Parking lot parking space information monitoring method and system based on ultrasonic detection

Abstract

The invention discloses a parking lot parking space information monitoring method and system based on ultrasonic detection, and relates to the technical field of information monitoring. According to the method, an independent file is established for each ultrasonic sensor and initial calibration is carried out, so that an accurate judgment threshold value is obtained; by collecting multi-modal data such as distance, echo intensity and signal-to-noise ratio and combining a stability index obtained by variance analysis, the parking space state is accurately judged by using a fusion judgment rule; furthermore, the pre-judgment of the state that the parking space is about to be idle is realized through a driving-away prediction mechanism combining behavior trajectory tracking, payment signals and sensor mode recognition; based on the aggregated parking space state data, an optimal parking path is provided for a user through a multi-target weighted path planning algorithm; according to the invention, the problems of weak anti-interference capability and single function in the prior art are effectively overcome, and the accuracy, prediction capability and data utilization rate of parking space state identification are significantly improved, so that the operation efficiency and user experience of the parking lot are comprehensively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of information monitoring, and in particular relates to a parking lot space information monitoring method and system based on ultrasonic detection. BACKGROUND

[0002] With the acceleration of urbanization, the number of motor vehicles has increased dramatically, and "parking difficulty" has become a common problem for city managers and citizens. The existing parking lot space monitoring technology mainly has the following several kinds: ultrasonic detection technology, which is the most popular technology at present; it installs ultrasonic sensors above each parking space, calculates the distance from the object below by emitting and receiving ultrasonic waves, and thus judges whether the parking space is occupied. However, the traditional scheme is susceptible to interference, for example: pedestrians walking by, shopping carts staying, vehicles passing by the adjacent lane, etc., which may be misjudged as having a car, resulting in false positives. In addition, the sensor may drift over a long period of use, affecting the detection accuracy; geomagnetic detection technology, which detects the disturbance of the vehicle to the earth's magnetic field to determine the parking space status, has the disadvantages of being susceptible to interference from surrounding large metal objects or underground pipelines, and requiring a high parking position for the vehicle; video recognition technology, which collects images through a camera and uses computer vision algorithms to identify the parking space status, has a high cost, is greatly affected by light, weather, and obstructions, and involves privacy issues.

[0003] In summary, the existing technology, especially the most economical ultrasonic detection technology, generally has the problems of weak anti-interference ability, single function (only providing a binary "car / no car" state), and inability to predict and deeply utilize the parking space status. SUMMARY

[0004] (I) Technical problems solved In view of the problems in the related art, the present application provides a parking lot space information monitoring method and system based on ultrasonic detection to overcome the above technical problems existing in the prior art.

[0005] (II) Technical solutions To solve the above technical problems, the present application is realized by the following technical solutions: S1, an independent configuration file is established for each parking space ultrasonic sensor to obtain a set of parking space sensor files; in the empty state of the parking lot, an initial judgment threshold is obtained through initial threshold calibration; and the initial judgment threshold is stored in the set of parking space sensor files to obtain a first set of parking space sensor files; S2, distance data, echo signal strength, and signal-to-noise ratio (SNR) data are collected by the ultrasonic sensor to obtain multi-modal data; and stability index data is obtained by processing the multi-modal data through variance calculation; S3. Based on the first parking space sensor file set, multimodal data, and stability index data, determine whether there is a car in each parking space by fusing judgment rules to obtain the initial parking space status data; S4. Update the initial parking space status data by tracking behavior trajectory and predicting departure to obtain the updated parking space status data; S5. The updated parking space status data is aggregated and displayed, and the optimal route is planned to obtain the optimal route; when the vehicle parks in the parking space according to the optimal route, the updated parking space status data is updated again. This invention lays the foundation for accurate detection by establishing an independent profile and performing initial calibration for each ultrasonic sensor; it obtains stability indicators by collecting multimodal data such as distance, echo intensity, and signal-to-noise ratio, combined with variance analysis, effectively filtering out instantaneous interference; it accurately determines the parking space status based on fusion judgment rules; further, it predicts the imminent vacancy of parking spaces through behavior trajectory tracking and departure prediction mechanisms; and finally, it uses a multi-objective weighted path planning algorithm to intelligently recommend the optimal parking space for car owners. This method greatly improves the accuracy and anti-interference capability of parking space status identification, while realizing a leap from passive monitoring to active prediction and guidance, significantly improving parking lot operation efficiency and user parking experience.

[0006] Preferably, step S1 includes the following steps: S11. Install ultrasonic sensor nodes directly above each parking space in the parking lot to obtain a set of ultrasonic sensor nodes for each parking space. S12. On the management platform of the central server, upload or draw the map of the parking lot to obtain the initial digital twin map; on the initial digital twin map, according to the actual installation location of each sensor node in the ultrasonic sensor node set of the parking space, record each sensor node and assign a unique identifier to obtain the final digital twin map. Based on the final digital twin map, a stable communication connection is established between each sensor node in the parking space ultrasonic sensor node set and the regional gateway. In the central server, an independent configuration file is created for each sensor node to obtain the parking space sensor file set. S13. When the parking lot is vacant, the initial reference distance is obtained by collecting and calculating the reference data; the initial judgment threshold is calculated based on the initial reference distance and the height compensation value; the initial judgment threshold is stored in the parking space sensor file set to obtain the first parking space sensor file set. This invention constructs a complete sensor network by installing ultrasonic sensor nodes directly above each parking space and assigning a unique identifier to each node in a digital twin map to establish a precise physical information mapping. By collecting reference distances in an vacant state and combining them with vehicle height compensation values, an adaptive initial judgment threshold is calculated for each sensor. This initialization process lays a solid foundation for achieving high-precision parking space status recognition and also realizes the system's accurate digital modeling of the parking lot environment and individualized adaptive calibration of the sensors.

[0007] Preferably, step S13 includes the following steps: S131. After ensuring that there are no vehicles, people or other obstacles in all target parking spaces, issue an initial calibration command for the entire site through the management platform. After receiving the initial calibration command for the entire site, the server broadcasts the calibration command to each sensor node in the ultrasonic sensor node set of the parking space through the gateway; after receiving the command, each node collects a set of distance readings to obtain the distance reading set. S132. Remove gross errors from the readings of each sensor node in the distance reading set and calculate the arithmetic mean to obtain an initial reference distance set; the initial reference distance set contains the distance from each sensor to the ground; S133. Based on the common vehicle types in the parking lot and the initial reference distance set, calculate the initial judgment threshold for each node in the ultrasonic sensor node set of the parking space using the initial judgment threshold formula to obtain the initial judgment threshold set. S134. Write the initial judgment threshold set and the initial reference distance set into the configuration file of each node in the parking space ultrasonic sensor node set to obtain the first parking space sensor file set. This invention automatically triggers a full-site calibration command when the parking lot is vacant, controlling each sensor node to collect raw distance data. After removing gross errors and calculating the mean, an accurate initial reference distance set is obtained. Then, combined with a preset vehicle height compensation value, a personalized initial judgment threshold is calculated for each node, ultimately updating the sensor profile. This automatic calibration process effectively eliminates the impact of individual sensor differences and installation errors on detection accuracy, while achieving standardization and unification of the system's measurement benchmark, laying a solid foundation for subsequent stable and reliable parking space status recognition.

[0008] Preferably, step S2 includes the following steps: S21. Set the acquisition period; based on the acquisition period, collect distance data, echo signal intensity, and signal-to-noise ratio (SNR) data through the ultrasonic sensor to obtain multimodal data; S22. Set the number of continuous acquisition cycles; calculate the variance of the multimodal data under the number of continuous acquisition cycles to obtain the stability index data; This invention constructs multimodal data by periodically collecting distance, echo intensity, and signal-to-noise ratio, and obtains stability indices by performing variance analysis on continuous periodic data. While effectively capturing the differences in surface features and motion states of objects, it achieves intelligent filtering of instantaneous interference, providing a key basis for accurately distinguishing between stationary vehicles and dynamic interference such as pedestrians.

[0009] Preferably, step S3 includes the following steps: S31. Set stability threshold and signal strength threshold; based on stability threshold and signal strength threshold, set fusion judgment rules; S32. Use a filtering algorithm to preprocess the multimodal data to obtain the processed multimodal data; S33. Based on the first parking space sensor file set, the processed multimodal data, and the stability index data, determine whether there is a car in each parking space by fusing judgment rules, obtain the judgment result, update the status of each parking space in the parking lot according to the judgment result, and obtain the initial parking space status data. This invention constructs a fusion judgment rule by setting stability and signal strength thresholds, and uses a filtering algorithm to preprocess sensor data. It integrates archive data, multimodal data and stability indicators, and accurately determines the parking space status according to the rule sequence. This significantly improves the accuracy of vehicle recognition, effectively filters out interference from pedestrians and small objects, and realizes real-time and reliable updates of parking space status.

[0010] Preferably, step S31 includes the following steps: S311. Set a vehicle presence determination rule; the vehicle presence determination rule is that the distance data is less than the initial determination threshold, the stability data is greater than the stability threshold, and the echo signal strength is greater than the signal strength threshold. S312, Set interference / invalidation rules; the dynamic interference / invalidation rules and the static interference / invalidation rules; the dynamic interference / invalidation rules are when the distance data is less than the initial judgment threshold and the stability data is less than the stability threshold; the static interference / invalidation rules are when the distance data is less than the initial judgment threshold, the stability data is less than the stability threshold, and the echo signal strength is less than the signal strength threshold; S313. Set a vacancy rule; the vacancy rule is that the distance data is greater than the initial judgment threshold. S314. Set the order of rule execution; the order of rule execution is as follows: empty space rule is judged first, followed by dynamic interference / invalid rule, then static interference / invalid rule, and finally the vehicle detection rule is triggered. This invention constructs a multi-condition, hierarchical fusion judgment logic by setting a complete rule sequence that includes vehicle presence determination, dynamic / static interference elimination, and empty space confirmation, and stipulating that the empty space rule takes priority, the interference rule takes second place, and the vehicle presence determination takes last. This significantly improves the accuracy of vehicle recognition, effectively filters instantaneous interference such as pedestrians and small objects, and achieves a leap from simple distance judgment to comprehensive recognition based on object motion characteristics and physical features, ensuring the high reliability of parking space status judgment results.

[0011] Preferably, step S4 includes the following steps: S41. Extract parking space data from the initial parking space status data that shows parking spaces changing from empty to occupied, to obtain changed parking space data; based on the final digital twin map, find sensor nodes that are spatially adjacent to the parking spaces in the changed parking space data and that may be passed by vehicles, to obtain the initial set of adjacent sensor nodes. Query the status change records of parking spaces in the initial adjacent sensor node set within the most recent time window to obtain node change data; Set the straight-ahead parking rules and the turning parking rules; query the node change data to see if there is data that matches the straight-ahead parking rules and the turning parking rules, and obtain the trajectory tracking data; S42. Departure prediction data is obtained by using payment departure prediction, sensor pattern recognition, and geomagnetic disturbance. S43. Update the initial parking space status data based on trajectory tracking data and departure prediction data to obtain updated parking space status data; This invention analyzes parking space status change data and uses digital twin maps to track the trajectories of adjacent nodes to verify the rationality of vehicle parking behavior. Simultaneously, it integrates payment system signals, sensor flashing phenomena, and geomagnetic disturbance data to predict vehicle departure intentions from multiple dimensions. This effectively distinguishes between genuine parking behavior and sensor false alarms, significantly improving the reliability of status judgments and achieving a strategic shift from passive detection to proactive prediction. This lays a key technological foundation for providing information on "soon-to-be-vacant" parking spaces and optimizing parking resource allocation.

[0012] Preferably, step S42 includes the following steps: S421. Establish a secure data connection with the parking lot's payment system. When a payment signal is received, collect vehicle information and determine that the parking space will soon be vacant. S422. Set the flickering phenomenon rule and the geomagnetic disturbance rule; the flickering phenomenon rule is that for an occupied parking space, the sensor data changes from a car present to a car absent and then back to a car present within 2 seconds; The geomagnetic disturbance rule is that a rapid and continuous disturbance occurs to the geomagnetic field of an occupied parking space; Continuously monitor the sensor data of all parking spaces with vehicles in the status to obtain vehicle data; when the vehicle data shows a phenomenon that matches the flashing phenomenon rule and the geomagnetic disturbance rule, collect vehicle information and determine that the parking space will soon be vacant; This invention obtains the intention to leave through a linked payment system and monitors the "flickering" phenomenon of sensors and geomagnetic disturbance patterns to capture characteristic signals of vehicles before starting from multiple dimensions. While upgrading the traditional parking space status detection to the prediction of leaving behavior, it also realizes the accurate prediction of "soon to be vacant" parking spaces, providing key data support for the real-time guidance system and significantly improving parking space turnover efficiency.

[0013] Preferably, in step S5, the updated parking space status data is aggregated and displayed, and the optimal path is calculated using a multi-objective weighted path planning algorithm to obtain the optimal path, which includes the following steps: All empty spaces and soon-to-be-empty spaces in the updated parking space status data are included in the candidate pool to obtain the candidate parking space set; Calculate the path distance and estimated travel time from the current location to each parking space in the candidate parking space set to obtain the candidate arrival dataset; assign weights to the empty spaces and those about to become available in the candidate parking space set, as well as time matching weights, to obtain weight data; and obtain the optimal path based on the candidate arrival dataset and the weight data. This invention includes real-time available parking spaces and soon-to-be-available parking spaces in a candidate set, comprehensively calculates multiple factors such as path distance, travel time, and walking distance, and uses a weighted algorithm to dynamically evaluate the overall value of each parking space, thereby planning the optimal parking route for car owners. This significantly reduces the time car owners spend searching for parking spaces and improves parking efficiency, while simultaneously achieving dynamic optimization of parking space resources and providing personalized services.

[0014] A parking space information monitoring system based on ultrasonic detection is used to implement the above-mentioned parking space information monitoring method based on ultrasonic detection. It includes a sensor initialization and calibration module, a multimodal data acquisition and processing module, a parking space status determination module, a behavior trajectory tracking and departure prediction module, and a data aggregation and path planning module. The sensor initialization and calibration module is used for the initial deployment and calibration of ultrasonic sensors. This includes installing sensor nodes directly above each parking space in the parking lot and creating an independent configuration file for each node in the central server. When the parking lot is vacant, the module ensures the accuracy and reliability of sensor data through an initial threshold calibration process. The module also supports periodic automatic calibration to cope with environmental changes and form a first set of parking space sensor files to provide basic data for subsequent monitoring. The multimodal data acquisition and processing module collects multimodal data, including distance data, echo signal strength, and signal-to-noise ratio (SNR) data, through ultrasonic sensors at fixed acquisition cycles. This data is processed by variance calculation to obtain stability index data, reflecting the data dispersion and measurement reliability. The module employs a low-power design, putting the sensor into sleep mode during non-acquisition periods to save energy while ensuring data continuity and stability, providing raw input for parking space status determination. The parking space status determination module determines the status of each parking space based on the first parking space sensor archive, processed multimodal data, and stability index data, by fusing judgment rules; it employs a filtering algorithm; and it preprocesses the data to reduce false alarms and interference, ultimately obtaining initial parking space status data; this module updates the parking space status in real time and records the status change timestamp, triggering subsequent behavior trajectory tracking and billing systems. The behavior trajectory tracking and departure prediction module analyzes parking space status change data to track the trajectory of vehicle parking behavior, and uses payment system signals, sensor pattern recognition, and geomagnetic disturbance data to predict departure. Based on the trajectory tracking and prediction results, the parking space status data is updated, parking spaces that are "about to be vacant" are marked, and a countdown threshold is set to handle vehicle departure situations. This improves the real-time performance and accuracy of parking space status updates and reduces misjudgments. The data aggregation and route planning module aggregates and displays updated parking space status data, and uses a multi-objective weighted route planning algorithm to calculate the optimal route. The algorithm comprehensively considers route distance, estimated travel time, walking distance, and parking space status weights to recommend the best parking space to the user. The system guides vehicles to park efficiently and updates parking space status information in real time, pushing it to platforms such as in-park guidance screens and mobile apps, improving parking lot management efficiency and user experience.

[0015] (III) Beneficial Effects The present invention has the following beneficial effects: This invention greatly improves the accuracy and reliability of parking space status monitoring. By using multimodal data, including distance, echo intensity, signal-to-noise ratio, and stability indicators, for fusion judgment, and combined with a carefully designed judgment rule sequence, the system can effectively distinguish different objects such as vehicles, pedestrians, and shopping carts. This significantly reduces the false alarms and missed alarms caused by instantaneous interference in traditional ultrasonic detection technology, and achieves high-precision identification of whether a parking space is occupied or vacant.

[0016] This invention achieves a leap from passive monitoring to active prediction of parking space status. The system not only focuses on the current status, but also verifies the rationality of parking events through behavioral trajectory tracking, and innovatively integrates payment signals, sensor "flash" patterns, and geomagnetic disturbance data to predict the vehicle's intention to leave. This enables the system to mark "soon to be vacant" parking spaces in advance, providing valuable data support for efficient parking guidance and resource scheduling, and improving the turnover efficiency of parking spaces.

[0017] This invention constructs a self-calibrating, continuously optimizing intelligent system. By establishing an independent profile for each sensor and implementing initial and periodic automatic calibration, the system can automatically compensate for sensor drift and environmental changes, ensuring long-term detection accuracy. This self-maintenance capability reduces subsequent maintenance costs and the frequency of manual intervention.

[0018] This invention deeply integrates monitoring data with user services, significantly improving user experience and parking lot operation efficiency. The system uses a multi-objective weighted path planning algorithm to dynamically recommend the best available or soon-to-be-available parking spaces to users and plan the optimal driving route. This not only reduces the time spent by car owners searching for parking spaces and alleviates congestion within the parking lot, but also maximizes the utilization of parking space information, ultimately providing a complete, intelligent, and efficient solution to the "parking difficulty" problem.

[0019] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a parking space information monitoring method based on ultrasonic detection according to the present invention. Figure 2 This is a schematic diagram of a parking space information monitoring system based on ultrasonic detection according to the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the invention, and not all embodiments. Based on the embodiments of the invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the invention.

[0023] In the description of this invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc., which indicate orientation or positional relationship, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements 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 on the invention.

[0024] Example 1: Please see Figure 1 This invention discloses a method for monitoring parking space information in a parking lot based on ultrasonic detection, comprising the following steps: S1. Establish an independent configuration file for each parking space ultrasonic sensor to obtain a parking space sensor file set; under the condition of the parking lot being vacant, obtain an initial judgment threshold through initial threshold calibration; store the initial judgment threshold into the parking space sensor file set to obtain the first parking space sensor file set. S1 includes the following steps: S11. Install ultrasonic sensor nodes directly above each parking space in the parking lot to obtain a set of ultrasonic sensor nodes for each parking space. S12. On the management platform of the central server, upload or draw the map of the parking lot to obtain the initial digital twin map; on the initial digital twin map, according to the actual installation location of each sensor node in the ultrasonic sensor node set of the parking space, record each sensor node and assign a unique identifier (such as NodeID: P1-A-05, which means parking space No. 5 in row A of area P1) to obtain the final digital twin map. Based on the final digital twin map, a stable communication connection is established between each sensor node in the parking space ultrasonic sensor node set and the regional gateway. In the central server, an independent configuration file is created for each sensor node, resulting in a parking space sensor file set. The parking space sensor file includes NodeID, installation location description, initial reference distance (Dbase initial) - to be calibrated, current judgment threshold (Tcurrent) - to be calculated, height compensation value (ΔH) - to be set, calibration status - uncalibrated, and communication status - online / offline. S13. When the parking lot is vacant, the initial reference distance is obtained by collecting and calculating the reference data; the initial judgment threshold is calculated based on the initial reference distance and the height compensation value; the initial judgment threshold is stored in the parking space sensor file set to obtain the first parking space sensor file set. S13 includes the following steps: S131. Conduct the calibration during the period when there are the fewest vehicles in the parking lot, such as from 02:00 to 04:00 at night. Ensure that there are no vehicles, people or other obstacles in all target parking spaces. If necessary, arrange staff to conduct on-site inspections and confirmations. The server administrator issues the initial calibration command for the entire site through the management platform. Upon receiving the initial calibration command for the entire site, the server broadcasts the calibration command to each sensor node in the parking space ultrasonic sensor node set through the gateway. After receiving the command, each node continuously and rapidly collects a set of distance readings (e.g., 100 times) under the condition that there are no cars directly below it, thus obtaining the distance reading set. The node uploads this raw data packet {NodeID, 100 distance readings} to the server. S132. After receiving the data, the server removes gross errors from the readings of each sensor node in the distance reading set and calculates the arithmetic mean to obtain an initial reference distance set; the initial reference distance set contains the distance from each sensor to the ground; The removal of gross errors uses the "Laida criterion (3σ criterion)" or other statistical methods to automatically identify and remove obviously impossible outliers (e.g., extremely short or extremely long distances caused by transient disturbances). In some embodiments, the mode can be used as the initial reference distance for stronger anti-interference capability, because the mode is less sensitive to occasional, non-continuous interference from small objects (such as flying insects passing by in an instant). S133. The administrator sets a reasonable "height compensation value" ΔH globally in the system based on the common vehicle types in the parking lot (such as mainly sedans or including SUVs). For example, the height of sedans is usually 1.4-1.5 meters. To ensure that all vehicles can be reliably detected, setting ΔH = 1.2 meters is a conservative and reliable choice. Based on the initial reference distance set, the server calculates an initial judgment threshold for each node in the ultrasonic sensor node set of the parking space using an initial judgment threshold formula, thus obtaining an initial judgment threshold set; the initial judgment threshold formula is as follows. T initial =D baseinitial -ΔH; Among them, T initial D represents the initial judgment threshold of the sensor node. baseinitial This represents the initial reference distance, and ΔH represents the height compensation; specifically, if the distance D from a sensor to the ground is the empty parking space... baseinitia If it is 3.0 meters, then T initial =3.0 - 1.2 = 1.8 meters; thereafter, the system will only start to consider the possibility of "a car" when the distance measured by the sensor is less than 1.8 meters; S134. The server writes the initial judgment threshold set and the initial reference distance set into the configuration file of each node in the parking space ultrasonic sensor node set, assigns the initial judgment threshold to the current judgment threshold, marking the current effective threshold of the node, updates the calibration status of the node to "calibrated", and obtains the first parking space sensor file set. After initial calibration is complete, a standard-sized vehicle can be parked sequentially in sample parking spaces in different zones; observe the status of the corresponding node on the server to determine whether it accurately changes from "empty space" to "car in use"; if there is any abnormality, check whether the D_base_initial of the node is reasonable, or check on-site whether the sensor installation is incorrect; The above process is a continuous update process. Specifically, the server automatically starts the calibration process during a preset low-traffic period (e.g., 3:00 AM daily); the server instructs all nodes to continuously collect 500 distance data points within 5 minutes; the server performs statistical analysis on the dataset uploaded by each node, removes obvious outliers, and calculates its mode as the new baseline distance D. basenew ; Calculate the new judgment threshold according to the initial judgment threshold formula and overwrite the current judgment threshold in the node file; Record the old threshold and the new threshold for this calibration for system maintenance and diagnosis; S2. Collect distance data, echo signal intensity, and signal-to-noise ratio (SNR) data using ultrasonic sensors to obtain multimodal data; process the multimodal data through variance calculation to obtain stability index data; S2 includes the following steps: S21. Set the acquisition period; a low-power timer is built into each sensor node in the parking space ultrasonic sensor node set, and a fixed acquisition period T is set (for example, T=1000 milliseconds, i.e. 1 second). During non-acquisition times, the node's main controller and ultrasonic module can enter sleep or low-power mode to save energy to the maximum extent; based on the acquisition period, distance data, echo signal strength, and signal-to-noise ratio (SNR) data are collected through the ultrasonic sensor to obtain multimodal data; The distance data represents the measured distance from the sensor node to the object below; the sensor obtains the straight-line distance to the object below by transmitting and receiving ultrasonic waves and calculating the time difference, in meters (m); for example, D = 2.35m; Echo signal intensity represents the strength of the reflected ultrasonic signal; the sensor also measures the amplitude or energy value of the received echo signal; this value is a relative quantity, usually expressed in analog voltage values ​​or digital units (e.g., 0-1023); it reflects the material, size, and surface structure of the reflecting object (metallic vehicles have strong echoes, while pedestrian legs have weak echoes). Signal-to-noise ratio (SNR) represents the ratio of the effective signal to the background noise. The chip or node microprocessor inside the sensor calculates the ratio of the effective echo signal to the ambient background noise, expressed in decibels (dB). A high SNR indicates a reliable measurement, while a low SNR indicates a noisy environment and low reliability of the measurement. S22. Set the number of continuous acquisition cycles; calculate the variance of the multimodal data under the number of continuous acquisition cycles to obtain the stability index data; The stability index directly reflects the dispersion of the data; the smaller the variance, the more stable the ranging results are over a number of consecutive collection periods. For example, if a car has come to a complete stop, the five ranging values ​​might be [1.52, 1.51, 1.53, 1.52, 1.52], with a very small variance; while if a pedestrian walks by, the data might be [1.55, 0.95, 1.80, 2.10, 1.60], with a very large variance or range. The sensor nodes transmit stability index data to the regional gateway via wireless communication; the regional gateway aggregates the data from all nodes within its jurisdiction and uploads it to the central processing server via the network. S3. Based on the first parking space sensor file set, multimodal data, and stability index data, determine whether there is a car in each parking space by fusing judgment rules to obtain the initial parking space status data; S3 includes the following steps: S31. Set stability threshold and signal strength threshold; based on stability threshold and signal strength threshold, set fusion judgment rules; S31 includes the following steps: S311. Set vehicle presence determination rules; the vehicle presence determination rules are that the distance data is less than the initial judgment threshold, the stability data is greater than the stability threshold, and the echo signal strength is greater than the signal strength threshold; according to the interpretation of the above rules, the object is close, the data is stable, and the echo is strong, which is consistent with the characteristics of a stationary vehicle; S312. Set interference / invalidation rules; the dynamic interference / invalidation rules and static interference / invalidation rules; the dynamic interference / invalidation rule is that the distance data is less than the initial judgment threshold and the stability data is less than the stability threshold; according to the above rule, although the distance is close, the data fluctuates greatly, which is likely due to a pedestrian or small animal crossing, and is judged as interference, maintaining the "empty" state; the static interference / invalidation rule is that the distance data is less than the initial judgment threshold, the stability data is less than the stability threshold, and the echo signal strength is less than the signal strength threshold; according to the above rule, although the distance is close, the signal is weak and the noise is high, which may be due to a low shopping cart or suitcase, and is judged as interference, maintaining the "empty" state; S313. Set a vacancy rule; the vacancy rule is that the distance data is greater than the initial judgment threshold; if the distance data is greater than the initial judgment threshold, it is clearly a vacancy. S314. Set the order of rule execution; the order of rule execution is as follows: empty space rule is judged first, followed by dynamic interference / invalid rule, then static interference / invalid rule, and finally the vehicle detection rule is triggered. S32. A filtering algorithm is used to preprocess the multimodal data to obtain processed multimodal data. Specifically, the server maintains a small historical data buffer for each node, storing distance data from the most recent M (e.g., M=3) received multimodal data. When new distance data arrives, the filter sorts the M data points, including the new value, and takes the median value as the effective distance value for this processing. For example, the original data sequence (unit: meters) is [1.55, 1.53, 0.45, 1.54, 1.53], where 0.45 may be a false alarm (such as interference from flying insects). After sorting: [0.45, 1.53, 1.53, 1.54, 1.55]; the median outputs the effective distance data as 1.53 meters. This effectively smooths the data and prevents a single false alarm from directly affecting the system's judgment. S33. Based on the first parking space sensor file set, the processed multimodal data, and the stability index data, determine whether there is a car in each parking space by fusing judgment rules, obtain the judgment result, update the status of each parking space in the parking lot according to the judgment result, and obtain the initial parking space status data. Specifically, based on the above judgment results, the server updates the current status field ("occupied" or "vacant") of the NodeID in the parking space status database; simultaneously, it records the status change timestamp and the data source that triggered this status (e.g., recording "occupied" triggered by rule A); if the status changes from "vacant" to "occupied", the system records the parking start time; triggers the subsequent "behavior trajectory tracking" logic; if the status changes from "occupied" to "vacant", the system calculates the parking duration, updates the database, and releases the "soon to be vacant" flag (if any); triggers the billing system to settle accounts (if linked); the server pushes the latest number and location information of vacant parking spaces, as well as the "soon to be vacant" parking space information, to the in-park guidance screens, official website, and mobile APP in real time, completing a complete monitoring-judgment-release cycle; S4. Update the initial parking space status data by tracking behavior trajectory and predicting departure to obtain the updated parking space status data; S4 includes the following steps: S41. Extract parking space data from the initial parking space status data that shows parking spaces changing from empty to occupied, to obtain changed parking space data; the server uses the final digital twin map to find sensor nodes that are spatially adjacent to the parking spaces in the changed parking space data and that may be passed by vehicles, to obtain an initial set of adjacent sensor nodes; for example, for a parking space at the end of a lane, its neighboring nodes may include node A in the lane in front of it and node B on its left. Query the status change records of parking spaces in the initial adjacent sensor node set within the most recent time window (e.g., the past Δt = 10 seconds) to obtain node change data; Set straight-ahead parking rules and turning parking rules; the straight-ahead parking rule is that [lane node A state change: "empty" -> "available" -> "empty"] occurs before [parking space X state change: "empty" -> "available"], and the time difference between the two is within a reasonable range (e.g., 1-3 seconds); the turning parking rule is [node B (left side) "empty" -> "available" -> "empty"] -> [node A (front) "empty" -> "available" -> "empty"] -> [parking space X "empty" -> "available"]; The system queries the node change data to see if there is data that conforms to the straight-ahead parking rule and the turning parking rule, thus obtaining trajectory tracking data. If such a logically consistent continuous triggering pattern exists, the system confirms that the "car present" state of parking space X is a valid "parking behavior." The system can record this trajectory log for data analysis. If no clear coordinated triggering pattern is found (for example, parking space X suddenly becomes "car present" while surrounding nodes do not react), the system will mark this event as "requiring attention." This may mean that the sensor is falsely reporting, or that a vehicle that has been parked for a long time has just been correctly identified by the system (such as after initialization). The system may maintain the "car present" state but will not record it as valid. S42. Departure prediction data is obtained by using payment departure prediction, sensor pattern recognition, and geomagnetic disturbance. S42 includes the following steps: S421. The central server and the parking lot's payment system (which can be a self-service payment machine, a contactless payment platform, or a mobile APP payment backend) establish a secure data connection through an API interface. When a payment signal is received, the system collects vehicle information and determines that the parking space will soon be vacant. The vehicle information includes the license plate number, payment time, and the corresponding parking space number X (or can be obtained by the payment system by looking up the license plate). S422. Set the flickering phenomenon rule and the geomagnetic disturbance rule; the flickering phenomenon rule is that for an occupied parking space, the sensor data changes from a car to no car and then back to a car within 2 seconds; the principle analysis is that this "flickering" is likely due to engine vibration or slight displacement of the vehicle body when the vehicle starts, causing the ultrasonic waves to fail to be effectively reflected back to the sensor at a certain moment, generating a false signal similar to "empty space", which is a very strong pre-departure signal; The geomagnetic disturbance rule is that a rapid and continuous disturbance occurs to the geomagnetic field of an occupied parking space. The principle analysis is that the movement of the vehicle's metal body or the start of the engine will change the surrounding magnetic field environment. This disturbance mode is significantly different from that when the vehicle is stationary. The system continuously monitors the sensor data of all parking spaces that are occupied to obtain vehicle occupancy data; when the vehicle occupancy data shows a flashing phenomenon rule and a geomagnetic disturbance rule, the system collects vehicle information and determines that the parking space will soon be vacant. S43. Update the initial parking space status data based on trajectory tracking data and departure prediction data to obtain updated parking space status data; specifically, based on the parking space number in the vehicle information, change the parking space status marker in the parking space status database of the first parking space sensor file set to "about to be vacant". Set a countdown threshold; if the sensor detects that the vehicle has left within the countdown threshold, the system clears the "about to be vacant" marker and changes the status to "vacant". If the countdown threshold is exceeded and the vehicle has not left, the system may restore the parking space status to occupied and may trigger an alarm message to notify the administrator that a vehicle may have exceeded the parking time limit. S5. The updated parking space status data is aggregated and displayed, and the optimal route is planned to obtain the optimal route; when the vehicle parks in the parking space according to the optimal route, the updated parking space status data is updated again. In step S5, the updated parking space status data is aggregated and displayed, and the optimal path is calculated using a multi-objective weighted path planning algorithm. The optimal path is obtained through the following steps: All empty spaces and soon-to-be-empty spaces in the updated parking space status data are included in the candidate pool to obtain the candidate parking space set; Calculate the path distance and estimated travel time from the current location to each parking space in the candidate parking space set to obtain the candidate arrival dataset; assign weights to the empty spaces and those about to become available in the candidate parking space set, as well as time matching weights, to obtain weight data; and obtain the optimal path based on the candidate arrival dataset and the weight data. Specifically, at the parking lot entrance, the system finds one empty parking space P and one soon-to-be-empty parking space Q; the candidate parking space set = {empty parking space P, soon-to-be-empty parking space Q}; to empty parking space P: driving distance 200 meters, time 3 minutes, walking distance 50 meters; to soon-to-be-empty parking space Q: driving distance 100 meters, time 1.5 minutes, walking distance 80 meters; Each indicator is converted into a score of 0-1, with higher values ​​being better; Distance score = 1 - (distance to this parking space / distance to the furthest parking space), Empty space P distance score = 1 - 200 / 200 = 0, i.e., empty space Q distance score = 1 - 100 / 200 = 0.5; Time score = 1 - (time to this parking space / time to the furthest parking space); Empty space P time score = 1 - 3 / 3 = 0; i.e., empty space Q time score = 1 - 1.5 / 3 = 0.5; Walking score = 1 - (walking distance to this parking space / walking distance to the furthest parking space); Empty space P walking score = 1 - 50 / 80 = 0.375; i.e., empty space Q walking score = 1 - 80 / 80 = 0; Overall score = State weight × (Distance weight × Distance score + Time weight × Time score + Walking weight × Walking score); Assuming weight allocation: Distance weight = 0.3, Time weight = 0.4, Walking weight = 0.3; State weight of vacant spot P = 1.0 (fixed value); State weight of soon-to-be-vacant spot Q = 0.7 + 0.5 × Time matching factor, Time matching factor = Maximum value (0, 1 - |Estimated arrival time - Estimated vacancy time| / Buffer time), Estimated arrival time of soon-to-be-vacant spot Q = Current time + 1.5 The estimated idle time is 2 minutes, and the time matching factor is 1 - |1.5 - 2| / 2 = 0.75. Therefore, the state weight of the idle position Q is 0.7 + 0.5 × 0.75 = 1.075. The final score calculation for the idle position P is 1.0 × (0.3 × 0 + 0.4 × 0 + 0.3 × 0.375) = 0.1125; the score for the idle position Q is 1.075 × (0.3 × 0.5 + 0.4 × 0.5 + 0.3 × 0) = 1.075 × 0.35 = 0.376. Since the score of the soon-to-be-vacant parking space Q (0.376) is greater than the score of the vacant parking space P (0.1125), the system plans the soon-to-be-vacant parking space Q as the optimal path; the system will guide the user to the parking space and indicate that the parking space is expected to be vacant when you arrive.

[0025] Example 2: Please see Figure 2 A parking space information monitoring system based on ultrasonic detection is used to implement the above-mentioned parking space information monitoring method based on ultrasonic detection. It includes a sensor initialization and calibration module, a multimodal data acquisition and processing module, a parking space status determination module, a behavior trajectory tracking and departure prediction module, and a data aggregation and path planning module. The sensor initialization and calibration module is used for the initial deployment and calibration of ultrasonic sensors. This includes installing sensor nodes directly above each parking space in the parking lot and creating an independent configuration file for each node in the central server. When the parking lot is vacant, the module ensures the accuracy and reliability of sensor data through an initial threshold calibration process. The module also supports periodic automatic calibration to cope with environmental changes and form a first set of parking space sensor files to provide basic data for subsequent monitoring. The multimodal data acquisition and processing module collects multimodal data, including distance data, echo signal strength, and signal-to-noise ratio (SNR) data, through ultrasonic sensors at fixed acquisition cycles. This data is processed by variance calculation to obtain stability index data, reflecting the data dispersion and measurement reliability. The module employs a low-power design, putting the sensor into sleep mode during non-acquisition periods to save energy while ensuring data continuity and stability, providing raw input for parking space status determination. The parking space status determination module determines the status of each parking space based on the first parking space sensor archive, processed multimodal data, and stability index data, by fusing judgment rules; it employs a filtering algorithm; and it preprocesses the data to reduce false alarms and interference, ultimately obtaining initial parking space status data; this module updates the parking space status in real time and records the status change timestamp, triggering subsequent behavior trajectory tracking and billing systems. The behavior trajectory tracking and departure prediction module analyzes parking space status change data to track the trajectory of vehicle parking behavior, and uses payment system signals, sensor pattern recognition, and geomagnetic disturbance data to predict departure. Based on the trajectory tracking and prediction results, the parking space status data is updated, parking spaces that are "about to be vacant" are marked, and a countdown threshold is set to handle vehicle departure situations. This improves the real-time performance and accuracy of parking space status updates and reduces misjudgments. The data aggregation and route planning module aggregates and displays updated parking space status data, and uses a multi-objective weighted route planning algorithm to calculate the optimal route. The algorithm comprehensively considers route distance, estimated travel time, walking distance, and parking space status weights to recommend the best parking space to the user. The system guides vehicles to park efficiently and updates parking space status information in real time, pushing it to platforms such as in-park guidance screens and mobile apps, improving parking lot management efficiency and user experience.

[0026] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. 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.

[0027] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for monitoring parking space information in a parking lot based on ultrasonic detection, characterized in that, Includes the following steps: S1. Establish an independent configuration file for each parking space ultrasonic sensor to obtain a parking space sensor file set; under the condition of the parking lot being vacant, obtain an initial judgment threshold through initial threshold calibration; store the initial judgment threshold into the parking space sensor file set to obtain the first parking space sensor file set. S2. Collect distance data, echo signal intensity, and signal-to-noise ratio (SNR) data using ultrasonic sensors to obtain multimodal data; process the multimodal data through variance calculation to obtain stability index data; S3. Based on the first parking space sensor file set, multimodal data, and stability index data, determine whether there is a car in each parking space by fusing judgment rules to obtain the initial parking space status data; S4. Update the initial parking space status data by tracking behavior trajectory and predicting departure to obtain the updated parking space status data; S5. The updated parking space status data is aggregated and displayed, and the optimal path is planned to obtain the optimal path. When the vehicle parks in the parking space according to the optimal path, the updated parking space status data is updated again.

2. The parking space information monitoring method based on ultrasonic detection according to claim 1, characterized in that, S1 includes the following steps: S11. Install ultrasonic sensor nodes directly above each parking space in the parking lot to obtain a set of ultrasonic sensor nodes for each parking space. S12. On the management platform of the central server, upload or draw the map of the parking lot to obtain the initial digital twin map; on the initial digital twin map, according to the actual installation location of each sensor node in the ultrasonic sensor node set of the parking space, record each sensor node and assign a unique identifier to obtain the final digital twin map. Based on the final digital twin map, a stable communication connection is established between each sensor node in the parking space ultrasonic sensor node set and the regional gateway. In the central server, an independent configuration file is created for each sensor node to obtain the parking space sensor file set. S13. When the parking lot is vacant, the initial reference distance is obtained by collecting and calculating the reference data; the initial judgment threshold is calculated based on the initial reference distance and the height compensation value; the initial judgment threshold is stored in the parking space sensor file set to obtain the first parking space sensor file set.

3. The parking space information monitoring method based on ultrasonic detection according to claim 2, characterized in that, S13 includes the following steps: S131. After ensuring that there are no vehicles, people or other obstacles in all target parking spaces, issue an initial calibration command for the entire site through the management platform. After receiving the initial calibration command for the entire site, the server broadcasts the calibration command to each sensor node in the ultrasonic sensor node set of the parking space through the gateway; after receiving the command, each node collects a set of distance readings to obtain the distance reading set. S132. Remove gross errors from the readings of each sensor node in the distance reading set and calculate the arithmetic mean to obtain an initial reference distance set; the initial reference distance set contains the distance from each sensor to the ground; S133. Based on the common vehicle types in the parking lot and the initial reference distance set, calculate the initial judgment threshold for each node in the ultrasonic sensor node set of the parking space using the initial judgment threshold formula to obtain the initial judgment threshold set. S134. Write the initial judgment threshold set and the initial reference distance set into the configuration file of each node in the parking space ultrasonic sensor node set to obtain the first parking space sensor file set.

4. The parking space information monitoring method based on ultrasonic detection according to claim 1, characterized in that, S2 includes the following steps: S21. Set the acquisition period; based on the acquisition period, collect distance data, echo signal intensity, and signal-to-noise ratio (SNR) data through the ultrasonic sensor to obtain multimodal data; S22. Set the number of continuous acquisition cycles; calculate the variance of the multimodal data under the number of continuous acquisition cycles to obtain the stability index data.

5. The parking space information monitoring method based on ultrasonic detection according to claim 1, characterized in that, S3 includes the following steps: S31. Set stability threshold and signal strength threshold; based on stability threshold and signal strength threshold, set fusion judgment rules; S32. Use a filtering algorithm to preprocess the multimodal data to obtain the processed multimodal data; S33. Based on the first parking space sensor file set, the processed multimodal data, and the stability index data, determine whether there is a car in each parking space by fusing judgment rules, obtain the judgment result, update the status of each parking space in the parking lot according to the judgment result, and obtain the initial parking space status data.

6. The parking space information monitoring method based on ultrasonic detection according to claim 5, characterized in that, S31 includes the following steps: S311. Set a vehicle presence determination rule; the vehicle presence determination rule is that the distance data is less than the initial determination threshold, the stability data is greater than the stability threshold, and the echo signal strength is greater than the signal strength threshold. S312, Set interference / invalidation rules; the dynamic interference / invalidation rules and the static interference / invalidation rules; the dynamic interference / invalidation rules are when the distance data is less than the initial judgment threshold and the stability data is less than the stability threshold; the static interference / invalidation rules are when the distance data is less than the initial judgment threshold, the stability data is less than the stability threshold, and the echo signal strength is less than the signal strength threshold; S313. Set a vacancy rule; the vacancy rule is that the distance data is greater than the initial judgment threshold. S314. Set the order of rule execution; the order of rule execution is as follows: empty space rule is judged first, followed by dynamic interference / invalid rule, then static interference / invalid rule, and finally the vehicle detection rule is triggered.

7. The parking space information monitoring method based on ultrasonic detection according to claim 1, characterized in that, S4 includes the following steps: S41. Extract parking space data from the initial parking space status data that shows parking spaces changing from empty to occupied, to obtain changed parking space data; based on the final digital twin map, find sensor nodes that are spatially adjacent to the parking spaces in the changed parking space data and that may be passed by vehicles, to obtain the initial set of adjacent sensor nodes. Query the status change records of parking spaces in the initial adjacent sensor node set within the most recent time window to obtain node change data; Set the straight-ahead parking rules and the turning parking rules; query the node change data to see if there is data that matches the straight-ahead parking rules and the turning parking rules, and obtain the trajectory tracking data; S42. Departure prediction data is obtained by using payment departure prediction, sensor pattern recognition, and geomagnetic disturbance. S43. Update the initial parking space status data based on trajectory tracking data and departure prediction data to obtain updated parking space status data.

8. A parking space information monitoring method based on ultrasonic detection according to claim 7, characterized in that, S42 includes the following steps: S421. Establish a secure data connection with the parking lot's payment system. When a payment signal is received, collect vehicle information and determine that the parking space will soon be vacant. S422. Set the flickering phenomenon rule and the geomagnetic disturbance rule; the flickering phenomenon rule is that for an occupied parking space, the sensor data changes from a car present to a car absent and then back to a car present within 2 seconds; The geomagnetic disturbance rule is that a rapid and continuous disturbance occurs in the geomagnetic field of an occupied parking space; Continuously monitor the sensor data of all parking spaces that are occupied to obtain vehicle occupancy data; when the vehicle occupancy data shows a flashing phenomenon rule and a geomagnetic disturbance rule, collect vehicle information and determine that the parking space will soon be vacant.

9. A parking space information monitoring method based on ultrasonic detection according to claim 1, characterized in that, In step S5, the updated parking space status data is aggregated and displayed, and the optimal path is calculated using a multi-objective weighted path planning algorithm. The optimal path is obtained through the following steps: All empty spaces and soon-to-be-empty spaces in the updated parking space status data are included in the candidate pool to obtain the candidate parking space set; Calculate the path distance and estimated travel time from the current location to each parking space in the candidate parking space set to obtain the candidate arrival dataset; assign weights to the vacant spaces and those about to become vacant in the candidate parking space set, as well as time matching weights, to obtain weight data; and obtain the optimal path based on the candidate arrival dataset and the weight data.

10. A parking lot space information monitoring system based on ultrasonic detection, characterized in that, The parking space information monitoring method based on ultrasonic detection as described in any one of claims 1-9 includes a sensor initialization and calibration module, a multimodal data acquisition and processing module, a parking space status determination module, a behavior trajectory tracking and departure prediction module, and a data aggregation and path planning module.

Images