Parking lot idle state estimation method and system and vehicle

Abstract

The invention relates to the technical field of parking lot monitoring, in particular to a parking lot idle state estimation method and system and a vehicle, and the estimation method comprises the steps: obtaining a sensing data packet from a crowdsourcing vehicle in a target parking lot, and determining whether the sensing data packet is an effective sensing data packet; based on each effective sensing data packet, calculating a local idle parking space proportion of the effective sensing data packet and a first weight of the effective sensing data packet in weighted fusion calculation; and generating a global idle parking space proportion representing the overall idle state of the target parking lot through weighted fusion calculation based on the local idle parking space proportion and the first weight of the plurality of effective sensing data packets. Through obtaining the sensing data packet from crowdsourcing vehicles, the problems of high cost and insufficient coverage of deployment depending on a fixed sensor of a parking lot are avoided, and automatic parking space state monitoring which does not need large-scale infrastructure reconstruction and does not depend on manual intervention is realized.

Description

Technical Field

[0001] This application relates to the field of parking lot monitoring technology, and in particular to a method, system and vehicle for estimating the vacancy status of a parking lot. Background Technology

[0002] With the continuous growth of car ownership in my country, urban parking resources are becoming increasingly strained, making parking a prominent issue affecting travel experience and traffic efficiency. Drivers often struggle to obtain accurate real-time parking availability information when heading to their destinations, leading to aimless parking searches, exacerbating surrounding traffic congestion, and even forcing them to change their itineraries due to full parking lots, severely reducing travel convenience and time utilization. Therefore, providing drivers with real-time and accurate parking space availability information has become an urgent need to be addressed in the fields of intelligent transportation and smart mobility.

[0003] Currently, the main way to obtain parking space status is to deploy sensors such as geomagnetic, ultrasonic, or camera sensors inside the parking lot to monitor the occupancy status of each parking space in real time and publish the number of available spaces through in-park displays or network platforms.

[0004] However, the existing technical solutions mentioned above have limitations. Although they can ensure real-time data, they require high costs for hardware procurement, installation, and subsequent maintenance. As a result, they can only cover a few large parking lots that have completed smart transformation, and cannot be adapted to open-air parking lots, old residential parking lots, and other areas without the conditions for smart transformation, making it difficult to meet the needs of comprehensive parking status monitoring. Summary of the Invention

[0005] This application provides a parking lot vacancy estimation method, system, and vehicle. Its purpose is to address the limitations of existing parking lot vacancy information acquisition schemes by providing a parking lot vacancy estimation scheme, solving the problem of unknown parking space information when users travel, and improving travel convenience and traffic efficiency.

[0006] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application provides a method for estimating the vacancy status of a parking lot, comprising: Acquire perception data packets from crowdsourced vehicles located in the target parking lot, and determine whether the perception data packets are valid perception data packets; Based on each of the aforementioned valid sensing data packets, the local vacant parking space ratio of that valid sensing data packet and its first weight in the weighted fusion calculation are calculated. Based on the local vacant parking space percentage and its first weight of multiple valid sensing data packets, a global vacant parking space percentage representing the overall vacancy status of the target parking lot is generated through the weighted fusion calculation.

[0007] This method can obtain perception data packets from crowdsourced vehicles, avoiding the high cost and insufficient coverage of relying on fixed sensors in parking lots. It also overcomes the shortcomings of manual reporting or historical data prediction, such as update lag and poor accuracy, and realizes automated parking space status monitoring without large-scale infrastructure transformation or human intervention.

[0008] By implementing effective verification rules for sensing data packets, off-site data, expired data, and abnormal interference data can be eliminated, avoiding the interference of invalid information on the global estimation results. By calculating the local vacant parking space ratio through local parking space information in a single effective sensing data packet, the local scanning results of a single vehicle are transformed into standardized data that can be directly used for fusion, solving the data integration problem caused by inconsistencies in the scanning range and number of different vehicles. Based on crowdsourced local data from multiple vehicles, this solution integrates scattered sensing information through a weighted fusion algorithm, avoiding the problem of limited scanning range of a single vehicle. This allows the global vacant parking space ratio to dynamically approximate the actual vacancy status of the target parking lot, solving the problem in existing technologies where local data cannot reflect the overall situation and the prediction results deviate significantly from the actual state.

[0009] In some embodiments of this application, the effective sensing data packet includes at least local parking space information sensed and statistically analyzed by the vehicle, as well as corresponding data time information; Based on the local parking space information, determine whether the vehicle belongs to the target parking lot; based on the data time information, determine whether the sensing data packet is within the preset validity period; if both determinations are yes, then it is the valid sensing data packet.

[0010] The local parking space information in this application has regional scene attributes, which can reflect whether the physical location of the vehicle is within the target parking lot. It eliminates invalid data scanned outside the parking lot and avoids global proportion estimation errors caused by cross-regional data mixing. The parking lot space status has high-frequency change characteristics, with frequent vehicle entry and exit. Expired data can no longer reflect the current real status. Such data can be directly discarded through timeliness verification to avoid misleading results caused by using past data to estimate the current status. The data time information is linked to the subsequent timeliness weight calculation. It is related to the time decay coefficient λ. The data retained after timeliness verification can obtain weight in the weighted fusion, making the global estimation result more consistent with the real-time status of the parking lot.

[0011] In some schemes of this application, the second weight is determined based on the local parking space information. The second weight of the m-th data It is the ratio of the total number of scanning parking spaces corresponding to the valid sensing data packet to the sum of the total number of scanning parking spaces corresponding to all valid sensing data packets.

[0012] The second weight in this application is used to determine the sample size reflected by the local parking space information. By using the proportion of the total number of scanned parking spaces as the basis for weight allocation, the effective sensing data packets with a wider scanning range are given higher weights, thereby improving the accuracy and reliability of the parking lot vacancy state estimation results. It also reasonably balances the contribution of different data, avoids excessive influence of local small-scale scanning data on the overall estimation results, enhances the robustness of the vacancy state estimation model, and better adapts to practical application scenarios where the scanning range of crowdsourced vehicles varies greatly.

[0013] The second weight is based on the total number of parking spaces scanned in a single valid data point. It converts the data's information coverage into a weighted proportion by comparing its own total scan count with the sum of all valid data scan counts. A single valid data point with a larger total number of scanned parking spaces indicates a wider local perception coverage for the vehicle, and the corresponding local parking space information is more representative of the overall parking lot status, thus receiving a higher second weight. This avoids the unreasonable situation where narrow, partial data and comprehensive data with wide scan counts are weighted equally, ensuring that the data contribution matches its actual information value, and guaranteeing the rationality of data utilization through weight allocation.

[0014] The limited coverage of localized scanning by crowdsourced vehicles means that a single vehicle cannot scan all parking spaces in a parking lot at once. However, the calculation logic of the second weight can overcome this limitation by weighted integration of data from multiple vehicles: effective data with a wider scanning range has a higher second weight and accounts for a larger proportion in subsequent fusion calculations, and the parking space information it covers can more fully reflect the local state of the parking lot. After the scanning data from multiple vehicles is weighted according to the coverage range, the scattered local perception information can be gradually pieced together into a more complete global parking space status, offsetting the problem of narrow scanning range and one-sided information of a single vehicle, and making the estimated result of the global vacant parking space ratio closer to the actual situation.

[0015] The calculation of this second weight relies solely on two values: the total number of parking spaces scanned in a single data entry and the sum of the total number of scans of all valid data entries. This offers the advantage of low computational power consumption. It eliminates the need for complex feature extraction, model training, or scene recognition algorithms, enabling rapid weight calculation for a single data entry. Adaptable to scenarios with massive amounts of crowdsourced vehicles uploading data in real time, it can synchronously iterate weights as data is updated, avoiding delays caused by complex calculations and ensuring the real-time nature of global idle state estimation.

[0016] In some schemes of this application, a third weight is determined based on the data time information, and the third weight is calculated as follows:

[0017] in, As the third weight, The time decay coefficient is a preset value and 0 < λ < 1. For the current time, The data collection time is the m-th data point indicated by the data time information.

[0018] The third weight in this application reflects the timeliness determined by the data's timing information. Parking space status is dynamically changing, and older data cannot reflect the current situation. Time difference. The larger the value, the older the data. Combined with the attenuation coefficient (0 < λ < 1), The smaller the result, the higher the weight; conversely, the closer the data collection time is to the current time, the smaller the time difference, and the higher the weight.

[0019] The third weight complements the second weight based on the total number of scanned parking spaces. When the two are combined, effective data with a wide scanning range and newer data will receive a higher composite weight. This avoids the excessive proportion of perceived information with a large coverage but outdated data, and also avoids the limited impact of information with new data but a narrow coverage. By selecting high-value data based on the spatial coverage and timeliness of the data, the weighted fusion result is more in line with the real global state of the parking lot.

[0020] In some schemes of this application, the composite weight of the m-th valid perceived data packet in the weighted fusion calculation is determined, and the composite weight is the second weight. With the third weight The product of the values; and the first weight is obtained by normalizing the composite weights of all valid sensed data. .

[0021] This application's composite weighting is approved. This approach links the sample size with the timeliness of data: only valid data that simultaneously meets the criteria of a large total number of scanned parking spaces and new data collection time can obtain a higher first weight; this avoids the limitations of a single-dimensional weight and ensures that the contribution of each data point matches its actual information value.

[0022] After normalization, the sum of the composite weights of all valid data is 1, ensuring that the subsequent calculation of the global vacant parking space ratio always falls within the range of 0% to 100%, avoiding abnormal results caused by excessive weight accumulation. The normalized data weights are all under a uniform quantification scale, making the weighted fusion results of different amounts of data comparable and consistent, thus improving the stability of the algorithm.

[0023] In some embodiments of this application, the weighted fusion calculation is implemented using the following formula:

[0024] in, The percentage of available parking spaces globally. The first weight is the weight corresponding to the m-th valid sensing data packet. The percentage of available parking spaces corresponds to the m-th valid sensing data packet.

[0025] The percentage of vacant parking spaces in each valid data point in this application represents the vacancy status of the vehicle scanning area. Combined with the differentiated allocation of the first weight, the scattered vacancy information of local areas can be stitched together to cover the entire parking lot. This avoids the problem that the scanning range of a single vehicle is limited and local data cannot reflect the whole situation, so that the global vacancy percentage R can approximate the true vacancy status of the target parking lot.

[0026] In some schemes of this application, the percentage of locally vacant parking spaces is... It is the ratio of the number of available parking spaces in the sensing data packet described in the m-th entry to the total number of scanned parking spaces represented by the local parking space information described in the m-th entry.

[0027] This application transforms the local perception data of a single crowdsourced vehicle into a standardized idle status indicator by using the ratio of the number of vacant parking spaces to the total number of scanned parking spaces. By uniformly quantifying the idle level of local areas using a scale of 0 to 1, the impact of differences in scan scale on data comparability is eliminated; this places local perception data from different vehicles under the same measurement dimension, laying a standardized data foundation for weighted fusion calculations of multiple vehicles. This reflects the actual idle rate within the scanning area of ​​a single crowdsourced vehicle.

[0028] In some embodiments of this application, the calculation of scanning coverage is also included, wherein the scanning coverage is the ratio of the sum of the total number of scanning parking spaces corresponding to all the valid sensing data packets to the total number of parking spaces in the target parking lot.

[0029] This application quantifies the coverage of crowdsourced data to a target parking lot by using the ratio of the total number of valid parking spaces scanned to the total number of parking spaces in the parking lot. If the scan coverage is high, it indicates that the valid data has covered most of the parking spaces in the parking lot, and the global vacant parking space ratio obtained by weighted fusion has higher data support reliability; if the scan coverage is low, it indicates that the valid data only covers a small area, and the estimated result of the global ratio may be biased and cannot reflect the status of the uncovered areas. When the scanning coverage is low, the system can specifically incentivize crowdsourced vehicles in the target parking lot to upload data, or prioritize the use of valid data with a large number of scanned parking spaces, gradually expanding the data coverage. As the scanning coverage increases, the area coverage of valid data becomes more complete, and the estimation error of the global vacant parking space ratio will decrease accordingly, forming a virtuous cycle of more accurate results from improved coverage, and achieving dynamic self-optimization of the technical solution.

[0030] The second aspect provides a parking lot vacancy state estimation system, including: The data receiving module is used to acquire sensing data packets located in the target parking lot; The data verification module is used to determine whether the sensing data packet is a valid sensing data packet; The weight calculation module is used to calculate the local vacant parking space ratio of each valid sensing data packet and its first weight in the weighted fusion calculation based on each valid sensing data packet. The fusion computing module is used to generate a global vacant parking space percentage that represents the overall vacant status of the target parking lot through the weighted fusion computing based on the local vacant parking space percentage and its first weight of multiple valid sensing data packets.

[0031] The data receiving module of this application is responsible for data input, the data verification module is responsible for data filtering, the weight calculation module is responsible for data value standardization, and the fusion calculation module is responsible for global result generation. From the upload of the sensing data packet to the output of the global idle ratio, no manual intervention is required, which improves the efficiency of data processing, avoids errors caused by manual operation, and ensures the stability and consistency of system operation.

[0032] By leveraging vehicle visual perception capabilities, there is no need to deploy additional dedicated hardware such as fixed cameras in parking lots, saving the procurement, installation, and maintenance costs of traditional solutions and lowering the barrier to system implementation. It is not limited by whether the parking lot has completed intelligent transformation. Even in old parking lots or open-air parking lots, as long as crowdsourced vehicles enter the site and upload data, it breaks through the scene coverage limitations of traditional fixed sensor solutions.

[0033] In addition, this application also provides a vehicle including the parking lot vacancy estimation system described in the second aspect.

[0034] After the vehicle in this application integrates this system, it can directly reuse its own widely available intelligent sensing hardware vision camera to autonomously complete the collection operations of scanning local parking space information, counting the number of vacant spaces, and generating sensing data packets. There is no need to install additional dedicated hardware or for users to manually submit data. This avoids the hardware costs and manual operation barriers of crowdsourced data collection, allowing each intelligent vehicle to become a crowdsourced node for parking lot status monitoring, thus expanding the base of data contribution participants.

[0035] Once equipped with this system, vehicles not only contribute data but also directly obtain the global vacancy rate output by the system. By combining their own location information, vehicles can autonomously plan better parking routes, avoid blindly circling around to find a parking space, and shorten parking time. If the global vacancy rate is extremely low, the system can alert the vehicle in advance that the parking lot is about to be full, helping users to adjust their destination in time and reduce unnecessary trips.

[0036] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating the overall process of estimating parking lot vacancy status provided in this application embodiment; Figure 2 This is a block diagram of the effective sensing data packet structure provided in the embodiments of this application; Figure 3 This is a flowchart of the effective sensing data packet judgment provided in the embodiments of this application; Figure 4 This is a flowchart of the weighted and weighted fusion calculation of the parking lot vacancy state estimation method provided in the embodiments of this application; Figure 5 This is a structural block diagram of the parking lot vacancy state estimation system provided in the embodiments of this application.

[0038] In the above diagrams: 100, data receiving module; 200, data verification module; 300, weight calculation module; 400, fusion calculation module; 500, effective sensing data packet; 510, local parking space information; 520, data time information. Detailed Implementation

[0039] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application. In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between components; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication between two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0040] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0041] In this application, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0042] Additionally, if the meaning of "and / or" in the text is that it includes three parallel options, taking "A and / or B" as an example, it includes option A, option B, or an option that satisfies both A and B.

[0043] The present application will now be described in detail through exemplary embodiments. However, it should be understood that, without further description, elements, structures, and features in one embodiment may be advantageously incorporated into other embodiments.

[0044] It should be noted that with the continuous growth of car ownership in my country, urban parking resources are becoming increasingly strained, making parking a prominent issue affecting travel experience and traffic efficiency. Drivers often struggle to obtain accurate real-time parking availability information when heading to their destinations, leading to aimless parking searches, exacerbating surrounding traffic congestion, and even forcing them to change their itineraries due to full parking lots, severely reducing travel convenience and time utilization. Therefore, providing drivers with real-time and accurate parking availability information has become an urgent need to be addressed in the fields of intelligent transportation and smart mobility.

[0045] Currently, the main way to obtain parking space status is to deploy sensors such as geomagnetic, ultrasonic, or camera sensors inside the parking lot to monitor the occupancy status of each parking space in real time and publish the number of available spaces through in-park displays or network platforms.

[0046] However, the existing technical solutions mentioned above have limitations. Although they can ensure real-time data, they require high costs for hardware procurement, installation, and subsequent maintenance. As a result, they can only cover a few large parking lots that have completed smart transformation, and cannot be adapted to open-air parking lots, old residential parking lots, and other areas without the conditions for smart transformation, making it difficult to meet the needs of comprehensive parking status monitoring.

[0047] Based on this, this application proposes a parking lot vacancy estimation method, system, and vehicle. By obtaining effective perception data packets from crowdsourced vehicles in the target parking lot, calculating the local vacancy rate and first weight, and then generating the global vacancy rate through weighted fusion, this technical solution achieves the effect of estimating the parking lot vacancy status in real time and accurately without relying on fixed sensors. This solves the problems of high cost, limited coverage, and poor timeliness in obtaining parking lot vacancy information in the prior art.

[0048] In the following, embodiments of this application will be described in detail with reference to the accompanying drawings.

[0049] As attached Figures 1 to 5 As shown in an illustrative embodiment of this application, in a first aspect, this application provides a parking lot vacancy state estimation method, which includes: S1. Obtain the perception data packet from the crowdsourced vehicle located in the target parking lot, and determine whether the perception data packet is a valid perception data packet 500.

[0050] Specifically, the perception data packet is automatically generated by the crowdsourced vehicle through its own configured visual sensors, parking radar and other hardware: when the vehicle enters the target parking lot, after the judgment is triggered by the geofence, Bluetooth beacon and other area markers at the entrance and exit of the parking lot, it will scan the surrounding parking spaces based on the driving path, and simultaneously count the total number of scanned parking spaces and the number of vacant parking spaces. This information, along with the vehicle identifier, data collection timestamp and the vehicle's current area status identifier, is used to mark whether the vehicle is in the target parking lot, and is encapsulated into a perception data packet, which is then uploaded to the cloud system through vehicle network communication technology at a preset period.

[0051] Among them, the perception data packet needs to pass the validity judgment. First, the area validity judgment is based on the vehicle area status identifier in the perception data packet, or combined with the matching result of the vehicle's real-time positioning and the geofence of the target parking lot, to confirm that the vehicle is in the target parking lot; second, the time validity judgment is calculated by calculating the difference between the current time and the data collection timestamp. If the difference is less than the preset validity period, the timeliness is judged to be qualified. In addition, the data must have a certain degree of integrity. Check whether the sensing data packet contains necessary fields such as the total number of scanned parking spaces and the number of available parking spaces. Data packets with missing fields or incorrect formats are directly determined to be invalid.

[0052] In addition, to improve the efficiency and quality of data acquisition, optimization operations will be performed simultaneously to retain only the latest data packet uploaded repeatedly by the same crowdsourced vehicle within a short period of time, thus avoiding data redundancy.

[0053] S2. Based on each valid sensing data packet 500, calculate the local vacant parking space ratio of the valid sensing data packet 500 and its first weight in the weighted fusion calculation.

[0054] Specifically, the local vacant parking space ratio is first calculated for a single valid sensing data packet: the number of vacant parking spaces recorded in the data packet is extracted. Total number of scanned parking spaces Through formula After completing the calculation, the absolute number of parking spaces scanned by vehicles is converted into a standardized proportion index in the range of 0 to 1, eliminating the interference of differences in the scanning scale of different vehicles; then, the first weight is calculated step by step: first, the second weight is calculated based on the total number of scanned parking spaces. The second weight is used to reflect the sample size weight, i.e. , The third weight is calculated based on the sum of the total number of parking spaces scanned by all currently valid sensing data packets. Then, based on the difference between the data acquisition timestamp and the current time, and combined with a preset time decay coefficient λ, a third weight is calculated. This third weight reflects the timeliness weight of the data. Finally, the second weight is multiplied by the third weight to obtain the composite weight, and the composite weight of all valid data is normalized so that the sum of all data weights is 1, thus obtaining the first weight of the data.

[0055] The calculation process requires binding with data identifiers. A unique data identifier is assigned to each valid sensing data packet. The local vacant parking space ratio, second weight, third weight, and first weight are associated with and stored to the corresponding identifiers for easy matching of data and weights during subsequent weighted fusion. Simultaneously, a time decay coefficient... Supports configuration based on parking lot scenarios, such as parking lots in commercial areas during peak hours. Set to 0.5, the off-peak parking lot in the office area will Set to 0.9 to adapt to the timeliness requirements of different scenarios.

[0056] In addition, to improve the rationality and efficiency of the calculation, the rationality of the proportion of vacant parking spaces in a certain area is checked. If the calculation result exceeds the reasonable range of 0 to 1, the data is automatically marked as pending review and will not be included in the weight calculation for the time being.

[0057] S3. Based on the local vacant parking space ratio and its first weight of multiple valid sensing data packets, a global vacant parking space ratio representing the overall vacant status of the target parking lot is generated through weighted fusion calculation.

[0058] Specifically, first extract the percentage of available parking spaces corresponding to all currently valid sensor data packets. With the first weight Each data With the corresponding Perform multiplication, sum all the product results to obtain the base value, and finally multiply the base value by 100% to convert it into a percentage form, i.e., using the formula... The calculation of the global vacant parking space ratio is completed. This process is executed in real time as the valid data changes dynamically. Each time a valid data point is added, the weight of a data point is updated, or an expired data point is removed, the weighted fusion calculation is immediately re-executed to ensure that the output results are synchronized with the current state of the parking lot. For example, if there are currently 3 valid data points: Data 1 ( Data 2 ), Data 3 ( If R = (0.6 × 0.4 + 0.7 × 0.3 + 0.5 × 0.3) × 100% = 60%.

[0059] The output of the global vacant parking space percentage needs to be verified in conjunction with the scan coverage: if the scan coverage is ≥70%, the result will be marked as sufficient data coverage and the result is reliable; if the coverage is between 30% and 70%, the data coverage will be marked as moderate and the result is for reference only; if the coverage is <30%, the current data coverage will be insufficient, and it is recommended to supplement the judgment by referring to the parking lot signs; at the same time, the calculation result needs to be constrained within a range. If the result exceeds the reasonable range of 0% to 100%, the abnormal data investigation process will be automatically triggered, invalid data will be removed, and the calculation will be recalculated.

[0060] In addition, the global vacancy rate will be synchronized in real time to surrounding navigation software and the in-vehicle screens of vehicles in the parking lot via cloud interface, and status labels will be displayed to show whether there are enough vacant spaces or whether the parking spaces are in short supply and about to be full. At the same time, the system will record the percentage change data and generate a trend curve to be pushed synchronously with the current results, which is used to predict changes in parking space availability.

[0061] Through the above scheme, this method can obtain perception data packets from crowdsourced vehicles, avoiding the high cost and insufficient coverage problems of relying on fixed sensors in parking lots. It also overcomes the shortcomings of update lag and poor accuracy caused by manual reporting or historical data prediction, and realizes automated parking space status monitoring without large-scale infrastructure transformation and without relying on manual intervention.

[0062] Furthermore, by using effective sensing data packet verification rules, off-site data, expired data, and abnormal interference data can be eliminated, avoiding the interference of invalid information on the global estimation results. By calculating the local vacant parking space ratio through local parking space information in a single effective sensing data packet, the local scanning results of a single vehicle are transformed into standardized data that can be directly used for fusion, solving the data integration problem caused by inconsistencies in the scanning range and number of different vehicles. Based on crowdsourced local data from multiple vehicles, this solution integrates scattered sensing information through a weighted fusion algorithm, avoiding the problem of limited scanning range of a single vehicle. This allows the global vacant parking space ratio to dynamically approximate the actual vacancy status of the target parking lot, solving the problem in existing technologies where local data cannot reflect the overall situation and the prediction results deviate significantly from the actual state.

[0063] In some embodiments, the valid sensing data packet 500 includes at least local parking space information 510 sensed and counted by the vehicle and corresponding data time information 520; it is determined whether the data packet belongs to the target parking lot based on the local parking space information 510; it is determined whether the sensing data packet is within the preset validity period based on the data time information 520; if both determinations are yes, it is a valid sensing data packet 500, otherwise it is an invalid sensing data packet.

[0064] Specifically, the collection and composition of the perception data packet have a clear execution subject and standardized format: the perception behavior is autonomously completed by crowdsourced vehicles in the target parking lot. The vehicles, using their own surround-view cameras, parking vision sensors, and other hardware, dynamically scan surrounding parking spaces while moving or stationary, automatically counting the total number of scanned parking spaces and the number of vacant parking spaces within the scanning range. Simultaneously, they extract the parking space layout features of the scanned area and encapsulate this information along with the precise timestamp of the data collection time into a perception data packet; wherein, the local parking space information includes the total number of scanned parking spaces. Number of available parking spaces Key information such as parking space layout and local features is presented in the form of timestamps in the format of year, month, day, hour, minute, second, and millisecond.

[0065] The system's regional attribution determination based on local parking space information relies on parking space layout feature matching logic. The system pre-stores a baseline parking space layout feature library for the target parking lot. Upon receiving a perception data packet, it extracts local parking space layout features and performs similarity matching with the baseline feature library. If a match is successful, the vehicle is determined to be within the target parking lot. This design does not rely on positioning signals; even in scenarios where positioning fails, such as underground parking lots, it can still complete regional verification through the scene features of the parking space itself, improving scene adaptability. For timeliness determination based on data time information, the default value for the validity period is 10 seconds. The system calculates the time difference between the current time and the data timestamp. ,like A validity period of 0 seconds is considered valid. In scenarios with high traffic flow, such as business districts and transportation hubs, the validity period can be shortened to 5 seconds to avoid old data affecting the judgment. In scenarios with smooth traffic flow, such as residential areas and office areas, the validity period can be extended to 15 seconds to ensure the sufficiency of valid data.

[0066] Through the above scheme, the local parking space information in this application possesses regional scene attributes, reflecting whether the physical location of the vehicle is within the target parking lot. It eliminates invalid data scanned outside the parking lot, avoiding global proportion estimation errors caused by cross-regional data mixing. Parking lot status exhibits high-frequency changes, with frequent vehicle entry and exit. Expired data can no longer reflect the current true status; timeliness verification allows for the direct discarding of such data, avoiding misleading results caused by estimating the current status using past data. Data time information is correlated with subsequent timeliness weight calculations, such as through a time decay coefficient λ. Data retained after timeliness verification can obtain weight in weighted fusion, making the global estimation result more closely match the real-time status of the parking lot.

[0067] In some embodiments, the second weight is determined based on local parking space information. The second weight of the m-th data The second weight of the m-th data point is calculated as the ratio of the total number of scanned parking spaces corresponding to this valid sensing data point to the sum of the total number of scanned parking spaces corresponding to all valid sensing data points.

[0068] in, The second weight for the m-th data point. This represents the total number of parking spaces scanned corresponding to this valid sensing data packet. This is the sum of the total number of scanning parking spaces corresponding to all valid sensing data packets.

[0069] Specifically, the calculation of the second weight is a step performed after the effective sensing data packets are filtered: the system first extracts all effective sensing data packets that have passed the area attribution and timeliness verification, and then reads the total number of scanned parking spaces corresponding to the local parking space information in each data packet. Then, calculate the sum of the total number of scanned parking spaces based on all currently valid data. Then, the values ​​are substituted into the formula to calculate the second weight for each data point. For example, if there are currently 3 valid data points: Data 1: Total number of parking spaces scanned. Data 2 scans the total number of parking spaces. Data 3 scans the total number of parking spaces. ,but =15+20+5=40, corresponding to the second weight of data 1. The second weight of data 2 The second weight of data 3 This process directly converts the spatial coverage of a single data point into a weight ratio, ensuring that data with a wider scanning area has a more reasonable weight in subsequent fusion.

[0070] Furthermore, to improve the adaptability and rationality of the second weight, overlapping scan areas are handled with compatibility. If the scan areas of multiple vehicles overlap, the total number of scanned parking spaces is still counted independently as a single data point without deduplication, thus avoiding underestimation of coverage due to overlapping areas. The information value of overlapping areas will be reflected through the complementarity of multiple data in subsequent weighted fusion, preventing over-allocation of weights. When new valid data is added, or old data is removed due to timeliness or regional verification failure, the system will automatically recalculate. The second weight of all currently valid data is updated synchronously to ensure that the weight allocation always matches the coverage of the latest valid data.

[0071] Through the above scheme, the second weight in this application is used to determine the sample size reflected by the local parking space information. By using the proportion of the total number of scanned parking spaces as the basis for weight allocation, the effective perception data packets with a wider scanning range are given higher weights, thereby improving the accuracy and reliability of the parking lot vacancy state estimation results. The contribution of different data is reasonably balanced to avoid excessive influence of local small-scale scanning data on the overall estimation results, thereby enhancing the robustness of the vacancy state estimation model and better adapting to actual application scenarios where the scanning range of crowdsourced vehicles varies greatly.

[0072] Furthermore, the second weight is based on the total number of parking spaces scanned for a single valid data point. It converts the data's information coverage into a weight ratio by comparing its own total scan count with the sum of all valid data scan counts. A single valid data point with a larger total number of scanned parking spaces indicates a wider local perception coverage for the vehicle, and the corresponding local parking space information is more representative of the overall parking lot status, thus receiving a higher second weight. This avoids the unreasonable situation where narrow, partial data and comprehensive data with wide scan counts are weighted equally, ensuring that the data contribution matches its actual information value, and guaranteeing the rationality of data utilization through weight allocation.

[0073] It should be noted that the local scanning of crowdsourced vehicles has the limitation of limited coverage. A single vehicle cannot scan all parking spaces in a parking lot at once. However, the calculation logic of the second weight can make up for this limitation by weighted integration of data from multiple vehicles: effective data with a wide scanning range has a high second weight and accounts for a higher proportion in subsequent fusion calculations. The parking space information it covers can more fully reflect the local state of the parking lot. After the scanning data of multiple vehicles are weighted according to the coverage range, the scattered local perception information can be gradually pieced together into a more complete global parking space status, which can offset the problem of narrow scanning range and one-sided information of a single vehicle, and make the estimated result of the global vacant parking space ratio closer to the actual situation.

[0074] Furthermore, the calculation of this second weight relies solely on two values: the total number of parking spaces scanned in a single data entry and the sum of the total number of scans of all valid data entries, offering the advantage of low computational power consumption. It eliminates the need for complex feature extraction, model training, or scene recognition algorithms, enabling rapid weight calculation for a single data entry. Adaptable to scenarios with massive amounts of crowdsourced vehicles uploading data in real time, it can synchronously iterate weights as data is updated, avoiding delays caused by complex calculations and ensuring the real-time nature of global idle state estimation.

[0075] In some embodiments, a third weight is determined based on data time information, and the third weight is calculated as follows:

[0076] in, As the third weight, The time decay coefficient is a preset value and 0 < λ < 1. For the current time, The data collection time is the m-th data point indicated by the data time information.

[0077] Specifically, the timeliness quantification step performed for each valid sensing data packet is as follows: the system first reads the acquisition time of the data. Synchronously obtain the current system time Calculate the time difference between the two. Then, the preset time decay coefficient is substituted. Perform the exponentiation operation to obtain the third weight of the data. For example, if If we set the weight to 0.8, and the time difference between the time a certain data point was collected and the current time is 2 seconds, then its third weight... If the time difference increases to 5 seconds, the third weight becomes... This intuitively demonstrates that the older the data, the lower its weight.

[0078] In addition, if the time difference If the parking lot's preset validity period has expired, the system will directly set the third weight of the data to 0, and it will no longer participate in subsequent weighted fusion calculations to avoid expired data interfering with the global results. At the same time, if the third weight result is lower than 0.1, the system will mark the data as old in the data label, and the weight can be reduced in subsequent calculations based on this label.

[0079] Through the above scheme, the third weight in this application reflects the timeliness determined by the data's time information. Parking lot space status has a dynamic change characteristic, and older data cannot reflect the current reality. Time difference. The larger the value, the older the data. Combined with the attenuation coefficient (0 < λ < 1), The smaller the result, the higher the weight; conversely, the closer the data collection time is to the current time, the smaller the time difference, and the higher the weight.

[0080] In summary, the third weight complements the second weight based on the total number of scanned parking spaces. When combined, effective data with a wide scanning range and newer data will receive a higher composite weight. This avoids the excessive proportion of perceived information with a large coverage but outdated data, and also avoids the limited impact of information with new data but a narrow coverage. By selecting high-value data based on the spatial coverage and timeliness of the data, the weighted fusion result is more in line with the real global state of the parking lot.

[0081] In some embodiments, a composite weight is determined for the valid sense data packet in the weighted fusion calculation, and the composite weight is the second weight. With the third weight The product of these factors and the composite weights satisfy the following formula:

[0082] in, Let be the composite weight of the m-th valid sensing data packet; normalize the composite weights of all valid sensing data packets so that... The first weight of the m-th valid sensing data packet is calculated. .in, , It is the sum of the composite weights of all valid sense data packets.

[0083] Specifically, the first step is to extract the calculated second and third weights of a single valid sensing data packet, and multiply them to obtain the initial composite weight of that data. The second step is to sum the initial composite weights of all current valid data, and then divide the initial composite weight of each data packet by this sum to complete the normalization process, ultimately obtaining the first weight. For example, if there are currently two valid data packets: data 1... , The initial composite weight is 0.5 × 0.8 = 0.4; Data 2 , The initial composite weight is 0.3 × 0.6 = 0.18; the total composite weight is 0.4 + 0.18 = 0.58, so the first weight of data 1 is... Data 2 Ensure that the sum of the first weights of all valid data is 1.

[0084] Furthermore, if the initial composite weight of a data point is lower than a preset threshold, the system will mark it as low-value data. During subsequent fusion calculations, its priority can be selectively reduced or it can be temporarily excluded from the calculation to avoid low-value data interfering with the overall results. When new valid data is added or old data is removed due to timeliness / regional verification failure, the system will automatically recalculate the initial composite weight and normalized results of all currently valid data to ensure that the first weight always matches the value of the latest valid data. This achieves dynamic data updates and real-time weight adjustments, ensuring the real-time nature of the overall estimation results.

[0085] Through the above scheme, the composite weighting of this application is approved. The sample size and timeliness are correlated, and the first weight is obtained by normalization. Only valid data that simultaneously meets the criteria of a large total number of scanned parking spaces and is new at the time of data collection can obtain a higher first weight; this avoids the limitations of single-dimensional weighting and ensures that the contribution of each data point matches its actual information value.

[0086] In addition, after normalization, the sum of the weights of all valid data is 1, which makes the subsequent global vacancy rate of parking spaces... The calculation results always remain within the range of 0% to 100%, avoiding abnormal results caused by excessive weight accumulation. The normalized data weights are all under a uniform quantization scale, ensuring the comparability and consistency of the weighted fusion results for different amounts of data, thus improving the algorithm's stability.

[0087] In some embodiments, the weighted fusion calculation is implemented using the following formula:

[0088] in, This represents the percentage of available parking spaces globally. The first weight is the weight corresponding to the m-th valid sensing data packet. The percentage of available parking spaces corresponds to the m-th valid sensing data packet.

[0089] Using the above scheme, the local vacancy rate of each valid data packet in this application Represents the idle state of the vehicle scanning area, combined with the first weight. The differentiated allocation can stitch together scattered local vacancy information to cover the entire parking lot; it avoids the problem that the scanning range of a single vehicle is limited and local data cannot reflect the whole, so that the global vacancy rate R can approximate the actual vacancy status of the target parking lot.

[0090] Specifically, the weighted fusion calculation is a global state integration step performed after the first weight of all valid sensing data and the proportion of local vacant parking spaces have been calculated. It first extracts all currently valid data from the system database. and Ensure that every piece of data and One-to-one correspondence, with no mismatches. Then, for each data point, the product of the first weight and the local vacant parking space percentage is performed. Finally, all product results are summed to obtain the base value of R, which is then converted to a percentage. For example, there are currently 3 valid data points: Data 1 ( Data 2 ), Data 3 ( If R = 0.4 × 0.6 + 0.3 × 0.7 + 0.3 × 0.5 = 0.6, then the global vacancy rate is 60%.

[0091] Furthermore, if the calculated R value exceeds the 0-1 range, such as due to data errors, the system will automatically trigger a process of re-filtering valid data and recalculating weight ratios. After excluding abnormal data, the fusion calculation will be performed again to ensure the reasonableness of the output results. The system will output R in sync with the current parking lot's scan coverage. If the scan coverage is ≥70%, the labeling results are based on high-coverage data and have high reliability. If the scan coverage is <30%, the current data coverage is insufficient, and the results are for reference only, helping users make more rational decisions based on R.

[0092] In some embodiments, the percentage of partially vacant parking spaces The ratio of the number of available parking spaces in the m-th sensing data packet to the total number of scanned parking spaces represented by the m-th local parking space information is calculated using the following formula:

[0093] in, Let m be the number of available parking spaces in the sensing data packet mentioned in the m-th packet. The total number of scanned parking spaces represented by the local parking space information described in the m-th entry.

[0094] Specifically, the calculation of the percentage of vacant parking spaces in a given area is a standardized procedure automatically executed by the crowdsourced vehicle after it completes a scan of a local area: as the vehicle scans the surrounding parking spaces using its visual sensors, it simultaneously counts the total number of clearly identified parking spaces. The number of parking spaces that are currently vacant. The system encapsulates these two values ​​into a sensing data packet. After the data packet passes validity verification, the system directly extracts these two fields and substitutes them into the formula for calculation. For example, if a vehicle scans 15 identifiable parking spaces, and 6 of those spaces are empty, then the percentage of available parking spaces for that data entry is calculated. This represents 40%, which visually quantifies the vacancy rate of the vehicle's scanned area.

[0095] in, The range of values ​​is The number of available parking spaces is a subset of the total number of parking spaces scanned. If the sensing data packet contains... If the sensor fails to recognize the data, it will be considered abnormal and the data will be marked as invalid. Must be a positive integer and not less than 1: Only if the vehicle scans at least one identifiable parking space, Only then does it have practical significance, if (If the vehicle is in a passageway area without parking spaces), then this data packet will not participate in the calculation of the local vacant parking space ratio, thus avoiding meaningless calculations in the formula.

[0096] Furthermore, if the same vehicle scans the same area multiple times within a short period of time, the system will calculate the results of the multiple scans. The average value, as the final local vacancy rate of the area, offsets occasional recognition errors of the sensor in a single scan.

[0097] Through the above scheme, this application transforms the local perception data of a single crowdsourced vehicle into a standardized idle status indicator by using the ratio of the number of vacant parking spaces to the total number of scanned parking spaces. By uniformly quantifying the idle level of local areas using a scale of 0 to 1, the impact of differences in scan scale on data comparability is eliminated; this places local perception data from different vehicles under the same measurement dimension, laying a standardized data foundation for weighted fusion calculations of multiple vehicles. It directly reflects the actual idle rate within the scanning area of ​​a single crowdsourced vehicle.

[0098] In some embodiments, the parking lot vacancy estimation method further includes scanning coverage calculation:

[0099] in, Where N represents the scan coverage, and N is the total number of parking spaces in the target parking lot. This is the sum of the total number of scanning parking spaces corresponding to all the aforementioned valid sensing data packets.

[0100] Specifically, the calculation of scanning coverage is performed simultaneously with the global percentage of vacant parking spaces. This is a quantification of the coverage of the current effective data. The total number of parking spaces N in the target parking lot needs to be entered into the system in advance. The number of parking spaces N can be obtained from the information provided by the parking lot management and the structured data from the map service provider to ensure consistency with the actual number of parking spaces. This involves real-time statistics of the total number of scanned parking spaces corresponding to all valid sensor data packets that have passed regional and time-sensitive verification. Each valid data packet can be used to calculate the total number of valid parking spaces. The summation is used to obtain the total number of parking spaces. For example, if the actual total number of parking spaces in the target parking lot is N=80, the sum of the total number of parking spaces scanned by the currently valid sensing data packets is used to obtain the total number of parking spaces. Then the scan coverage If the scanned areas of multiple vehicles overlap, It may reach 90, at this point This indicates that the current valid data covers the entire parking lot area, and some areas may have been scanned repeatedly. To avoid the impact of repeated scanning on the accuracy of the results, the system records the unique identifier of each parking space scanned by each vehicle, such as the parking space number and relative coordinates. The system also counts the number of distinct parking spaces that have been scanned at least once. , combined Calculate the actual physical coverage, actual coverage = For example: The target parking lot actually has a total of 80 parking spaces, N=80. ,but The actual physical coverage is 87.5%, clearly indicating incomplete coverage; if If the actual physical coverage is 100%, it means that full coverage has been achieved. When the same parking space is scanned by multiple vehicles, its vacancy status will be verified multiple times, which can offset the recognition error of a single vehicle. The more samples in a local area, the more accurate the proportion of vacant parking spaces in that area will be, and the more reliable the final fusion calculation result will be.

[0101] In the above scheme, this application quantifies the coverage of the target parking lot by scanning the ratio of the total number of parking spaces to the total number of parking spaces in the parking lot using all valid data. A higher Coverage indicates that the effective data has covered most of the parking spaces in the parking lot, and the global vacant parking space ratio obtained by weighted fusion has higher data support credibility; if the Coverage is low, it indicates that the effective data only covers a small part of the area, and the estimated result of the global ratio may be biased and cannot reflect the status of the uncovered area. When the coverage is low, the system can specifically incentivize crowdsourced vehicles in the target parking lot to upload data, or prioritize the use of valid data with a large number of scanned parking spaces, gradually expanding the data coverage. As the coverage increases, the area coverage of valid data becomes more complete, and the estimation error of the global vacant parking space ratio will decrease accordingly, forming a virtuous cycle of improved coverage and more accurate results, thus achieving dynamic self-optimization of the technical solution.

[0102] A second aspect of this application provides a parking lot vacancy estimation system, including a data receiving module 100, a data verification module 200, a weight calculation module 300, and a fusion calculation module 400. The data receiving module 100 is used to acquire sensing data packets located within a target parking lot; the data verification module 200 is used to determine whether the sensing data packets are valid sensing data packets 500; the weight calculation module 300 is used to calculate, based on each valid sensing data packet 500, the local vacant parking space ratio of that valid sensing data packet 500 and its first weight in the weighted fusion calculation; the fusion calculation module 400 is used to generate a global vacant parking space ratio representing the overall vacancy status of the target parking lot through weighted fusion calculation based on the local vacant parking space ratios and their first weights of multiple valid sensing data packets 500.

[0103] Specifically, the data receiving module 100 obtains perception data packets from the crowdsourced vehicles in the target parking lot. The trigger condition is that the vehicle enters the parking lot. After activation, the vehicle automatically collects surrounding parking space information through visual sensors and encapsulates it into a standardized data packet containing the total number of scanned parking spaces, the number of vacant parking spaces, local features of the parking space layout, and the collection timestamp. The data packet is then uploaded using vehicle-to-everything (V2X) communication or a mobile network.

[0104] The data verification module 200 performs filtering of valid data, matching the parking space layout features in the data packet with the parking lot baseline feature library, or confirming whether it is located in the target parking lot through the satellite positioning system, excluding off-site data; calculates the difference between the current time and the collection timestamp, and if it is less than the preset validity period, the validity period is qualified; checks the required fields such as the total number of scanned parking spaces and the number of vacant spaces, and directly marks the data packet that is missing or has abnormal values ​​as invalid.

[0105] The weight calculation module 300 first uses the formula Calculate the percentage of vacant parking spaces in a given area; then calculate the second weight in sequence. Third weight The two are multiplied and then normalized to obtain the first weight; all results are bound to the data packet identifier and stored to ensure accurate subsequent matching.

[0106] The fusion computing module 400 extracts the percentage of available parking spaces from valid data packets. With the first weight ,pass The weighted summation is completed and converted into a percentage of the global vacant parking space ratio. This calculation iterates with the addition / expiration of data, while verifying whether the result is in the 0%-100% range. If an anomaly occurs, data re-screening and secondary calculation are triggered.

[0107] The output of the data receiving module is the input of the data verification module. The valid data after verification is pushed to the weight calculation module, and the weight result is then transferred to the fusion calculation module. The functions of the modules are independent of each other and can be upgraded independently without affecting the operation of other modules, thus improving the maintainability of the system.

[0108] Through the above scheme, the data receiving module of this application is responsible for data input, the data verification module is responsible for data filtering, the weight calculation module is responsible for data value standardization, and the fusion calculation module is responsible for global result generation. From the upload of the perception data packet to the output of the global idle ratio, no manual intervention is required, which improves the efficiency of data processing, avoids errors caused by manual operation, and ensures the stability and consistency of system operation.

[0109] By leveraging vehicle visual perception capabilities, there is no need to deploy additional dedicated hardware such as fixed cameras in parking lots, saving the procurement, installation, and maintenance costs of traditional solutions and lowering the barrier to system implementation. It is not limited by whether the parking lot has completed intelligent transformation. Even in old parking lots or open-air parking lots, as long as crowdsourced vehicles enter the site and upload data, it solves the limitations of traditional fixed sensor solutions.

[0110] A third aspect of this application also provides a vehicle including the parking lot vacancy estimation system described in the second aspect.

[0111] In the above embodiments, after the vehicle of this application integrates the system, it can directly reuse its own widely used intelligent sensing hardware vision camera to autonomously complete the collection operations of scanning local parking space information, counting the number of vacant spaces, and generating sensing data packets, without the need for additional dedicated hardware or manual data submission by users; it avoids the hardware costs and manual operation thresholds of crowdsourced data collection, allowing each intelligent vehicle to become a crowdsourced node for parking lot status monitoring, thus expanding the base of data contribution participants.

[0112] In addition, once a vehicle is equipped with this system, it not only becomes a data provider but also directly obtains the global vacancy rate output by the system. Combined with its own location information, the vehicle can autonomously plan a better parking route, avoiding blindly circling around to find a parking space and shortening parking time. If the global vacancy rate is extremely low, the system can alert the vehicle in advance that the parking lot is about to be full, helping users to adjust their destination in time and reduce unnecessary travel.

[0113] In the overall data flow, this technical solution achieves real-time acquisition and sharing of parking space information through a closed-loop process of vehicle-side data collection, cloud-based aggregation and processing, and finally data sharing services. Vehicles entering the target parking lot use their own visual sensors to scan and identify surrounding parking spaces, completing the total number of scanned parking spaces and determining the occupancy / fullness status on the vehicle side. The vehicle then uploads information including parking space statistics and collection timestamps to the cloud's central processor. The cloud receives parking space data uploaded by multiple vehicles entering the parking lot, and uses algorithms to aggregate, integrate, and analyze the multi-source data, outputting the real-time parking space occupancy / vacancy status of the parking lot. The cloud shares the processed parking space vacancy data to public network platforms, such as navigation and map applications, for reference by other users visiting the parking lot.

[0114] Furthermore, this technical solution addresses the technical challenges of traditional parking lot monitoring solutions, achieving low cost, wide coverage, and high timeliness. Adaptable to diverse scenarios, it eliminates the need for additional dedicated hardware such as geomagnetic sensors and cameras in parking lots, directly reusing the vehicle's existing visual sensors. This avoids hardware procurement, installation, and maintenance costs, making it suitable for open-air parking lots, older residential parking lots, and other areas lacking intelligent upgrades, filling coverage gaps in traditional solutions. Ensuring real-time performance, it matches dynamic changes in parking spaces. Real-time scanning and processing of data from the vehicle upon entry, along with cloud-based aggregation and analysis of multi-source data, ensures data updates are synchronized with parking space status, resolving the issue of poor timeliness in reporting parking space statistics by management. Optimizing travel efficiency and alleviating traffic congestion, users can obtain real-time parking space availability information in advance, avoiding blindly searching for spaces or changing their itinerary due to full capacity. This improves personal travel convenience and time utilization, reduces regional traffic congestion caused by circling around for parking spaces, and contributes to improving overall traffic efficiency.

[0115] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for estimating the vacancy status of a parking lot, characterized in that, It includes: Acquire perception data packets from crowdsourced vehicles located in the target parking lot, and determine whether the perception data packets are valid perception data packets; Based on each of the aforementioned valid sensing data packets, the local vacant parking space ratio of that valid sensing data packet and its first weight in the weighted fusion calculation are calculated. Based on the local vacant parking space percentage and its first weight of multiple valid sensing data packets, a global vacant parking space percentage representing the overall vacancy status of the target parking lot is generated through the weighted fusion calculation.

2. The parking lot vacancy estimation method according to claim 1, characterized in that, The effective perception data packet includes at least the local parking space information perceived and statistically analyzed by the vehicle, as well as the corresponding data time information; Based on the local parking space information, determine whether the vehicle belongs to the target parking lot; based on the data time information, determine whether the sensing data packet is within the preset validity period; If both conditions are met, then it is considered a valid sensing data packet.

3. The parking lot vacancy estimation method according to claim 2, characterized in that, The second weight is determined based on the local parking space information. The second weight of the m-th data It is the ratio of the total number of scanning parking spaces corresponding to the valid sensing data packet to the sum of the total number of scanning parking spaces corresponding to all valid sensing data packets.

4. The parking lot vacancy estimation method according to claim 3, characterized in that, The third weight is determined based on the data time information, and the calculation method of the third weight is as follows: in, As the third weight, The time decay coefficient is a preset value and 0 < λ < 1. For the current time, The data collection time is the m-th data point indicated by the data time information.

5. The parking lot vacancy estimation method according to claim 4, characterized in that, The composite weight of the m-th valid perceived data packet in the weighted fusion calculation is determined, and the composite weight is the second weight. With the third weight The product of the values; and the first weight is obtained by normalizing the composite weights of all valid sensed data. .

6. The parking lot vacancy estimation method according to claim 5, characterized in that, The weighted fusion calculation is achieved through the following formula: in, The percentage of available parking spaces globally. The first weight corresponds to the m-th valid perceived data. The percentage of available parking spaces corresponds to the m-th valid sensing data packet.

7. The estimation method according to claim 6, characterized in that, The percentage of vacant parking spaces in the local area It is the ratio of the number of available parking spaces in the sensing data packet described in the m-th entry to the total number of scanned parking spaces represented by the local parking space information described in the m-th entry.

8. The parking lot vacancy estimation method according to claim 1, characterized in that, It also includes the calculation of scanning coverage, which is the ratio of the sum of the total number of scanned parking spaces corresponding to all the valid sensing data packets to the total number of parking spaces in the target parking lot.

9. A parking lot vacancy status estimation system, characterized in that, include: The data receiving module is used to acquire sensing data packets located in the target parking lot; The data verification module is used to determine whether the sensing data packet is a valid sensing data packet; The weight calculation module is used to calculate the local vacant parking space ratio of each valid sensing data packet and its first weight in the weighted fusion calculation based on each valid sensing data packet. The fusion computing module is used to generate a global vacant parking space percentage that represents the overall vacant status of the target parking lot through the weighted fusion computing based on the local vacant parking space percentage and its first weight of multiple valid sensing data packets.

10. A vehicle, characterized in that, Includes the parking lot vacancy estimation system as described in claim 9.

Images