Vehicle safety inspection equipment dispatching method based on intelligent parking lot

Abstract

The application discloses a vehicle safety inspection equipment scheduling method based on a smart parking lot, and relates to the technical field of smart parking; first, basic space data of the parking lot is acquired, a parking lot space topology model is constructed, and the parking lot is divided to form a plurality of monitoring areas; then, quantitative analysis is performed on key parameters of the monitoring areas to obtain environmental risk weights of the monitoring areas; then, vehicle risk factors are obtained in combination with vehicle stay duration, vehicle categories and abnormal activities near the vehicles, and the vehicle risk factors are risk-weighted based on the environmental risk weights to generate vehicle inspection demand indexes, and inspection task clusters are formed according to the inspection demand indexes; and finally, the inspection task clusters are dynamically distributed to corresponding inspection equipment based on dynamic matching between the inspection equipment and the inspection task clusters, so that the timeliness of inspection of high-risk vehicles, the utilization rate of inspection resources and the continuous inspection capability in complex scenes are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of smart parking technology, specifically to a method for scheduling vehicle safety inspection equipment based on smart parking lots. Background Technology

[0002] As the level of smart parking lot construction continues to improve, parking lots are gradually being equipped with license plate recognition equipment, video surveillance equipment, parking space detection equipment, environmental sensing equipment, and mobile inspection equipment to achieve real-time perception of the parking status, traffic status, and environmental status of vehicles within the lot. Especially in scenarios such as large commercial complexes, transportation hubs, residential communities, and logistics parks, where there are a large number of vehicles, scattered parking areas, and varying parking durations, some areas may also have safety hazards such as insufficient nighttime lighting, complex pedestrian flow, blocked fire lanes, unlocked vehicles, and open windows, periodic vehicle inspections are of great significance.

[0003] Current parking lot inspections typically employ manual patrols or inspection robots that patrol along fixed routes and at fixed time intervals. Manual patrols are inefficient, heavily reliant on the experience of the inspectors, and difficult to cover all vehicles in the parking lot. Fixed-route patrols are prone to overlooking high-risk vehicles and key areas, and cannot be dynamically adjusted based on vehicle parking duration, environmental changes, and abnormal events. Summary of the Invention

[0004] In response to the problems in related technologies, this invention provides a method for scheduling vehicle safety inspection equipment based on smart parking lots to overcome the aforementioned technical problems in existing related technologies.

[0005] To solve the aforementioned technical problem, the present invention is achieved through the following technical solution:

[0006] A method for scheduling vehicle safety inspection equipment based on smart parking lots, characterized by the following steps:

[0007] Step 1: Obtain basic spatial data of the parking lot, and perform spatial topology modeling of the parking lot based on the basic spatial data to divide it into several monitoring areas;

[0008] Step 2: According to the preset monitoring cycle, continuously collect and quantitatively analyze the status of the monitoring area and the vehicle parking status in the parking lot to obtain the environmental risk weight and vehicle risk factor of each monitoring area. Based on the environmental risk weight corresponding to the monitoring area where the vehicle is located, conduct risk gain analysis on the vehicle risk factor to obtain the vehicle inspection demand index. Then, form an inspection task cluster based on the inspection demand index.

[0009] Step 3: Collect the current location, current load, and historical inspection records of the inspection equipment. Based on this, analyze the final matching degree of the inspection equipment with the inspection task cluster, and assign the inspection equipment with the highest final matching degree and meeting the power constraints to the corresponding inspection task cluster; the current load refers to the inspection task cluster that the inspection equipment has been assigned to.

[0010] Step 4: Determine the inspection sequence based on the inspection demand index of vehicles within the assigned inspection task clusters by the inspection equipment, and select the shortest accessible route between adjacent inspection objects.

[0011] Preferably, in step one, dividing the parking lot into several monitoring zones includes:

[0012] The basic spatial data includes at least the parking space layout map, lane topology map, entrance and exit locations, fire lane locations, surveillance camera locations, environmental sensor locations, and access path information for inspection equipment.

[0013] The basic spatial data of the parking lot is mapped to a unified spatial coordinate system, and the boundaries of parking spaces, lanes, fire lanes, entrances and exits, and the passage paths of inspection equipment are vectorized. The parking lot is initially divided into areas using the main lanes, branch lanes, circular lanes, ramps, and fire lanes as primary area division boundaries. Based on the coverage of surveillance cameras, the coverage of environmental sensors, and the continuous reachable paths of inspection equipment, adjacent parking spaces are merged or split into monitoring areas.

[0014] Preferably, in step two, the quantitative analysis of the status of the monitored area includes:

[0015] For monitoring area i, where i represents the index of the monitoring area within the parking lot; obstacles within monitoring area i are identified using video surveillance sensors, and the number of obstacles and the total area occupied by obstacles are counted; the total length and feasible length of the inspection passage within monitoring area i are calculated, where the feasible length refers to the length of the passage that ensures the normal passage of the inspection equipment at the current moment, and the total area occupied by obstacles refers to the sum of the areas obstructed by temporary parking objects, construction barriers, debris, vehicles occupying the road, etc.; then the passage congestion degree T of monitoring area i is expressed as:

[0016]

[0017] Among them, L free L is the passable length of the passage, and A is the total length of the inspection passage within the monitoring area. obs A represents the total area occupied by obstacles. path N represents the total area of ​​the inspection channels within the monitoring area. max To pre-determine the normalized upper limit of the number of obstacles, N obsThis represents the number of obstacles in the monitoring area; η1, η2, and η3 are weighting coefficients, and satisfy: η1+η2+η3=1. In this example, η1, η2, and η3 are taken as 0.4, 0.3, and 0.3, respectively.

[0018] For monitoring area i, calculate the total area of ​​the entire monitoring area. At the same time, take the area that can be covered by video surveillance within the monitoring area as the effective monitoring area. Then, subtract the effective monitoring area from the total area of ​​the entire monitoring area and divide by the total area of ​​the entire monitoring area to obtain the monitoring blind zone degree of monitoring area i. The monitoring blind zone degree is used to characterize the size of the monitoring blind zone.

[0019] For monitoring area i, the number of people and vehicles passing through monitoring area i during the monitoring period is counted. Then the dynamic interference degree F is expressed as:

[0020]

[0021] Where ρ1 and ρ2 are weighting coefficients, and satisfy ρ1 + ρ2 = 1. In this embodiment, ρ1 and ρ2 are taken as 0.6 and 0.4 respectively, Δt represents the monitoring duration of this monitoring cycle, and N per To determine the number of people i in the monitored area, N veh This indicates the number of vehicles passing through the monitored area; the dynamic interference level is used to characterize the intensity of dynamic interference within the area caused by pedestrian and vehicle traffic.

[0022] For monitoring area i, smoke concentration, temperature, and humidity are acquired. Temperature and humidity deviations are calculated by comparing them to preset normal temperature and humidity thresholds, respectively. These deviations are then normalized to ensure they are on the same scale. Finally, they are linearly weighted and fused according to preset weights to obtain the environmental anomaly score, with the weights summed to one. The environmental anomaly score characterizes the impact of environmental anomalies such as smoke, temperature, and humidity on safety inspections. Smoke concentration is used to identify fires, smoke, or abnormal equipment heating; temperature deviation is used to identify localized temperature rise anomalies; and humidity deviation is used to identify problems such as dampness, condensation, or abnormal ventilation. In this embodiment, the weights for smoke concentration, temperature deviation, and humidity deviation are set to 0.5, 0.25, and 0.25, respectively, to prioritize the identification of smoke anomalies. A higher environmental anomaly score indicates a greater likelihood of potential safety hazards in the area, thus requiring a higher inspection priority.

[0023] The channel blockage degree, monitoring blind spot degree, dynamic interference degree, and environmental anomaly degree are then normalized to obtain standardized parameters under a unified dimension. Based on the importance of each parameter, a weighted fusion is performed to obtain the environmental risk weight of the monitoring area. In this embodiment, the weights of the channel blockage degree, monitoring blind spot degree, dynamic interference degree, and environmental anomaly degree are set to 0.30, 0.25, 0.15, and 0.30, respectively.

[0024] Preferably, in step two, the quantitative analysis of the vehicle's stationary status includes:

[0025] The system acquires vehicle entry time, vehicle type, and information on target activity near the vehicle; it obtains a vehicle dwell time factor based on the time difference between the vehicle entry time and the current time; it obtains a vehicle type factor based on the vehicle type; it constructs a vehicle neighborhood centered on the target vehicle and identifies the trajectory of targets entering the neighborhood, obtaining the minimum approach distance, dwell time, average approach speed, and number of repeated approaches corresponding to the trajectory, and calculates the trajectory abnormal activity factor; it summarizes all trajectory abnormal activity factors within the monitoring period to obtain the vehicle vicinity abnormal activity factor; and it weights and fuses the dwell time factor, vehicle type factor, and vehicle vicinity abnormal activity factor to obtain the vehicle risk factor.

[0026] Preferably, the vehicle category includes at least ordinary vehicles, new energy vehicles, and high-value vehicles. Values ​​are assigned according to vehicle type. Specifically, ordinary vehicles, new energy vehicles, and high-value vehicles correspond to different type values, with the type value of ordinary vehicles < the type value of new energy vehicles < the type value of high-value vehicles. In this embodiment, these values ​​are set to 0.3, 0.6, and 0.9, respectively.

[0027] The monitoring radius of the vehicle's neighborhood is determined based on the target vehicle's dimensions, the distance between adjacent parking spaces, and the safety buffer distance; the major axis dimension of the target vehicle's circumscribed rectangle is obtained. The monitoring radius is calculated using the formula. δ1 is the safety buffer distance around the vehicle, and δ2 is the target recognition compensation distance. In this embodiment, the values ​​are 0.5m and 0.3m, respectively. The safety buffer distance is used to cover the safety recognition range outside the true boundary of the vehicle, and the target recognition compensation distance is used to compensate for the deviation of the vehicle recognition box, the position calibration error and the trajectory extraction error.

[0028] Preferably, the generation of the inspection task cluster in step two includes:

[0029] Vehicles are classified into high-priority, medium-priority, and low-priority inspection targets based on the vehicle inspection demand index. When multiple vehicles are located in the same or adjacent monitoring areas, their inspection demand indices are in the same priority range, and the spatial distance between the vehicles is less than a preset distance threshold, they are grouped into the same inspection task cluster. When the vehicle inspection demand index is greater than or equal to a preset high-risk threshold, the corresponding vehicles are formed into independent inspection task clusters.

[0030] Preferably, the calculation of the final matching degree in step three includes:

[0031] The shortest path length from the current location of the inspection equipment to the target inspection task cluster is calculated based on the parking lot spatial topology model, and the distance matching degree is obtained by normalization. The number of inspection task clusters currently assigned to the inspection equipment and the cumulative duration of the current inspection tasks are counted, and the load matching degree is calculated. The historical inspection records of the inspection equipment are obtained, and the execution coefficient is obtained based on the packet loss rate, inspection completion rate, rescheduling rate and failure rate. Then, the running history coefficient is obtained after time decay. The distance matching degree, load matching degree and running history coefficient are weighted and fused to obtain the final matching degree.

[0032] Preferably, the load matching degree is determined by the number of task clusters to be inspected and the cumulative duration, and the two are fused according to preset weights, as shown in the following formula:

[0033]

[0034] Where M represents the number of task clusters to be inspected, τ represents the cumulative duration of the inspection equipment currently executing inspection tasks, γ1 and γ2 are the weights of the number of task clusters to be inspected and the cumulative duration, respectively, and γ1+γ2=1, which are taken as 0.6 and 0.4 in this embodiment, respectively; Mmax and τmax are the normalized upper limits of the number of task clusters to be inspected and the cumulative duration, respectively; the load matching degree is used to characterize the current resource occupancy of the inspection equipment, the higher the load, the lower the matching degree; the running history coefficient is obtained by time decay based on the execution coefficient corresponding to the historical monitoring cycle; the time decay calculation formula is:

[0035]

[0036] in The time interval between the monitoring time of detection period j and the current time is represented by λ, which is the time decay coefficient and λ > 0. The running history coefficient is used to characterize the continuous execution capability of the inspection equipment in the most recent historical period and serves as the basis for the allocation of subsequent inspection task clusters. After a task cluster is allocated, the load matching degree of the relevant inspection equipment is updated, and the final matching degree of the remaining inspection task clusters is recalculated.

[0037] The present invention has the following beneficial effects:

[0038] 1. By constructing a parking lot spatial topology model based on basic parking lot spatial data, and combining the coverage of video surveillance, environmental sensors, and continuous reachable paths of inspection equipment, the parking lot is divided into monitoring areas. This creates a unified and clear boundary relationship between the parking lot monitoring areas in terms of spatial structure, sensor coverage, and inspection accessibility. This avoids problems such as incomplete coverage by inspection equipment, repeated detours, or distorted area division. It provides unified regional basic data support for subsequent environmental risk analysis, vehicle risk identification, and inspection task clustering, improving the accuracy and continuity of subsequent risk assessment and task scheduling.

[0039] 2. By synchronously collecting multi-dimensional status data of the monitoring area, vehicles, and inspection equipment within a preset monitoring period, and quantitatively analyzing environmental risks, vehicle risks, and inspection needs in the monitoring area, dynamic risk identification and hierarchical management of parking lot inspection targets are achieved. Specifically, by integrating and analyzing channel congestion, blind spot coverage, dynamic interference, and environmental anomalies, the real-time risk status of different areas of the parking lot can be accurately reflected. Furthermore, by combining vehicle dwell time, vehicle type, and abnormal activity around the vehicle, potentially high-risk vehicles can be further identified. Based on this, environmental risk weights are used to apply risk amplification to vehicle risk factors, forming an inspection demand index. Inspection task clusters are then generated based on spatial proximity and risk priority, transforming the traditional fixed-frequency inspection method into an intelligent inspection method based on dynamic risk generation. This effectively avoids the problem of low-risk areas occupying excessive inspection resources, allowing high-risk vehicles and high-risk areas to receive higher inspection priority, thereby improving the real-time performance, targeting, and resource utilization efficiency of parking lot security inspections.

[0040] 3. By comprehensively analyzing the matching relationship between inspection equipment and inspection task clusters based on the current location of the inspection equipment, task load status, and historical operation history, intelligent and dynamic allocation of inspection tasks is achieved. Specifically, distance matching ensures that inspection equipment prioritizes responses to nearby tasks, reducing inspection response time; load matching prevents overload of individual inspection equipment, improving the overall balance of inspection resources; and the operation history coefficient dynamically evaluates the actual execution capability of inspection equipment by combining recent communication stability, task completion ability, and fault status, thus preventing poorly performing equipment from undertaking high-risk tasks. During task allocation, the system can also update equipment load and recalculate the matching degree of remaining tasks in real time after task allocation, thereby avoiding subsequent scheduling imbalances caused by the allocation of previous tasks. The system can dynamically adjust the inspection task allocation strategy based on the real-time risk status of the parking lot and the equipment operating status, improving the continuity, stability, and overall scheduling efficiency of the inspection system in complex scenarios.

[0041] 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

[0042] 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.

[0043] Figure 1 The present invention provides a flowchart of a method for scheduling vehicle safety inspection equipment based on a smart parking lot. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] Application Scenarios: Existing parking lot inspections typically rely on manual patrols or inspection robots that patrol along fixed routes and at fixed time intervals. Manual patrols are inefficient, heavily reliant on the experience of the inspectors, and struggle to cover all vehicles in the parking lot. Fixed-route patrols easily overlook high-risk vehicles and key areas, and cannot dynamically adjust based on vehicle parking duration, environmental changes, and abnormal events. To address these technical issues, this solution is applied to smart parking lots equipped with parking space sensing, video surveillance, environmental monitoring, and mobile inspection equipment. Inspection equipment includes, but is not limited to, mobile inspection robots, inspection vehicles with cameras, unmanned inspection platforms, or other devices with mobile inspection capabilities. By collecting data through sensors, and then performing data aggregation, risk analysis, task generation, equipment matching, and dynamic scheduling, this solution enables refined identification, prioritization, and intelligent scheduling of parking lot vehicle safety inspection tasks. This allows inspection equipment to move beyond the crude operation of fixed routes and frequencies, dynamically allocating tasks based on vehicle risk, equipment status, and environmental changes. This improves the efficiency of high-risk vehicle detection, reduces inspection resource waste, and enhances the parking lot's safety inspection capabilities in complex scenarios.

[0046] like Figure 1 As shown, this embodiment of the invention provides a method for scheduling vehicle safety inspection equipment based on smart parking lots, specifically including:

[0047] Step 1: Obtain basic spatial data of the parking lot. This data should include at least the parking space layout map, lane topology map, entrance and exit locations, fire lane locations, surveillance camera locations, environmental sensor locations, and access path information for inspection equipment. Based on this data, the entire parking area is divided into several monitoring zones. The specific process for dividing these zones includes:

[0048] The basic spatial data of the parking lot is mapped to a unified spatial coordinate system, and the boundaries of parking spaces, lanes, fire lanes, entrances and exits, and the access paths of inspection equipment are vectorized to form a parking lot spatial topology model. The main lanes, branch lanes, circular lanes, ramps, and fire lanes within the parking lot are used as primary area division boundaries to initially segment the parking spaces on both sides of the lanes. Parking spaces on both sides of the fire lanes are separately classified as independent area units. Parking lot entrances and exits, payment gates, turnstiles, elevator lobbies, stairwells, and charging area entrances are used as functional nodes. Area division lines are generated perpendicular to the lane extension direction at the corresponding functional node locations, further dividing the parking lot into multiple connected area units. The area units are then merged according to the arrangement of parking spaces. Specifically, consecutive parking spaces on both sides of the same lane segment are traversed and analyzed according to their parking space numbers. When adjacent parking spaces simultaneously meet the following conditions: located in the same lane segment, within the continuous coverage area of ​​the same set of surveillance cameras, within the effective acquisition range of the same set of environmental sensors, and accessible by inspection equipment along the same path, the parking space is considered merged. The system first merges the corresponding parking spaces into the same monitoring area. If the coverage set of the monitoring cameras or the coverage set of the environmental sensors for adjacent parking spaces changes, a boundary is generated between the corresponding parking spaces, dividing them into different monitoring areas. The boundary of the monitoring area is then corrected based on the accessible path of the inspection equipment. If the inspection equipment cannot continuously cover all parking spaces in a certain area without backing up, the area is split into multiple sub-monitoring areas. If there is a continuous accessible path between multiple areas, and the inspection equipment can enter from the same entrance point and complete the inspection sequentially along the same path, the corresponding areas are merged back into the same monitoring area. Each monitoring area after division is assigned a unique area number, and the set of parking spaces, boundary coordinates, video monitoring sensors, environmental sensors, inspection entrance point, and inspection equipment access path included in the corresponding monitoring area is recorded. Through the above division of monitoring areas, the parking lot monitoring area forms a clear boundary in terms of space, perception, and accessibility, providing a unified regional basis for subsequent environmental risk analysis, vehicle risk analysis, and inspection task cluster planning.

[0049] Step two involves continuously collecting and quantitatively analyzing the status of monitored areas and vehicle parking status within the parking lot through a preset monitoring cycle. This yields the environmental risk weights and vehicle risk factors for each monitored area. Combining the environmental risk weights corresponding to the monitored areas where vehicles are located, a risk gain analysis is performed on the vehicle risk factors to obtain the vehicle inspection demand index. Then, based on the inspection demand index, vehicle spatial distribution, and inspection path continuity, vehicles within the parking lot are prioritized and inspection tasks are clustered to form clusters of inspection tasks to be assigned. This enables dynamic identification and refined inspection task generation for high-risk vehicles, high-risk areas, and areas of abnormal activity within the parking lot. Specifically, this includes:

[0050] Step 201: Based on the sensors in each monitoring area, a quantitative analysis of the status of the monitoring area is performed to accurately assess the risk of the monitoring area and obtain the environmental risk weight of the monitoring area. Then, based on the distribution of parking spaces in the monitoring area, further analysis is performed to obtain the location risk weight of each parking space in the monitoring area, specifically including:

[0051] For monitoring area i, where i represents the index of the monitoring area within the parking lot; obstacles within monitoring area i are identified using video surveillance sensors, and the number of obstacles and the total area occupied by obstacles are counted; the total length and feasible length of the inspection passage within monitoring area i are calculated, where the feasible length refers to the length of the passage that ensures the normal passage of the inspection equipment at the current moment, and the total area occupied by obstacles refers to the sum of the areas obstructed by temporary parking objects, construction barriers, debris, vehicles occupying the road, etc.; then the passage congestion degree T of monitoring area i is expressed as:

[0052]

[0053] Among them, L free L is the passable length of the passage, and A is the total length of the inspection passage within the monitoring area. obs A represents the total area occupied by obstacles. path N represents the total area of ​​the inspection channels within the monitoring area. max To pre-determine the normalized upper limit of the number of obstacles, N obs This represents the number of obstacles within the monitoring area; η1, η2, and η3 are weighting coefficients, satisfying: η1 + η2 + η3 = 1. In this example, η1, η2, and η3 are taken as 0.4, 0.3, and 0.3, respectively. The first term in the above channel congestion calculation formula reflects the proportion of the channel being directly occupied, the second term reflects the degree of encroachment of obstacles on the path space, and the third term reflects the detour and congestion impact caused by the number of obstacles. The greater the channel congestion, the more difficult it is for the inspection equipment to pass smoothly in the area, and the higher the risk of inspection delay.

[0054] For monitoring area i, calculate the total area of ​​the entire monitoring area. At the same time, take the area that can be covered by video surveillance within the monitoring area as the effective monitoring area. Then, subtract the effective monitoring area from the total area of ​​the entire monitoring area and divide by the total area of ​​the entire monitoring area to obtain the monitoring blind zone degree of monitoring area i. The monitoring blind zone degree is used to characterize the size of the monitoring blind zone.

[0055] For monitoring area i, the number of people and vehicles passing through monitoring area i during the monitoring period is counted. Then the dynamic interference degree F is expressed as:

[0056]

[0057] Where ρ1 and ρ2 are weighting coefficients, and satisfy ρ1 + ρ2 = 1. In this embodiment, ρ1 and ρ2 are taken as 0.6 and 0.4 respectively, Δt represents the monitoring duration of this monitoring cycle, and N per To determine the number of people i in the monitored area, N veh This indicates the number of vehicles passing through the monitoring area i; the values ​​of ρ1 and ρ2 are determined by differentiating the weighting coefficients based on the varying degrees of impact of pedestrian and vehicle traffic on the inspection. In this embodiment, since pedestrian traffic is more random and has a greater impact on the passage of inspection equipment and vehicle safety, the value of ρ1 is increased; the dynamic interference degree is used to characterize the intensity of dynamic interference within the area caused by pedestrian and vehicle traffic.

[0058] In a preferred embodiment, the dynamic interference degree is calculated using a peak correction term to further enhance the dynamic performance. The specific calculation method is as follows:

[0059]

[0060] Wherein, ρ1, ρ2 and ρ3 are weighting coefficients, and satisfy ρ1+ρ2+ρ3=1. In this embodiment, ρ1, ρ2 and ρ3 are taken as 0.45, 0.35 and 0.20 respectively. σ represents the fluctuation range of pedestrian and vehicle flow in monitoring area i, that is, the fluctuation range of the number of people and vehicles, which is used to reflect short-term congestion and sudden gathering. The greater the fluctuation, the more unstable the inspection path and the surrounding environment of the vehicles are, and the higher the inspection risk.

[0061] For monitoring area i, smoke concentration, temperature, and humidity are acquired. Temperature and humidity deviations are calculated by comparing them to preset normal temperature and humidity thresholds, respectively. These deviations are then normalized to ensure they are on the same scale. Finally, they are linearly weighted and fused according to preset weights to obtain the environmental anomaly score, with the weights summed to one. The environmental anomaly score characterizes the impact of environmental anomalies such as smoke, temperature, and humidity on safety inspections. Smoke concentration is used to identify fires, smoke, or abnormal equipment heating; temperature deviation is used to identify localized temperature rise anomalies; and humidity deviation is used to identify problems such as dampness, condensation, or abnormal ventilation. In this embodiment, the weights for smoke concentration, temperature deviation, and humidity deviation are set to 0.5, 0.25, and 0.25, respectively, to prioritize the identification of smoke anomalies. A higher environmental anomaly score indicates a greater likelihood of potential safety hazards in the area, thus requiring a higher inspection priority.

[0062] The channel obstruction degree, monitoring blind spot degree, dynamic interference degree, and environmental anomaly degree are then normalized to obtain standardized parameters under a unified dimension. Based on the importance of each parameter, a weighted fusion is performed to obtain the environmental risk weight of the monitoring area. It should be noted that in this example, channel obstruction degree and monitoring blind spot degree are set to higher weights, as these directly affect the accessibility and risk identification capability of the inspection equipment; dynamic interference degree is set to a medium weight because it reflects the degree of dynamic interference in the area; and environmental anomaly degree is set to a higher weight because smoke, abnormal temperature and humidity are often more strongly associated with actual safety incidents. Specifically, the weights of channel obstruction degree, monitoring blind spot degree, dynamic interference degree, and environmental anomaly degree are set to 0.30, 0.25, 0.15, and 0.30, respectively.

[0063] Step 202: Quantitatively analyze the vehicle's stationary status to obtain the vehicle's own risk factors, specifically including:

[0064] For each vehicle within monitoring area i, its entry time, vehicle type, and target activity information near the vehicle are obtained. The vehicle type includes at least ordinary vehicles, new energy vehicles, and high-value vehicles. The target activity information near the vehicle includes the trajectory of people and vehicles within the preset monitoring range around the vehicle, the duration of target stay, the number of times the target approaches, and the target loitering behavior.

[0065] The dwell time is calculated by the time difference between the vehicle's entry time and the current time. The dwell time is then normalized to obtain a dwell time factor. Values ​​are assigned based on vehicle type; specifically, ordinary vehicles, new energy vehicles, and high-value vehicles correspond to different type values, with the type value for ordinary vehicles < that for new energy vehicles < that for high-value vehicles. In this embodiment, these are set to 0.3, 0.6, and 0.9, respectively. The higher the vehicle's value and the greater the need for protection, the larger its vehicle type factor value. A monitoring radius is set centered on the target vehicle to obtain its neighborhood range. The monitoring radius is preferably determined based on the target vehicle's dimensions, the distance between adjacent parking spaces, and the safety buffer distance. The major axis dimension of the target vehicle's circumscribed rectangle is then obtained. The monitoring radius is calculated using the formula. Where δ1 is the vehicle perimeter safety buffer distance and δ2 is the target recognition compensation distance, which are 0.5m and 0.3m respectively in this embodiment. If the parking lot has narrow spaces and a lot of people are moving around, the buffer distance can be appropriately increased. The vehicle perimeter safety buffer distance is used to cover the safety recognition range outside the actual boundary of the vehicle. The target recognition compensation distance is used to compensate for the vehicle recognition frame deviation, position calibration error and trajectory extraction error. Thus, a neighborhood range of the target vehicle is constructed within this monitoring radius, and each target trajectory entering the neighborhood range is identified within a preset monitoring period. The minimum approach distance, dwell time in the neighborhood, average approach speed towards the vehicle and number of repeated approaches are obtained for each trajectory. The minimum approach distance, dwell time in the neighborhood, average approach speed towards the vehicle and number of repeated approaches are normalized and linearly fused according to preset weights to obtain the abnormal activity factor of a single trajectory. The sum of the weights is one. The abnormal activity factors of all trajectories entering the neighborhood range within the monitoring period are summarized to obtain the abnormal activity factor near the target vehicle. The summation can preferably be determined by the mean, maximum value or a fusion of the mean and maximum value.

[0066] The vehicle risk factor is obtained by weighting and fusing the dwell time factor, type value, and abnormal activity factor.

[0067] Step 203: Based on the environmental risk weight of the monitored area where the vehicle is located, the vehicle risk factors are weighted to obtain the vehicle's inspection demand index, and an inspection task cluster is generated accordingly; specifically including:

[0068] The environmental risk weight corresponding to the monitoring area where the vehicle is located is obtained. Then, the environmental risk weight is used as a weighted gain coefficient, i.e., (1 + environmental risk weight) multiplied by the vehicle's vehicle risk factor to obtain the vehicle's inspection demand index. When the environmental risk of the monitoring area is higher, the vehicle's inspection demand index is amplified accordingly, thus giving vehicles in high-risk areas a higher inspection priority. After obtaining the inspection demand index of all vehicles, they are sorted from high to low according to the inspection demand index, and vehicles are divided into high-priority inspection objects, medium-priority inspection objects, and low-priority inspection objects according to a preset demand range. The demand range is set based on the historical inspection task distribution, the maximum carrying capacity of the inspection equipment, and the risk response timeliness requirements. In this embodiment, the demand range is set to [0.3, 0.6]. When the vehicle's inspection demand index is greater than the upper limit of the demand range, the vehicle is classified as a high-priority inspection object. When the vehicle's inspection demand index is within the demand range, the vehicle is classified as a medium-priority inspection object. When the vehicle's inspection demand index is less than the lower limit of the demand range, the vehicle is classified as a low-priority inspection object. Vehicles are classified as low-priority inspection objects. If multiple vehicles simultaneously meet the following conditions: located in the same or adjacent monitoring areas, with inspection demand indices in the same priority range, and the spatial distance between vehicles is less than a preset distance threshold, they are grouped into the same inspection task cluster. The distance threshold is determined based on the distance between adjacent parking spaces, the minimum safe gap for inspection equipment passage, and the continuous coverage requirement of the same inspection path. In this embodiment, it is set to 4m. If the standardized inspection demand index of a vehicle is greater than or equal to the high-risk threshold, in this embodiment, the high-risk threshold is set to 0.9, indicating that the vehicle has a strong need for immediate inspection. In this case, the vehicle is formed into an independent inspection task cluster. When the inspection demand of a vehicle is much higher than that of other vehicles in the same area, forming it into an independent cluster can prevent its high risk from being diluted by ordinary clustering tasks, thereby ensuring that high-risk vehicles receive priority, independent, and timely inspection processing. This forms a set of inspection task clusters to be assigned, and based on the adjacency of vehicle spatial locations, the continuity of inspection paths, and the task time window, vehicles that meet the clustering conditions are aggregated into inspection task clusters.

[0069] Step 3: For each inspection device, collect the current location of the inspection device and the currently assigned clusters of tasks to be inspected, and calculate the matching degree of the inspection device with each cluster of tasks to intelligently allocate each cluster of tasks to the respective inspection device; specifically including:

[0070] Arbitrarily select an inspection task cluster as the target task cluster, calculate the shortest path length d from the current location of the inspection equipment to the target task cluster based on the parking lot spatial topology model, and normalize it to obtain the distance matching degree P. d The normalization formula is:

[0071]

[0072] Where dmin and dmax represent the shortest and longest path lengths from all current inspection equipment to the target task cluster, respectively; ε is a very small positive number to prevent division by zero; distance matching degree is used to characterize the proximity of the inspection equipment to the target location of the task cluster, and the shorter the distance, the larger its value.

[0073] The system counts the number of task clusters (M) currently assigned to the inspection equipment, the cumulative duration (τ) of the current inspection tasks performed by the equipment, and the load matching degree (P). f The calculation formula is:

[0074]

[0075] Wherein, γ1 and γ2 are the weights of the number of task clusters to be inspected and the cumulative duration, respectively, and γ1+γ2=1, which are 0.6 and 0.4 respectively in this embodiment; Mmax and τmax are the normalized upper limits of the number of task clusters to be inspected and the cumulative duration, respectively. The load matching degree is used to characterize the current resource occupancy of the inspection equipment. The higher the load, the lower the matching degree.

[0076] Obtain historical inspection records of the inspection equipment. These records should include at least the historical task number, task start time, task end time, task execution status, communication data return records, scheduling change records, and equipment fault alarm records. Within each monitoring cycle, count the total number of communication data return packets and the number of packets that were not successfully received. Divide the number of packets that were not successfully received by the total number of communication data return packets to obtain the packet loss rate. The packet loss rate is used to characterize the communication stability of the inspection equipment during the inspection process. The lower the packet loss rate, the more stable the data return capability of the inspection equipment.

[0077] The inspection completion rate is calculated by counting the total number of inspection task clusters assigned to the inspection equipment and the number of inspection task clusters that were successfully completed. The number of successfully completed inspection task clusters is divided by the total number of inspection task clusters assigned to the inspection equipment. A successfully completed inspection task is one in which the inspection equipment completes all inspection points and data transmission according to the scheduling requirements without any abnormal interruption. The higher the inspection completion rate, the stronger the task performance capability of the inspection equipment.

[0078] By recording scheduling changes within each monitoring cycle, the number of times an inspection device was reassigned or exited a task midway through execution is counted, resulting in the inspection rescheduling count. Inspection rescheduling refers to the situation where an inspection device's current task is reassigned to another device due to reasons such as communication abnormalities, insufficient power, path congestion, or device malfunctions—that is, due to the device's own reasons. The higher the inspection rescheduling count, the worse the stability of the inspection device's execution. The rescheduling rate is obtained by dividing the inspection rescheduling count by the total number of inspection task clusters assigned to the inspection device.

[0079] By recording the equipment fault alarms in each monitoring cycle, the number of fault alarms generated by the inspection equipment in the monitoring cycle is counted to obtain the number of equipment fault alarms. The fault rate is obtained by dividing the number of fault alarms by the total number of inspection task clusters assigned to the inspection equipment.

[0080] The packet loss rate for each monitoring period Inspection completion rate Rescheduling rate and failure rate The execution coefficient of the inspection equipment is obtained by weighted fusion calculation. The calculation formula is as follows:

[0081]

[0082] Where α1, α2, α3, and α4 are weighting coefficients, which in this embodiment are set to 0.25, 0.40, 0.15, and 0.20 respectively. The inspection completion rate is given the highest weight to highlight the equipment's actual task fulfillment capability; packet loss rate and the number of equipment fault alarms are given medium weights to comprehensively reflect the equipment's communication stability and operational reliability; and the number of inspection reschedulings is given a lower weight to help reflect the stability of the equipment during execution. A larger execution coefficient indicates a stronger overall execution capability of the inspection equipment; therefore, the execution coefficient corresponding to each historical monitoring cycle is denoted as... , where j is the index of any monitoring period between the current monitoring periods, with the end time of each monitoring period as the monitoring time of that period, and then the execution coefficients of all historical monitoring periods are used. The operating history coefficient of the inspection equipment is obtained by applying time decay, which can more accurately reflect the actual performance of the inspection equipment at the current stage, rather than being overly influenced by its historical performance from a long time ago; the time decay calculation formula is:

[0083]

[0084] Where R jThe time interval between the monitoring time of detection cycle j and the current time is represented by λ, which is the time decay coefficient and λ > 0. The running history coefficient is used to characterize the continuous execution capability of the inspection equipment in the most recent historical cycle and serves as the basis for the allocation of subsequent inspection task clusters.

[0085] The distance matching degree, load matching degree, and operation history coefficient are linearly weighted and fused according to preset weights to obtain the final matching degree. The sum of the weights is one. In this embodiment, the weights of distance matching degree, load matching degree, and operation history coefficient are set to 0.35, 0.30, and 0.35, respectively. Distance directly affects arrival efficiency, operation history coefficient directly reflects the recent actual execution capability of the equipment, and load matching degree is mainly used to avoid equipment overload. It is a constraint indicator but should not be too heavy. The inspection equipment with the highest final matching degree and meeting the power constraint is assigned to the corresponding inspection task cluster. After a task cluster is assigned, the current load of the related equipment will change. It is necessary to update the load matching degree of these equipment and recalculate the final matching degree for the remaining task clusters. This can avoid the distortion of subsequent assignments after the high-priority tasks have been assigned in the past.

[0086] Step four: The inspection equipment determines the inspection order according to the vehicle inspection demand index in the assigned inspection task cluster from high to low, and selects the shortest accessible passage path between adjacent inspection objects. When there are closed areas, obstacles or temporary restricted areas in the path, the accessible passage path between adjacent inspection objects is re-determined based on the current parking lot traffic status, and an alternative passage is used to complete the subsequent inspection, so as to ensure that the inspection equipment can continuously cover all assigned tasks.

[0087] 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.

[0088] 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 scheduling vehicle safety inspection equipment based on smart parking lots, characterized in that, Includes the following steps: Step 1: Obtain basic spatial data of the parking lot, and perform spatial topology modeling of the parking lot based on the basic spatial data to divide it into several monitoring areas; Step 2: According to the preset monitoring cycle, continuously collect and quantitatively analyze the status of the monitoring area and the vehicle parking status in the parking lot to obtain the environmental risk weight and vehicle risk factor of each monitoring area. Based on the environmental risk weight corresponding to the monitoring area where the vehicle is located, conduct risk gain analysis on the vehicle risk factor to obtain the vehicle inspection demand index. Then, form an inspection task cluster based on the inspection demand index. Step 3: Collect the current location, current load, and historical inspection records of the inspection equipment. Based on this, analyze the final matching degree of the inspection equipment to the inspection task cluster, and assign the inspection equipment with the highest final matching degree and meeting the power constraints to the corresponding inspection task cluster; the current load refers to the inspection task cluster that the inspection equipment has been assigned to. Step 4: Determine the inspection sequence based on the inspection demand index of vehicles within the assigned inspection task clusters by the inspection equipment, and select the shortest accessible route between adjacent inspection objects.

2. The method for scheduling vehicle safety inspection equipment based on smart parking lots according to claim 1, characterized in that, In step one, dividing the parking lot into several monitoring zones includes: The basic spatial data of the parking lot is mapped to a unified spatial coordinate system, and the boundaries of parking spaces, lanes, fire lanes, entrances and exits, and the passage paths of inspection equipment are vectorized. The parking lot is initially divided into areas using the main lanes, branch lanes, circular lanes, ramps, and fire lanes as primary area division boundaries. Based on the coverage of surveillance cameras, the coverage of environmental sensors, and the continuous reachable paths of inspection equipment, adjacent parking spaces are merged or split into monitoring areas.

3. The method for scheduling vehicle safety inspection equipment based on smart parking lots according to claim 2, characterized in that, Step two, which involves quantitative analysis of the status of the monitored area, includes: Obstacles within the monitored area are identified using video surveillance sensors, and the number of obstacles and the total area occupied by obstacles are counted. The total length of the inspection channel, the passable length of the channel, and the total area of ​​the inspection channel are calculated to obtain the channel congestion degree. The total area of ​​the monitored area and the effective coverage area of ​​the monitoring are calculated to obtain the monitoring blind spot degree. The number of people and vehicles within the monitoring period are counted to obtain the dynamic interference degree. Smoke concentration, temperature deviation, and humidity deviation are obtained to obtain the environmental anomaly degree. The channel congestion degree, monitoring blind spot degree, dynamic interference degree, and environmental anomaly degree are normalized and weighted and fused to obtain the environmental risk weight of the monitored area.

4. The vehicle safety inspection equipment scheduling method based on smart parking lots according to claim 3, characterized in that, Step two, which involves quantitative analysis of the vehicle's stationary status, includes: The system acquires vehicle entry time, vehicle type, and information on target activity near the vehicle; it obtains a vehicle dwell time factor based on the time difference between the vehicle entry time and the current time; it obtains a vehicle type factor based on the vehicle type; it constructs a vehicle neighborhood centered on the target vehicle and identifies the trajectory of targets entering the neighborhood, obtaining the minimum approach distance, dwell time, average approach speed, and number of repeated approaches corresponding to the trajectory, and calculates the trajectory abnormal activity factor; it summarizes all trajectory abnormal activity factors within the monitoring period to obtain the vehicle vicinity abnormal activity factor; and it weights and fuses the dwell time factor, vehicle type factor, and vehicle vicinity abnormal activity factor to obtain the vehicle risk factor.

5. The vehicle safety inspection equipment scheduling method based on smart parking lots according to claim 4, characterized in that, The vehicle categories include at least ordinary vehicles, new energy vehicles, and high-value vehicles. The monitoring radius of the vehicle's neighborhood range is determined jointly based on the target vehicle's external dimensions, the distance between adjacent parking spaces, and the safety buffer distance. The safety buffer distance is used to cover the safe identification range outside the vehicle's true boundary, and the target identification compensation distance is used to compensate for vehicle identification box deviation, position calibration error, and trajectory extraction error.

6. The method for scheduling vehicle safety inspection equipment based on smart parking lots according to claim 5, characterized in that, The step two process of generating the inspection task cluster includes: Vehicles are classified into high-priority, medium-priority, and low-priority inspection targets based on the vehicle inspection demand index. When multiple vehicles are located in the same or adjacent monitoring areas, their inspection demand indices are in the same priority range, and the spatial distance between the vehicles is less than a preset distance threshold, they are grouped into the same inspection task cluster. When the vehicle inspection demand index is greater than or equal to a preset high-risk threshold, the corresponding vehicles are formed into independent inspection task clusters.

7. The method for scheduling vehicle safety inspection equipment based on smart parking lots according to claim 6, characterized in that, The calculation of the final matching degree in step three includes: The shortest path length from the current location of the inspection equipment to the target inspection task cluster is calculated based on the parking lot spatial topology model, and the distance matching degree is obtained by normalization. The number of inspection task clusters currently assigned to the inspection equipment and the cumulative duration of the current inspection tasks are counted, and the load matching degree is calculated. The historical inspection records of the inspection equipment are obtained, and the execution coefficient is obtained based on the packet loss rate, inspection completion rate, rescheduling rate and failure rate. Then, the running history coefficient is obtained after time decay. The distance matching degree, load matching degree and running history coefficient are weighted and fused to obtain the final matching degree.

8. The method for scheduling vehicle safety inspection equipment based on smart parking lots according to claim 7, characterized in that, The load matching degree is determined by the number of task clusters to be inspected and the cumulative duration, and the two are fused according to preset weights; the running history coefficient is obtained by time decay based on the execution coefficient corresponding to the historical monitoring cycle. Once a task cluster is assigned, the load matching degree of the relevant inspection equipment is updated, and the final matching degree of the remaining inspection task clusters is recalculated.

Images