance from the target vehicle to the parking space, the parking price of the parking space, and the walking distance from the parking space to the destination as attribute indicators, including:

[0017] Construct a set of available parking spaces in the area, perform dimensionless processing on the above attribute indicators based on the preset standardized formula, and calculate the entropy weight of each attribute indicator, which is expressed as:

[0018]

[0019] in, is the value of the i-th parking space on the j-th attribute index; r′ ij is the standardized attribute value; J1 is the benefit index, J2 is the cost index; η′ j is the information entropy of the jth indicator; I is the number of available parking spaces; is the entropy weight of the j-th indicator;

[0020] The ideal solution set is determined based on the principle that the larger the benefit-type index, the better, while the smaller the cost-type index, the better. The correlation coefficient between each parking space and the ideal solution is calculated based on the grey correlation coefficient formula, and the initial total correlation degree is obtained by weighted entropy weight, which is expressed as:

[0021]

[0022] Where rj is the ideal value and ρ is the resolution coefficient.

[0023] In one possible implementation, for manually driven HVs, determining the optimal parking space for the target vehicle based on the vehicle type and parking space status includes:

[0024] The initial total correlation is modified according to the parking space type and vacancy status, and is expressed as:

[0025]

[0026] Where s i Represents the occupancy status of the parking space, 0 if occupied and 1 if free; M i c Used to indicate whether parking space i meets the needs of HVs; N i represents the initial total correlation;

[0027] Select the vacant parking space with the highest correlation from the ordinary parking spaces as the optimal parking space for the manually driven vehicle, expressed as:

[0028]

[0029] In one possible implementation, for autonomous vehicles (AVs), determining the optimal parking space for a target vehicle based on the vehicle type and parking space status includes:

[0030] The initial total correlation is modified based on the lowest parking cost of a regular parking space and the cost of driving without a passenger in a PUDO parking space to determine the optimal parking space for the autonomous vehicle, including:

[0031] If the autonomous vehicle AVs chooses to park in a regular parking space, the parking cost is the parking cost during the activity time, which is expressed as:

[0032]

[0033] Where, represents the estimated parking cost of autonomous vehicle a; p i represents the unit price of parking space i;

[0034] If the autonomous vehicle AVs chooses to wait for passengers to get off in the PUDO parking space and then enters the free driving phase with no passengers, the cost is the free driving cost during the activity time, which is expressed as:

[0035]

[0036] Where, represents the cost of autonomous vehicle a choosing to drive automatically; θ represents the travel cost per unit distance; Represents the average speed of roads within a 1.5 km radius; u a represents the activity time of the autonomous driving vehicle a;

[0037] The free driving distance during the activity time is determined by the average speed of the surrounding road segments and is expressed as:

[0038]

[0039] Where Υ(r, d) represents the Manhattan distance between two points; |r i -d i | represents the absolute value of the coordinate difference between point r and point d in the i-th dimension.

[0040] In one possible implementation, after determining the optimal parking space corresponding to the target vehicle based on the vehicle type and the parking space status, the method further includes:

[0041] When there are no available parking spaces, manually driven vehicles / autonomous driving vehicles enter a cruising state and detect available parking spaces in real time;

[0042] If an available parking space is detected, the BFS algorithm is used to dynamically plan a path to the corresponding optimal parking space.

[0043] In one possible implementation, after determining the optimal parking space corresponding to the target vehicle based on the vehicle type and the parking space status, the method further includes:

[0044] For manually driven vehicles (HVs) and autonomous vehicles (AVs) that choose to park in ordinary parking spaces, when the vehicle arrives at the corresponding optimal parking space, the parking timer is started and the actual parking time is monitored in real time until the vehicle's predefined activity time is reached;

[0045] For autonomous vehicles (AVs) that choose PUDO parking spaces, when the vehicle arrives at the target PUDO parking space and completes the passenger drop-off operation, the free travel timer is started and the vehicle travels randomly in the road network until the free travel time reaches the active time.

[0046] In one possible implementation, the real-time updating of the target vehicle's vehicle position, cumulative travel distance, vehicle status, and parking space occupancy status includes:

[0047] When the vehicle's cumulative driving distance exceeds the preset threshold, the route is switched, the vehicle's cruising and empty mileage / time are classified and counted, the road section flow and time are calculated in real time through the BPR function, and the parking space occupancy status is adjusted synchronously.

[0048] One possible implementation involves adjusting the ratio of PUD0 spaces to regular parking spaces to optimize the on-street parking configuration with the goal of minimizing total travel time, and also includes:

[0049] An optimization solution framework is constructed based on the simulated annealing algorithm. The ratio of PUDO parking spaces is used as the optimization parameter, and minimizing the total travel time is the optimization goal. The algorithm parameters are set, and the optimization is iterated and optimized until the solution end conditions are met, and then the optimal PUDO parking space configuration solution is output.

[0050] In one possible implementation, the method for optimizing parking space configuration in an on-street parking lot further includes:

[0051] Compare the three strategies of proportional matching strategy, static allocation strategy and dynamic allocation strategy, and determine the optimal allocation strategy based on comprehensive benefit indicators such as vehicle mileage, time and parking space utilization rate;

[0052] Among them, the comprehensive benefit indicators include system benefit indicators and parking lot benefit indicators; the system benefit indicators include total vehicle mileage VKT, invalid mileage VEK, total travel time VTT, and invalid travel time VET; the parking lot benefit indicators include PUDO parking space utilization rate, ordinary parking space utilization rate, total parking space utilization time TUT, and ordinary parking space revenue.

[0053] Compared with the existing technology, the technical solution provided by this application has the following beneficial effects:

[0054] By dynamically dividing the ratio of PUD0 parking spaces to regular parking spaces, combining the unladen driving characteristics of autonomous vehicles with the parking behavior differences of manually driven vehicles, and utilizing a gray entropy correlation model to optimize parking space selection by integrating multiple attributes such as driving distance, parking price, and walking distance, the system effectively reduces inefficient cruising and unladen mileage. By dynamically adjusting parking space allocation strategies using a simulated annealing algorithm and integrating comparative analysis of three strategies, namely proportional correspondence, static, and dynamic, the system achieves efficient adaptation of parking resources to fluctuating demand over multiple time periods, significantly improving parking space utilization and road network efficiency. Furthermore, based on real-time vehicle status tracking and a BPR road condition dynamic feedback mechanism, the system accurately calculates key indicators such as total travel time and inefficient energy consumption, providing data-driven optimization decision support for urban on-street parking management. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flowchart of a method for allocating on-street parking resources considering autonomous driving vehicles, provided in Example 1 of the present application.

[0056] Figure 2 A flowchart of a method for allocating on-street parking resources considering autonomous driving vehicles is provided in Example 2 of the present application.

[0057] Figure 3 This is a schematic diagram of the on-street parking road network model constructed in Example 2 of this application.

[0058] Figure 4 This is a flowchart of a method for searching for the best target parking space based on total correlation provided in the second embodiment of the present application.

[0059] Figure 5 A flowchart of another method for allocating on-street parking resources considering autonomous driving vehicles, provided in Example 2 of the present application.

[0060] Figure 6 This is a comparison chart of the PUDO parking space configuration solutions in three areas under the PCS, SC, and DC strategies provided in Example 2 of this application.

[0061] Figure 7 This is a schematic diagram of the detailed comparison of several evaluation indicators under different strategies provided in Example 2 of this application. DETAILED DESCRIPTION

[0062] The following will combine the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0063] To address the issues identified in the background, this application discloses a method for allocating on-street parking resources for autonomous vehicles (AVs). Considering that AVs carry passengers through the pickup and / or drop-off (PUDO) process, a simulation optimization approach is used to convert some on-street parking spaces into temporary PUDO spaces to accommodate the parking behavior characteristics of AVs. The method primarily includes the following steps: Inputting information required for on-street parking simulation in an autonomous driving scenario, including on-street parking space supply information and traveler demand information. Considering the varying demands of autonomous vehicles for on-street parking spaces, a multi-agent-based on-street parking space selection simulation framework is constructed. This framework simulates the interaction between vehicle parking and parking space occupancy, and outputs system efficiency and parking space operational efficiency evaluation parameters. Based on the simulation results, a method for allocating different types of on-street parking spaces is established. An optimization model is then solved to determine the number of parking spaces allocated for each type.

[0064] Below, the specific implementation steps of the method provided by this application are described in detail by way of specific examples.

[0065] Example 1

[0066] See also Figure 1 , is a flow chart of a method for allocating on-street parking resources considering autonomous driving vehicles provided in Example 1 of this application. Figure 1 As shown in , the specific implementation steps of the above method include:

[0067] Step 101: Establish a road network model and configure the layout of on-street parking spaces, parking prices, and parking space categories.

[0068] Step 102: Divide the historical data into multiple time periods and use normal distribution to generate travel demand for each time period, including departure point, destination, vehicle type and activity time.

[0069] Step 103: Plan the shortest path for the target vehicle to reach the road intersection closest to the destination, calculate the total relevance of available parking spaces using driving distance, parking price, and walking distance as attribute indicators, and determine the best parking space corresponding to the target vehicle based on the vehicle type and parking space status.

[0070] Human-driven vehicles (HVs) select only the vacant spaces with the highest overall correlation among regular parking spaces. Autonomous vehicles (AVs) comprehensively evaluate the parking costs of regular spaces and the cost of idling in PUDO spaces, selecting the spaces with the highest correlation and the best cost as the optimal parking spaces for the AVs.

[0071] Step 104: When there is no available parking space for the target vehicle, the target vehicle enters a cruising state and performs parking space detection in real time until an available parking space is detected.

[0072] Step 105: Update the target vehicle's location, mileage, vehicle status, and parking space occupancy status in real time. By dynamically calculating the cumulative time spent by AVs / HVs in various states, adjust the ratio of PUDO parking spaces to regular parking spaces, and optimize the on-street parking space configuration with the goal of minimizing total travel time.

[0073] Compared with the prior art, the technical solution provided in Example 1 of the present application has the following beneficial effects:

[0074] By dynamically dividing the ratio of PUD0 parking spaces to regular parking spaces, combining the unladen driving characteristics of autonomous vehicles with the parking behavior differences of manually driven vehicles, and utilizing a gray entropy correlation model to optimize parking space selection by integrating multiple attributes such as driving distance, parking price, and walking distance, the system effectively reduces inefficient cruising and unladen mileage. By dynamically adjusting parking space allocation strategies using a simulated annealing algorithm and integrating comparative analysis of proportional, static, and dynamic strategies, the system achieves efficient allocation of parking resources under multi-period demand fluctuations, significantly improving parking space utilization and road network efficiency. Furthermore, based on real-time vehicle status tracking and a BPR road condition dynamic feedback mechanism, the system accurately calculates key indicators such as total travel time and inefficient energy consumption, providing data-driven optimization decision support for urban on-street parking management.

[0075] Example 2

[0076] See also Figure 2 , is a flow chart of a method for allocating on-street parking resources considering autonomous driving vehicles, provided in Example 2 of this application. Figure 2 As shown in , the specific implementation steps of the above method include:

[0077] Step 201: Acquire scene information corresponding to the on-street parking lot and initialize the road grid and parking space status of the on-street parking lot.

[0078] Specifically, the above-mentioned scene information includes but is not limited to the distribution layout of on-street parking lots, the total number of parking spaces, and the parking status of each parking space.

[0079] like Figure 3As shown in , based on the above-mentioned scenario information of on-street parking obtained, the present embodiment constructs a multi-row and multi-column grid road network model, in which the distance between adjacent intersections is set to a fixed distance, and a destination node is set at the center of each block. On-street parking spaces are distributed evenly or according to preset rules on both sides of the road network. The total number of parking spaces is set and divided into multiple areas (such as Area 1, Area 2, and Area 3), each of which contains continuously distributed parking spaces.

[0080] Furthermore, this application also sets differentiated parking prices for each area (such as area 1 rate p1 = 4 yuan / hour, area 2 rate p2 = 6 yuan / hour, area 3 rate p3 = 8 yuan / hour).

[0081] Based on the above parking space distribution, parking space status and rate parameters, an initialization database containing parking space location coordinates, area affiliation, real-time status and rate is constructed for dynamic deduction of parking behavior in a simulation environment.

[0082] Step 202: Obtain the travel demand of the target vehicle.

[0083] Specifically, this embodiment of the application considers the differences in parking demand at different times of the day and divides the target vehicle's activity time range into several preset time periods. For each preset time period, a normal distribution function is used to generate the number of trips required for that time period. The mean and variance of the normal distribution are set based on historical demand characteristics during that time period.

[0084] For example, in a specific application scenario, 14 hours in a day are used as the activity time of the travelers, and the above activity time is divided into 5 preset time periods, including before the morning peak (BM, 6:00-8:00), the morning peak (MP, 8:00-10:00), the off-peak (OP, 10:00-17:00), the evening peak (EP, 17:00-19:00) and after the evening peak (AE, 19:00-20:00). Furthermore, as shown in Table 1, the embodiment of the present application classifies travelers into three types according to their travel purpose: work, shopping, and visiting. The demand proportion of each travel type in different time periods is determined based on a preset proportion parameter. The activity duration is generated for each traveler, where the work type travel activity time obeys the N(360, 60) normal distribution, the shopping type travel activity time obeys the N(120, 30) normal distribution, and the visiting type travel activity time obeys the N(20, 5) normal distribution.

[0085] Table 1 Traveler demand information

[0086]

[0087] Furthermore, this application randomly selects a proportion of k vehicles from the generated travel demand and defines them as autonomous vehicles AVs, where k is selected from the set {0.1, 0.3, 0.5, 0.7, 0.9}, and marks the target vehicles with proportion k as autonomous vehicles (b j =1), and the rest are marked as manually driven vehicles (b j =0). In the pre-built gridded road network model, each target vehicle j is randomly assigned a starting point coordinate o j and the end point coordinate d j , where d j Prioritize the available parking area closest to the destination node.

[0088] Step 203: According to the travel demand of the target vehicle, find the intersection closest to the destination, and determine the driving path of the target vehicle based on the breadth-first algorithm.

[0089] Specifically, when a vehicle needs to park, it first drives toward the intersection closest to the destination. Manhattan distance traversal is used to calculate the nearest intersection to the destination. Upon arrival at the intersection, a parking space is selected based on the vehicle type.

[0090] Step 204: When the target vehicle arrives at the nearest intersection, a parking space is selected based on the driving distance from the target vehicle to the parking space, the parking price of the parking space, and the walking distance of the passenger from the parking space to the destination as attribute indicators to search for the best target parking space.

[0091] This application selects three main attributes: driving distance, parking price, and walking distance as attribute indicators for parking space selection. The attribute values ​​of each available parking space i in the area are standardized, with driving distance and parking price defined as cost-based indicators, and walking distance as a benefit-based indicator. With the goal of minimizing the cost-based indicator and minimizing the benefit-based indicator, the optimal parking space is selected for the target vehicle. HVs select vacant ordinary parking spaces that meet the type constraints, while AVs dynamically select either parking in an ordinary parking space or driving in an empty space after connecting to a PUDO parking space, combining parking cost and unladen free travel cost optimization strategies.

[0092] Specifically, if Figure 4 As shown in , the above step 204 may further include:

[0093] Step 2041: Construct a solution set of available parking spaces in the area, perform dimensionless processing on the above attribute indicators based on a preset standardized formula, and calculate the entropy weight of each of the above attribute indicators.

[0094] Specifically, pre-defined is the solution set for parking space decision, and the attribute set of each parking space in the area is The evaluation values ​​of different parking space selection schemes on attributes are

[0095] Taking into account the differences in the numerical dimensions of attribute indicators, the data are standardized according to formulas (1) and (2), expressed as:

[0096]

[0097] This application converts the original attribute values ​​of driving distance, price, and walking distance of different dimensions into dimensionless standardized evaluation values ​​through a preset standardized formula, thereby forming a solution set matrix of available parking spaces.

[0098] Furthermore, the information entropy and weight of the attribute index j are calculated according to formulas (3) and (4). In this embodiment, the three indicators of driving distance, parking price, and walking distance are selected to calculate the entropy weight, that is, J=3.

[0099]

[0100] in, is the value of the i-th parking space on the j-th attribute index; r′ ij is the standardized attribute value; J1 is the benefit index, J2 is the cost index; η′ j is the information entropy of the jth indicator; I is the number of available parking spaces; is the entropy weight of the j-th indicator.

[0101] Step 2042: Determine the ideal solution set based on the principle that the larger the benefit-type indicator, the better, and the smaller the cost-type indicator, the better. Calculate the correlation coefficient between each parking space and the ideal solution based on the grey correlation coefficient formula, and obtain the initial total correlation degree through weighted entropy weighting.

[0102] Specifically, the ideal parking solution set is determined based on the principle that the larger the benefit index is, the better, while the smaller the cost index is, the better, and is expressed as: * ={r j * |r j * =max i r ij or r j * =min i r ij}.

[0103] Calculate the grey correlation coefficient according to formula (5) Expressed as:

[0104]

[0105] Using formula (6), the initial total correlation degree of parking space i is expressed as:

[0106]

[0107] Among them, r j * is the ideal value. ρ is the resolution coefficient, which takes values ​​in [0, 1].

[0108] Step 2043: For the manually driven vehicle HVS, the initial total correlation is modified according to the parking space type and vacancy status, and the vacant parking space with the highest correlation is selected from the ordinary parking spaces as the optimal parking space for the manually driven vehicle.

[0109] For manually driven HVs, the embodiment of the present application only allows the selection of ordinary parking spaces. By correcting the total correlation, the vacant parking space with the highest correlation is selected from the ordinary parking spaces as the best parking space for parking.

[0110] Specifically, for manually driven vehicles (HVs), the corrected total correlation degree is calculated according to formula (7) and is expressed as:

[0111]

[0112] Define m j 1 represents an ordinary parking space, m j If it is 0, it is a PUDO parking space. HVs can only choose ordinary parking spaces for parking. Its parking space type selection satisfies formula (8), which is expressed as:

[0113]

[0114] The parking space with the largest grey correlation degree after correction is selected as the final choice for the HVs vehicle, which is expressed as:

[0115]

[0116] in, Represents the corrected association between the manually driven vehicle c and the parking space i. i Represents the occupancy status of the parking space, 0 if occupied and 1 if free. i c It is used to indicate whether parking space i meets the needs of manually driven vehicle c. represents the optimal parking space selected by the manually driven vehicle c.

[0117] S2044. For autonomous vehicles (AVs), the above-mentioned initial total correlation is modified based on the lowest principle of the parking cost of ordinary parking spaces and the unladen driving cost of PUDO parking spaces to determine the optimal parking space for the autonomous vehicle.

[0118] In this embodiment, for autonomous driving vehicles (AVs), it is allowed to select ordinary parking spaces or PUDO parking spaces, correct the initial total correlation, and select a parking mode based on the parking cost optimization principle.

[0119] The total correlation degree of the corrected autonomous driving vehicle a to the parking space i is calculated according to formula (10) and is expressed as:

[0120]

[0121] Since the vehicle has the ability to drive automatically, in addition to choosing to park in an ordinary parking space, it can also drive freely without any load instead of parking. i a The calculation of is shown in formula (11), which is expressed as:

[0122]

[0123] If the autonomous vehicle a chooses to park in an ordinary parking space like the manually driven vehicles HVs, the parking cost is the parking cost during the activity time, which can be estimated according to formula (12) and expressed as:

[0124]

[0125] Alternatively, autonomous vehicle a chooses to wait for passengers to get off the PUDO parking space and then enters the free driving phase with no load. The cost is the free driving cost during the active time, as shown in formula (13), which is expressed as:

[0126]

[0127] It should be noted that the free driving distance during the activity time is estimated based on the average speed of the surrounding road segments. The free driving distance is estimated using the average speed of the surrounding road segments within 1.5 km of Manhattan, as shown in formulas (14) and (15). AVs select one of the parking modes based on the principle of minimum cost.

[0128]

[0129] If the autonomous driving vehicle a selects the normal parking mode (i.e. satisfies ), the autonomous vehicle a will choose a regular parking space when selecting a parking space. Otherwise, if the autonomous vehicle a chooses to wait for the passengers to get off in the PUDO parking space and then enters the free driving mode (i.e., it satisfies ), the autonomous vehicle a will choose the PUDO parking space type, as shown in formula (11). Finally, the parking space with the largest gray correlation degree after correction is selected as the final choice of the autonomous vehicle a, as shown in formula (16), which is expressed as:

[0130]

[0131] in, represents the corrected association degree of the autonomous driving vehicle a with the parking space i; M i a Used to indicate whether parking space i meets the needs of autonomous vehicle a; represents the estimated parking cost of autonomous vehicle a; p i represents the unit price of parking space i; u a represents the activity time of the autonomous vehicle a; Υ(r, d) represents the Manhattan distance between two points; represents the cost of autonomous vehicle a choosing to drive automatically; θ represents the travel cost per unit distance; Represents the average speed of roads within a 1.5 km radius; i -d i | represents the absolute value of the coordinate difference between point r and point d in the i-th dimension; represents the best parking space selected by autonomous vehicle a.

[0132] Furthermore, if the manually driven vehicle / autonomous driving vehicle does not find the best parking space that meets the requirements among the current parking spaces, the vehicle enters a cruise mode until a parking space that meets the requirements is found, which specifically includes the following steps:

[0133] Step 205: When there is no available parking space, the manually driven vehicle / autonomous driving vehicle enters a cruising state and detects available parking spaces in real time. If an available parking space is detected, the BES algorithm is used to dynamically plan a path to the corresponding optimal parking space.

[0134] When the vehicle is in the cruising stage, the autonomous vehicle a is driven to the state x at and the state x of the manually driven vehicle c ct are all set to 0, and then the vehicles randomly drive along the road direction and cruising range. At each time t in the cruising phase, both vehicles can re-determine in real time whether there is a parking space that meets the requirements based on the parking space selection process. If so, they set the temporary destination as that parking space, continue using the BFS algorithm to obtain a set of reachable paths, and drive along that path until they reach the parking space. If not, they continue cruising.

[0135] Step 206: For manually driven vehicles (HVs) and autonomous vehicles (AVs) that choose to park in ordinary parking spaces, when the vehicle arrives at the corresponding optimal parking space, the parking timer is started to monitor in real time whether the actual parking time reaches the vehicle's predefined activity time. If so, the vehicle is triggered to exit the system.

[0136] Step 207: For the AVs that choose to park in a PUDO parking space, after the vehicle arrives at the target PUDO parking space and completes the passenger drop-off operation, the vehicle starts the free travel timer and travels randomly in the road network until the free travel time reaches the active time.

[0137] Step 208: When the free travel time reaches the active time, the AVs are controlled to automatically navigate to a PUDO parking space near the destination, and the vehicle exits the system after completing the passenger pickup operation.

[0138] In the embodiment of this application, Figure 5 As shown in the figure, when the vehicle is in the parking phase, for HVs or AVs that choose to park in ordinary parking spaces, when the vehicle arrives at the parking space, the parking time of both types of vehicles begins to count down until the parking time reaches the vehicle activity time, at which point the vehicle completes the parking process and exits the system. For AVs that choose to drive freely but require the PUDO process, when the vehicle arrives at the PUDO parking space, it waits for the vehicle to complete the drop-off and then begins to drive randomly along the road. At this time, the vehicle is in an unloaded state, and the driving state is changed to x t If it is set to 2, until the free travel time reaches the active time, the AVs will automatically go to the vicinity of the destination and select a PUDO parking space to pick up the passenger. After completing the passenger picking process, the AVs will exit the system.

[0139] Step 209: Dynamically update the vehicle's position, status, and driving parameters. When the vehicle's cumulative driving distance exceeds a preset threshold, perform route switching, classify and count the vehicle's cruising and empty mileage / time, calculate the road section flow and time in real time through the BPR function, and synchronously adjust the parking space occupancy status.

[0140] In the embodiment of the present application, at each time t in the simulation, the position and status of the vehicle and the parking space status information are updated, and the cumulative driving distance l of the vehicle on the current road section is updated. at or l ct Among them, l at represents the cumulative driving distance of the autonomous driving vehicle a on the current road section, l c Indicates the cumulative driving distance of the manually driven vehicle c on the current road section. If this distance is greater than the length of the current road section l r , then drive to the next node in the reachable path set, and at or l ct Reset to 0 and update the total mileage L of all autonomous vehicles and manually driven vehicles at and L ct , the total travel time of all autonomous vehicles and manually driven vehicles B at and B ct If the autonomous vehicle or the manually driven vehicle is in the cruise state (x at or x ct=0), update the vehicle's total cruising mileage L at cruise and L ct cruise , total cruising time B at cruise and B ct cruise If the vehicle is in an unloaded state (x at 2), update the vehicle's total empty mileage B at empty , total no-load driving time B at empty Update the road flow q according to the BPR function rt , road section travel time e rt , road speed v rt If there is a vehicle occupying the parking space, update the parking space status s it .

[0141] Step 210: By dynamically calculating the cumulative time of AVs / HVs in various states, the ratio of PUDO parking spaces to ordinary parking spaces is adjusted to optimize the parking space configuration of the on-street parking lot with the goal of minimizing the total travel time.

[0142] In this application, the optimization parameter is to set the number of PUDO parking spaces in each on-street parking area. By configuring the number of PUDO parking spaces and ordinary parking spaces, the parking selection of vehicles and the travel efficiency of the system can be affected. Specifically, from the system perspective, minimizing the total travel time is selected as the optimization goal. The travel time of the autonomous vehicle a at time t and the travel time of the manually driven vehicle c at time t are defined as b, b, and c, respectively. at and b ct , then the optimization goal of minimizing the total travel time can be obtained according to formula (17). Define the state of the autonomous driving vehicle a as x at , using values ​​0, 1, and 2 to represent x at In the empty, loaded and parked states; define the state of manually driven vehicle c as x ct , using values ​​0, 1, and 2 to represent x ct In cruise, non-cruise and parking state, b at and b ct They can be obtained using formulas (18) and (19) respectively.

[0143]

[0144] Step 211: Calculate the optimal PUDO parking space configuration solution based on a simulated annealing algorithm.

[0145] In the embodiment of the present application, an optimization solution framework is constructed based on the simulated annealing algorithm, and the above optimization objective (minimizing the total travel time) and optimization parameters (the ratio of the number of PUDO parking spaces) are input into the above optimization solution framework. The algorithm parameters are set, including the initial temperature T0 is set to 10, the end temperature T f The algorithm continues to select and optimize until the solution condition is met and the optimal PUDO parking space configuration solution is output.

[0146] Step 212: Set a proportional correspondence strategy, a static allocation strategy, and a dynamic allocation strategy for comparison, and determine the optimal allocation strategy based on vehicle mileage, time, and parking space utilization as comprehensive efficiency indicators.

[0147] The Proportional Matching Strategy (PCS) simply allocates PUDO parking spaces based on the market share of AVs. The Static Allocation Strategy (SC) uses a fixed allocation of on-street parking spaces to achieve optimization goals, based on a prior understanding of traveler demand. The Dynamic Allocation Strategy (DC) periodically updates the parking allocation strategy based on travel demand and parking space usage over time, periodically optimizing the number of PUDO parking spaces to achieve the ultimate goal.

[0148] The aforementioned comprehensive efficiency indicators include system efficiency indicators and parking lot efficiency indicators. System efficiency indicators include total vehicle mileage (VKT), inefficient mileage (VEK) (including HVs cruising and AVs idle mileage), total travel time (VTT), and inefficient travel time (VET). Parking lot efficiency indicators include PUDO parking space utilization rate, general parking space utilization rate, total parking space utilization time (TUT), and general parking space revenue (TPR).

[0149] According to the above steps, the number of PUDO parking spaces in three areas under the PCS, SC, and DC strategies are as follows: Figure 6 As shown in the figure, the results show that under the SC strategy, when the proportion of AVs is at a low level, the number of PUDO parking spaces in the three areas is small and the difference is not obvious; when the proportion of AVs is at a high level (greater than 0.5), the number of PUDO parking spaces in the three areas increases with the increase of parking prices in the area; under the DC allocation strategy, the BM and MP stages have a medium level of PUDO parking spaces, followed by fewer PUDO parking spaces in the OP stage, and more PUDO parking spaces in the EP and AE stages. Figure 7 A detailed comparison of several evaluation metrics under different strategies is presented. The results show that VKT and VEK increase with the proportion of AVs under all three strategies. Compared with the PCS strategy, the SC and DC strategies can also significantly reduce VTT and VET, with the DC strategy achieving a greater reduction than the SC strategy.

[0150] Compared with the prior art, the technical solution provided in Example 2 of the present application has the following beneficial effects:

[0151] This application significantly improves the accuracy of parking space selection and the rationality of system resource allocation by constructing a gray entropy correlation correction model and a multi-attribute decision-making mechanism, combining the optimization of unloaded driving costs for autonomous vehicles with parking space type constraints for manually driven vehicles. A dynamic simulation framework tracks vehicle status in real time, updates road network parameters, and synchronizes parking space occupancy information, effectively reducing vehicle cruising and unloaded driving mileage and inefficient travel time. Furthermore, a simulated annealing algorithm is used to optimize the PUDO parking space allocation ratio. A multi-strategy comparative evaluation mechanism (PCS, SC, DC) integrates multiple dimensions, including vehicle driving efficiency, parking space utilization, and operating revenue, to achieve efficient dynamic adaptation of on-street parking resources.

[0152] Although the embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for allocating on-street parking resources considering autonomous vehicles, characterized in that: include: Establish a road network model and configure the total number of parking spaces, parking prices, and parking space categories in on-street parking lots, including regular parking spaces and PUDO parking spaces; Divide historical data into multiple time periods and use normal distribution to generate travel demand for each period, including departure point, destination, vehicle type, and activity time; The shortest path is planned for the target vehicle to reach the nearest intersection to the destination. The total correlation of available parking spaces is calculated using the driving distance from the target vehicle to the parking space, the parking price of the parking space, and the walking distance of the passenger from the parking space to the destination as attribute indicators. The optimal parking space for the target vehicle is determined based on the vehicle type and parking space status. The optimal parking space for human-driven vehicles (HVs) is the vacant parking space with the highest total correlation among ordinary parking spaces. The optimal parking space for autonomous vehicles (AVs) is the parking space with the highest total correlation and the optimal combination of the parking cost of ordinary parking spaces and the vacant driving cost of PUDO parking spaces. The target vehicle's location, cumulative travel distance, vehicle status, and parking space occupancy status are updated in real time. By dynamically calculating the cumulative time spent by AVs / HVs in various states, the ratio of PUDO parking spaces to regular parking spaces is adjusted to optimize the parking space configuration in on-street parking lots with the goal of minimizing total travel time.

2. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: Establish a road network model and configure the total number of parking spaces, parking prices, and parking space categories in on-street parking lots, including: Obtaining scene information corresponding to the on-street parking lot and initializing the road grid and parking space status of the on-street parking lot; wherein the scene information includes the distribution layout of the on-street parking lot, the total number of parking spaces, and the parking status of each parking space; Construct a multi-row and multi-column grid road network model, set the distance between adjacent intersections to a fixed distance, and set a destination node at the center of each block; On-street parking spaces are distributed evenly or according to preset rules on both sides of the road network. The total number of parking spaces is set and divided into multiple areas. Each area contains continuously distributed parking spaces, and different parking prices are set for the parking spaces in each area.

3. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: The total relevance of available parking spaces is calculated using the driving distance from the target vehicle to the parking space, the parking price of the parking space, and the walking distance of the passenger from the parking space to the destination as attribute indicators, including: Construct a set of available parking spaces in the area, perform dimensionless processing on the above attribute indicators based on the preset standardized formula, and calculate the entropy weight of each attribute indicator, which is expressed as: in, is the value of the i-th parking space on the j-th attribute index; r′ ij is the standardized attribute value; J1 is the benefit index, J2 is the cost index; η′ j is the information entropy of the jth indicator; I is the number of available parking spaces; is the entropy weight of the j-th indicator; The ideal solution set is determined based on the principle that the larger the benefit-type index, the better, while the smaller the cost-type index, the better. The correlation coefficient between each parking space and the ideal solution is calculated based on the grey correlation coefficient formula, and the initial total correlation degree is obtained by weighted entropy weight, which is expressed as: Among them, r * j is the ideal value, and ρ is the resolution coefficient.

4. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: For HVs, the optimal parking space for the target vehicle is determined based on the vehicle type and parking space status, including: The initial total correlation is modified according to the parking space type and vacancy status, and is expressed as: Where s i Represents the occupancy status of the parking space, 0 if occupied and 1 if free; M i c Used to indicate whether parking space i meets the needs of HVs; N i represents the initial total correlation; Select the vacant parking space with the highest correlation from the ordinary parking spaces as the optimal parking space for the manually driven vehicle, expressed as:

5. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: For autonomous vehicles (AVs), the optimal parking space for the target vehicle is determined based on the vehicle type and parking space status, including: The initial total correlation is modified based on the lowest parking cost of a regular parking space and the cost of driving without a passenger in a PUDO parking space to determine the optimal parking space for the autonomous vehicle, including: If the AVs choose to park in a regular parking space, the parking cost is the parking cost during the activity time, which is expressed as: Where, represents the estimated parking cost of autonomous vehicle a; p i represents the unit price of parking space i; If the autonomous vehicle AVs chooses to wait for passengers to get off in the PUDO parking space and then enters the free driving phase with no passengers, the cost is the free driving cost during the activity time, which is expressed as: Where, represents the cost of autonomous vehicle a choosing to drive automatically; θ represents the travel cost per unit distance; Represents the average speed of roads within a 1.5 km radius; u a represents the activity time of the autonomous driving vehicle a; The free driving distance during the activity time is determined by the average speed of the surrounding road segments and is expressed as: γ(r,d)=|r1-d1|+|r2-d2|+...|r n -d n | (15) Where γ(r, d) represents the Manhattan distance between two points; i -d i | represents the absolute value of the coordinate difference between point r and point d in the i-th dimension.

6. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: After determining the optimal parking space corresponding to the target vehicle according to the vehicle type and the parking space status, the method further includes: When there are no available parking spaces, manually driven vehicles / autonomous driving vehicles enter a cruising state and detect available parking spaces in real time; If an available parking space is detected, the BFS algorithm is used to dynamically plan a path to the corresponding optimal parking space.

7. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: After determining the optimal parking space corresponding to the target vehicle according to the vehicle type and the parking space status, the method further includes: For manually driven vehicles (HVs) and autonomous vehicles (AVs) that choose to park in ordinary parking spaces, when the vehicle arrives at the corresponding optimal parking space, the parking timer is started and the actual parking time is monitored in real time until the vehicle's predefined activity time is reached; For autonomous vehicles (AVs) that choose PUDO parking spaces, when the vehicle arrives at the target PUDO parking space and completes the passenger drop-off operation, the free travel timer is started and the vehicle travels randomly in the road network until the free travel time reaches the active time.

8. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: Real-time updates of the target vehicle's location, cumulative travel distance, vehicle status, and parking space occupancy status, including: When the vehicle's cumulative driving distance exceeds the preset threshold, the route is switched, the vehicle's cruising and empty mileage / time are classified and counted, the road section flow and time are calculated in real time through the BPR function, and the parking space occupancy status is adjusted synchronously.

9. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: Adjusting the ratio of PUDO parking spaces to regular parking spaces to optimize on-street parking configuration with the goal of minimizing total travel time, including: An optimization solution framework is constructed based on the simulated annealing algorithm. The ratio of PUDO parking spaces is used as the optimization parameter, and minimizing the total travel time is the optimization goal. The algorithm parameters are set, and the optimization is iterated and optimized until the solution end conditions are met, and then the optimal PUDO parking space configuration solution is output.

10. The method for allocating on-street parking resources considering autonomous driving vehicles according to claim 1, characterized in that: Optimizing parking space configuration in an on-street parking lot, the method further comprising: Compare the three strategies of proportional matching strategy, static allocation strategy and dynamic allocation strategy, and determine the optimal allocation strategy based on comprehensive benefit indicators such as vehicle mileage, time and parking space utilization rate; Among them, the comprehensive benefit indicators include system benefit indicators and parking lot benefit indicators; the system benefit indicators include total vehicle mileage VKT, invalid mileage VEK, total travel time VTT, and invalid travel time VET; the parking lot benefit indicators include PUDO parking space utilization rate, ordinary parking space utilization rate, total parking space utilization time TUT, and ordinary parking space revenue.

Images