Battery full charge alarm method and system under power-off charging state of airport automatic driving vehicle

Abstract

The invention discloses a battery full charge alarm method and system in a power-off charging state of an airport automatic driving vehicle, and relates to the technical field of vehicle battery management. The method comprises the following steps that: a battery management system performs full charge detection on a vehicle battery in a power-off charging state, and triggers a vehicle to execute a local alarm action and disconnects a charging loop after judging that the vehicle battery is fully charged; a sensing system arranged in the charging area collects camera and radar data of vehicles in the area in real time; the edge calculation unit carries out space-time alignment and fusion decision making on the sensing data, and judges whether a fully-charged vehicle triggering a local alarm exists or not; if yes, the alarm information is uploaded to a cloud alarm system to achieve remote early warning. Through cooperative work of the battery management system and the sensing system and in combination with edge calculation fusion decision, reliable detection and remote alarm of battery full charge in a power-off charging state are realized, the risk of overcharge is effectively avoided, and the operation safety and scheduling efficiency of airport vehicles are improved.

Description

Technical Field

[0001] This invention relates to the field of battery management technology, specifically to a method and system for providing a full charge alarm for an airport autonomous vehicle during a power outage charging state. Background Technology

[0002] With the development of smart airport construction, the application of autonomous vehicles within airports is becoming increasingly widespread, such as driverless shuttle buses and baggage tractors. These vehicles are typically driven entirely by electricity and require automatic charging in designated charging areas at the airport. Airport environments are characterized by high safety requirements, multi-vehicle collaborative operations, and 24 / 7 uninterrupted operation. During the charging process of autonomous vehicles, especially in unattended, power-off charging conditions, the vehicles are in a power-down sleep mode, with only the charging circuit remaining operational. Accurately determining whether the battery is truly fully charged and issuing timely alarm information is crucial. On one hand, timely alarms upon full charging can prevent safety hazards caused by overcharging, such as battery thermal runaway and lifespan degradation. On the other hand, full-charge information is a key basis for the airport dispatch system to reallocate vehicles, helping to improve the operational efficiency of airport ground services.

[0003] However, current full-charge alarms during power outage charging primarily rely on the battery management system's own detection and local vehicle alarm triggering. This information cannot be directly transmitted to remote monitoring systems, making it difficult for airport dispatch centers to promptly obtain vehicle full-charge status and conduct subsequent dispatching. Furthermore, relying solely on the battery management system for full-charge determination carries a risk of false alarms or missed alarms, especially under complex environmental conditions, making it difficult to guarantee reliable alarm information transmission. Therefore, how to achieve reliable detection and remote alarming of battery full charge during power outage charging has become a pressing technical problem to be solved in this field. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for alarming full battery charge in airport autonomous vehicles during power outage charging. This method and system enables reliable detection and remote alarming of full battery charge during power outage charging, effectively avoiding overcharging risks and improving airport vehicle operation safety and scheduling efficiency.

[0005] To achieve the above objectives, the embodiments of this invention provide the following technical solutions:

[0006] This application provides a method for a battery full-charge alarm in the power-off charging state of an airport autonomous vehicle, including the following steps: S1, the battery management system performs full-charge detection on the vehicle battery in the power-off charging state. When it is determined that the battery has reached full charge, the system triggers the vehicle to execute a local alarm action and disconnects the charging circuit; S2, a perception system set in the charging area collects perception data of vehicles in the area in real time and preprocesses the perception data. The perception system includes a camera and radar; S3, an edge computing unit receives the perception data, performs spatiotemporal alignment and fusion decision on the camera data and radar data, and determines whether there is a fully charged vehicle in the area that triggers a local alarm; S4, if it is determined that there is a fully charged alarm vehicle, the edge computing unit uploads the alarm information to the alarm system to achieve remote early warning.

[0007] Furthermore, the battery management system performs a full charge detection on the vehicle battery in a power-off charging state, specifically including the following steps: S11, after the vehicle is connected to the charging pile, the battery management system is woken up and performs a self-test, collecting the initial parameters of the battery. If the state of charge is <100%, charging is started and real-time monitoring is activated, wherein the initial parameters include voltage, state of charge, and temperature; S12, the battery management system continuously monitors the battery's voltage, current, temperature, and state of charge parameters and communicates with the charging pile in real time to adjust the charging power; S13, when the battery simultaneously meets the following conditions: state of charge ≥99%, voltage change rate <0.01V / min, charging current <0.01C, and battery temperature is normal, the battery management system determines that the battery has reached a full charge state, where C is the battery's rated capacity; S14, the battery management system disconnects the charging relay to stop charging, sends a full charge signal to the vehicle controller, and controls the battery to enter the equalization mode.

[0008] Furthermore, the triggering of the vehicle to perform a local alarm action specifically includes: after the vehicle controller receives the full charge signal sent by the battery management system, controlling the vehicle to turn on the hazard warning lights to flash and send a buzzer alarm sound, while automatically disconnecting the charging contactor.

[0009] Furthermore, the preprocessing of the sensing data includes coordinate transformation and time synchronization of the radar data. The coordinate transformation includes converting the polar coordinate data acquired by the radar into Cartesian coordinates, then obtaining the coordinates in the camera coordinate system through rotation and translation, and finally projecting them onto the image plane using camera intrinsic parameters. The projection relationship is expressed as follows: ,in Cartesian coordinates, Let be a rotation matrix. It is a translation vector. For the camera intrinsic parameter matrix, As a scale factor, The coordinates are the projected image coordinates; the time synchronization adopts a hardware trigger or software timestamp alignment algorithm to ensure that the radar data frame and the camera image frame correspond to the same moment.

[0010] Furthermore, the preprocessing of the perceived data also includes color space conversion, red region extraction, and adaptive threshold optimization of the camera data. The color space conversion includes converting the RGB image captured by the camera into an HSV image, and normalizing R, G, and B during the conversion process. The normalized values ​​range from [0,1]. The calculation formulas for chroma H, saturation S, and luminance V are as follows: If the calculated H < 0°, then H is corrected by adding 360°, and the final value range of H is [0°, 360°], while the value ranges of S and V are [0, 1]. The red region extraction includes setting a threshold based on the HSV color space, extracting the red pixel region in the image to generate a binary image, and since red is distributed in two discontinuous intervals in the chroma channel, the two intervals are extracted separately and then subjected to union processing. The extraction conditions are as follows: ,in The saturation threshold, The brightness threshold is used; the adaptive threshold optimization includes dividing the region of interest (ROI) of the HSV image into blocks and dynamically adjusting the threshold using a global thresholding method or a local adaptive thresholding method. and If a global thresholding method is used, the annular interval of the H channel is first linearized before calculating the global threshold, while also considering the periodicity of H. If a local adaptive thresholding method is used, the threshold is dynamically adjusted based on the pixel mean and standard deviation within each block. and To adapt to scenes with dynamically changing lighting.

[0011] Furthermore, the fusion decision includes: calculating the probability of the vehicle's hazard lights flashing by analyzing the timing of the vehicle lights in the camera images. ; Calculate the probability that the vehicle is in an abnormal state by using vehicle speed data collected by radar. And when the vehicle speed is less than 1 km / h and continues for a preset duration, Take the highest value; calculate the alarm confidence level using a fusion formula. The fusion formula is expressed as: ,in For visual probability weights, Let be the radar motion probability weight, and ;like If the value exceeds the preset alarm threshold, the vehicle is determined to be a fully charged vehicle.

[0012] Furthermore, the visual probability weight and radar motion probability weight are dynamically adjusted according to the environment of the airport charging station area. Specifically, when in daylight or well-lit environments, the proportion of visual probability weight is increased; when in nighttime, rain, fog, or other poor lighting or severe weather environments, the proportion of radar motion probability weight is increased.

[0013] Furthermore, the perception system employs a redundant perception mode consisting of cameras and radar. Under harsh environmental conditions, the data collected by the cameras and radar are mutually verified to ensure the accuracy of the perception data.

[0014] Furthermore, the alarm information includes the vehicle's location coordinates, vehicle information, full charging time, and charging pile number; the cloud alarm system receives the alarm signal, displays and stores it, and supports the push and query of alarm information.

[0015] Accordingly, this application also provides a battery full charge alarm system for airport autonomous vehicles in a power outage charging state, comprising: a battery management system, installed on the vehicle, for detecting the battery full charge status and sending a full charge signal after determining that it is full charge; a vehicle controller, connected to the battery management system, for receiving the full charge signal and controlling the vehicle to execute local alarm actions; a sensor perception system, deployed in the charging area, including cameras and radar, for collecting perception data of vehicles in the area; an edge computing unit, connected to the sensor perception system, for performing spatiotemporal alignment and fusion decision-making on the perception data to determine whether there is a vehicle with a full charge alarm in the area; and a cloud alarm system, communicatively connected to the edge computing unit, for receiving alarm information and realizing remote early warning.

[0016] The beneficial effects of this invention are as follows: By constructing a full-process alarm system that integrates a local detection and perception system, an edge computing unit, and a cloud alarm system within the battery management system, dual protection of local alarms and remote early warnings for fully charged batteries is achieved. This allows the charging area to promptly perceive the fully charged status and enables the airport dispatch and operations departments to monitor alarm information in real time. Standardized preprocessing and fusion decision-making of the perceived data improves the accuracy of identifying fully charged vehicles, avoiding false positives and false negatives. Real-time remote transmission of fully charged alarm information provides the airport dispatch system with accurate vehicle status data, enabling timely allocation and scheduling of fully charged vehicles, reducing idle time after charging, and optimizing the turnaround efficiency of airport ground autonomous vehicles. Furthermore, the closed-loop design from detection to alarm avoids the safety hazards of battery overcharging at the source, improving the overall safety of the airport autonomous vehicle charging process. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a method for providing a full charge alarm for an airport autonomous vehicle during a power outage and charging state, as provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a battery full charge alarm system for an airport autonomous vehicle in a power outage charging state, provided as an embodiment of this application. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0020] In this invention, the terms "system" and "network" are used interchangeably. "Multiple" refers to two or more; therefore, in this invention, "multiple" can also be understood as "at least two." "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / ", unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, it should be understood that in the description of this invention, terms such as "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or order.

[0021] Example 1:

[0022] Currently, when airport autonomous vehicles are charging after a power outage, they rely solely on the battery management system to complete basic full-charge detection, which can only trigger simple local alarms without a remote warning mechanism. Full-charge information cannot be transmitted to the airport dispatch and operations departments in a timely manner. At the same time, the vehicle status perception in the charging area lacks standardized processing and decision-making procedures, making it impossible to accurately identify vehicles with full-charge alarms. This results in fully charged vehicles being idle and charging safety hazards not being monitored remotely in a timely manner, leading to low overall safety and operational efficiency in charging management.

[0023] like Figure 1As shown in the figure, this application embodiment provides a battery full charge alarm method for an airport autonomous vehicle in a power-off charging state, including the following steps: S1, the battery management system performs full charge detection on the vehicle battery in a power-off charging state. When it is determined that the battery has reached a full charge state, the system triggers the vehicle to execute a local alarm action and disconnects the charging circuit; S2, a perception system set in the charging area collects perception data of vehicles in the area in real time and preprocesses the perception data. The perception system includes a camera and radar; S3, an edge computing unit receives the perception data, performs spatiotemporal alignment and fusion decision on the camera data and radar data, and determines whether there is a fully charged vehicle in the area that triggers a local alarm; S4, if it is determined that there is a fully charged alarm vehicle, the edge computing unit uploads the alarm information to the alarm system to realize remote early warning.

[0024] In another possible embodiment, the battery management system installed on the airport's autonomous vehicles first performs a comprehensive full-charge status detection on the vehicle's battery, which is in a power-off charging state. When the battery management system determines, based on multiple parameters, that the battery has reached full charge, it immediately triggers the vehicle to execute the corresponding local alarm action and directly disconnects the vehicle's charging circuit to prevent overcharging from the source. Subsequently, the sensor perception system deployed in the designated area of ​​the airport charging stations continuously collects various perception data from all vehicles in that area in real time. After collection, the data is immediately preprocessed to eliminate various interference factors generated during the data collection process, ensuring the validity of the perception data. Then, it communicates with the sensor perception system. The edge computing unit connected to the communication network receives pre-processed perception data and performs spatiotemporal alignment processing on the visual data collected by cameras and the motion perception data collected by radar to completely eliminate the deviation between the two types of data in the time and space dimensions. Then, based on the aligned perception data, it performs fusion decision analysis to accurately determine whether there are fully charged vehicles in the charging pile area that have triggered local alarms. If the edge computing unit clearly determines through fusion decision that there are such fully charged vehicles in the area, it will immediately upload the alarm information containing key vehicle-related information to the cloud alarm system through the network. The cloud alarm system will then realize remote early warning of the fully charged alarm, allowing the airport dispatch and operation terminals to obtain the vehicle fully charged alarm information in the charging pile area in real time.

[0025] By constructing a comprehensive alarm system that integrates a local detection and perception system for the battery management system, an edge computing unit, and a cloud alarm system, dual protection is achieved: local alarms and remote early warnings for fully charged batteries. This ensures that the charging area can promptly perceive the fully charged status, and that the airport dispatch and operations departments can monitor alarm information in real time. Standardized preprocessing and fusion decision-making of the perceived data improves the accuracy of identifying vehicles with fully charged alarms, avoiding false positives and false negatives. Real-time remote transmission of fully charged alarm information provides the airport dispatch system with accurate vehicle status data, enabling the system to promptly allocate and dispatch fully charged vehicles, reducing idle time after charging and optimizing the turnaround efficiency of autonomous vehicles on the airport ground. Furthermore, the closed-loop design from detection to alarm avoids the safety hazards of battery overcharging at the source, improving the overall safety of the charging process for autonomous vehicles at the airport.

[0026] If the battery management system relies solely on a single state of charge parameter to determine the battery's full charge status, it is prone to misjudgment or missed detection due to factors such as battery voltage fluctuations and abnormal current. Furthermore, simply stopping charging after detecting full charge without a robust battery protection and individual cell voltage balancing mechanism not only affects the battery's lifespan but also easily leads to safety hazards in subsequent use due to inconsistent individual cell voltages.

[0027] In the embodiments of this application, the battery management system performs a full charge detection on the vehicle battery in a power-off charging state, specifically including the following steps: S11, after the vehicle is connected to the charging pile, the battery management system is woken up and performs a self-test, collecting the initial parameters of the battery. If the state of charge is <100%, charging is started and real-time monitoring is activated, wherein the initial parameters include voltage, state of charge, and temperature; S12, the battery management system continuously monitors the battery's voltage, current, temperature, and state of charge parameters and communicates with the charging pile in real time to adjust the charging power; S13, when the battery simultaneously meets the following conditions: state of charge ≥99%, voltage change rate <0.01V / min, charging current <0.01C, and battery temperature is normal, the battery management system determines that the battery has reached a full charge state, where C is the battery's rated capacity; S14, the battery management system disconnects the charging relay to stop charging, sends a full charge signal to the vehicle controller, and controls the battery to enter the equalization mode.

[0028] In another possible embodiment, when an autonomous vehicle connects to an airport charging station, the vehicle's battery management system (BMS) is automatically activated and immediately performs a comprehensive self-test. After the self-test, it collects the battery's initial operating parameters, including total battery voltage, individual cell voltage, and state of charge (SOC). If the collected SOC does not meet the full charge standard, the BMS immediately sends a charging start signal to the charging station to initiate the charging process. Simultaneously, it activates the BMS's real-time monitoring mode. Throughout the charging process, the BMS continuously monitors various battery operating parameters, including real-time total battery voltage, individual cell voltage, charging current, battery temperature, and dynamically changing SOC. The system maintains real-time communication with the charging pile and dynamically sends power adjustment signals to the charging pile based on the monitored real-time battery status. The charging pile then adjusts the charging power accordingly. When the battery management system detects that the battery simultaneously meets four conditions—that the state of charge has reached the full charge range, the battery voltage change rate is within a stable threshold, the charging current is close to zero, and the battery temperature is within the normal operating range—it comprehensively determines that the battery has reached the full charge state. Immediately afterward, it controls the disconnection of the charging relay to physically cut off the charging circuit, stop the battery charging, and send a full charge signal to the vehicle controller. Then, it controls the battery pack to enter the equalization mode to perform voltage balance adjustment on all individual cells in the battery pack to ensure the consistency of the voltage of each individual cell.

[0029] The system comprehensively assesses multiple parameters, including state of charge, rate of change of voltage, charging current, and battery temperature, to detect the battery's full charge status. Compared to relying on a single parameter, this significantly improves the accuracy of full charge detection and effectively avoids misjudgments and missed detections caused by fluctuations in a single parameter. Upon detecting a full charge, the charging circuit is physically disconnected by disconnecting the charging relay, mitigating the risk of overcharging from a hardware perspective. Simultaneously, a battery balancing mode is activated to balance and adjust the voltage of individual cells within the battery pack, ensuring voltage consistency and improving overall battery performance while extending battery life. Furthermore, real-time communication between the battery management system and the charging station allows for dynamic adjustment of charging power based on battery status, ensuring a safe and efficient charging process.

[0030] The local alarm methods for vehicles after the battery is fully charged are relatively simple, mostly just simple light prompts, which have limited warning effect. Furthermore, there is no automatic disconnection mechanism for the charging circuit. Even if the alarm is triggered, there is still a safety hazard of the battery continuing to charge and causing overcharging, which cannot achieve dual protection for charging safety.

[0031] In the embodiments of this application, triggering the vehicle to perform a local alarm action specifically includes: after the vehicle controller receives the full charge signal sent by the battery management system, controlling the vehicle to turn on the hazard warning lights to flash and send a buzzer alarm sound, while automatically disconnecting the charging contactor.

[0032] In another possible embodiment, after receiving the battery full charge signal from the battery management system, the vehicle controller immediately performs three operations simultaneously. The first is to control the vehicle's hazard warning lights to flash continuously at a preset fixed frequency, creating a clear visual warning in the charging area. The second is to trigger the buzzer on the vehicle to continuously emit a low-frequency alarm sound, enhancing the on-site warning effect through auditory means, achieving dual local alarms of light and sound. The third is to control the charging contactor inside the vehicle to automatically disconnect, achieving a secondary physical cutoff of the charging circuit. This forms a dual hardware cutoff with the charging relay disconnected by the battery management system, completely eliminating the possibility of the battery continuing to charge. All of the above actions are started simultaneously without any sequential delay, and the local alarm actions and charging cutoff state will only be released after the staff manually resets the vehicle.

[0033] By designing a dual local alarm system combining light and sound, the visual and auditory warnings are significantly enhanced, improving the on-site warning effect in the charging area. This allows airport staff to promptly detect vehicles with full-charge alarms. Simultaneously, the charging contactor is automatically disconnected upon triggering the alarm, achieving a secondary disconnection of the charging circuit. This, along with the disconnection of the charging relay by the battery management system, forms a dual hardware disconnection guarantee, fundamentally eliminating the possibility of battery overcharging and further improving the safety of the charging process. Moreover, all local alarm actions and charging disconnection actions are executed synchronously by the vehicle controller, ensuring the timeliness and consistency of the actions without any execution delay.

[0034] In a sensor perception system, radar and cameras are different types of sensing devices. Radar collects data in polar coordinates, while camera collects data in image coordinates. The two coordinate systems are not consistent, and there is a slight deviation in the time of data collection. The existence of spatiotemporal deviation makes it impossible to directly fuse and analyze the two types of data, which can easily cause deviations in subsequent vehicle status determination and reduce the accuracy of alarm vehicle identification.

[0035] In the embodiments of this application, the preprocessing of the sensing data includes coordinate transformation and time synchronization of the radar data. The coordinate transformation includes converting the polar coordinate data acquired by the radar into Cartesian coordinates, then obtaining the coordinates in the camera coordinate system through rotation and translation, and finally projecting them onto the image plane in combination with the camera intrinsic parameters. The projection relationship is expressed as follows: ,in Cartesian coordinates, Let be a rotation matrix. It is a translation vector. For the camera intrinsic parameter matrix, As a scale factor, The coordinates are the projected image coordinates; the time synchronization adopts a hardware trigger or software timestamp alignment algorithm to ensure that the radar data frame and the camera image frame correspond to the same moment.

[0036] In another possible embodiment, when preprocessing the perceived data, the polar coordinate data acquired by the radar is first transformed into Cartesian coordinate data. Then, using a preset rotation matrix and translation vector, the Cartesian coordinate data is transformed into spatial coordinate data in the camera coordinate system. Finally, combined with the camera intrinsic parameter matrix, the spatial coordinate data in the camera coordinate system is projected onto the image plane of the camera, converting it into image coordinate data consistent with the data acquired by the camera, thus unifying the spatial coordinates of the radar and camera data. Subsequently, time synchronization processing of the radar and camera data is performed. If a hardware triggering method is used, a unified hardware trigger signal controls the radar and camera to start data acquisition at the same time to achieve time synchronization. If a software algorithm method is used, a timestamp alignment algorithm is used to time-calibrate the acquired data of the radar and camera, and interpolation correction is performed on data with time deviations to ensure that the radar data frame and the camera image frame are the acquisition results at the same time, thus completing the time synchronization of the two types of data.

[0037] Through a coordinate transformation process, a unified conversion from radar polar coordinate data to camera image coordinate data is achieved, enabling the analysis of two types of perception data in different coordinate systems within the same spatial dimension. This eliminates the analysis errors caused by spatial deviation. Simultaneously, time synchronization between radar and camera data is achieved through hardware triggering or software algorithms, ensuring that both types of data are collected at the same time, thus eliminating the analysis errors caused by time deviation. This dual elimination of spatiotemporal deviation provides a foundation for the fusion analysis of radar and camera perception data, significantly improving the accuracy of subsequent fusion decisions and providing reliable data support for the accurate identification of vehicles with full charge warnings.

[0038] The visual data collected by the camera is in RGB color space format. The three channels of this color space are highly correlated and are easily affected by factors such as changes in lighting and ambient reflection. Directly extracting the red area of ​​the vehicle's hazard warning light based on the RGB color space has extremely low accuracy. Moreover, existing technologies mostly use fixed thresholds for color area extraction, which cannot adapt to the dynamic changes in lighting in the charging pile area. This can easily lead to missed or false red area extraction, affecting the subsequent recognition of the alarmed vehicle.

[0039] In embodiments of this application, the preprocessing of the perceived data further includes color space conversion, red region extraction, and adaptive threshold optimization of the camera data. The color space conversion includes converting the RGB image captured by the camera into an HSV image, and normalizing R, G, and B during the conversion process. The normalized values ​​range from [0,1]. The calculation formulas for chroma H, saturation S, and luminance V are as follows: If the calculated H < 0°, then H is corrected by adding 360°, and the final value range of H is [0°, 360°], while the value ranges of S and V are [0, 1]. The red region extraction includes setting a threshold based on the HSV color space, extracting the red pixel region in the image to generate a binary image, and since red is distributed in two discontinuous intervals in the chroma channel, the two intervals are extracted separately and then subjected to union processing. The extraction conditions are as follows: ,in The saturation threshold, The brightness threshold is used; the adaptive threshold optimization includes dividing the region of interest (ROI) of the HSV image into blocks and dynamically adjusting the threshold using a global thresholding method or a local adaptive thresholding method. and If a global thresholding method is used, the annular interval of the H channel is first linearized before calculating the global threshold, while also considering the periodicity of H. If a local adaptive thresholding method is used, the threshold is dynamically adjusted based on the pixel mean and standard deviation within each block. and To adapt to scenes with dynamically changing lighting.

[0040] In another possible embodiment, the original RGB color space image captured by the camera is first converted to an HSV color space image. The HSV color space separates the image's chroma, saturation, and brightness features. During the conversion process, the pixel values ​​of the red, green, and blue channels of the RGB image are first normalized, and then the chroma, saturation, and brightness values ​​of each pixel are calculated separately to complete the color space conversion. Subsequently, based on the features of the HSV color space, the red pixel region corresponding to the vehicle's hazard warning light is extracted from the image. Since red is distributed in two discontinuous intervals in the chroma channel, corresponding saturation and brightness thresholds are set for each interval, and each... The red pixel regions within each interval are used to generate corresponding binary images. Then, the union of the two binary images is performed to obtain the complete binary image of the red region. Finally, the binary image of the red region is subjected to threshold optimization. On the one hand, the annular interval of the chroma channel is linearized and the global threshold is calculated using the OTSU algorithm to adapt to a wide range of illumination changes. On the other hand, the region of interest (ROI) in the image, specifically the vehicle detection area within the charging pile area, is divided into blocks. Based on the statistical information such as the pixel mean and standard deviation in each block, the saturation and brightness thresholds are dynamically adjusted to adapt to local uneven illumination scenes, thus completing the preprocessing of the camera visual data.

[0041] By converting the RGB color space to the HSV color space, which separates chroma, saturation, and brightness, the impact of lighting changes and ambient reflections on visual data is effectively reduced, making the extraction of the red area more stable and accurate. Addressing the characteristic that red is distributed across two discontinuous intervals in the HSV color space, a method of interval extraction followed by union is adopted to ensure complete extraction of the red area of ​​the hazard warning light without omission. Simultaneously, an optimization method combining global thresholding and local adaptive thresholding is employed, which can adapt to both large-scale lighting changes in the charging pile area and localized uneven lighting scenarios, significantly improving the accuracy and robustness of red area extraction and providing high-quality visual data support for subsequent vehicle warning light status recognition.

[0042] If the determination of vehicle status in the charging station area relies heavily on data from a single sensor, such as visual data or radar motion data, it is easily affected by environmental factors and may lead to misjudgments. Furthermore, without combining the vehicle's motion status and lighting status for comprehensive analysis, it is impossible to effectively distinguish between vehicles with full charging alarms and other stationary vehicles in the charging station area, resulting in low accuracy in identifying alarm vehicles.

[0043] In embodiments of this application, the fusion decision includes: calculating the probability of hazard lights flashing on a vehicle by analyzing the timing of vehicle lights in camera images. ; Calculate the probability that the vehicle is in an abnormal state by using vehicle speed data collected by radar. And when the vehicle speed is less than 1 km / h and continues for a preset duration, Take the highest value; calculate the alarm confidence level using a fusion formula. The fusion formula is expressed as: ,in For visual probability weights, Let be the radar motion probability weight, and ;like If the value exceeds the preset alarm threshold, the vehicle is determined to be a fully charged vehicle.

[0044] In another possible embodiment, after completing the spatiotemporal alignment of radar and camera perception data, the edge computing unit performs fusion decision analysis. First, it performs vehicle light timing analysis on the camera's visual data, continuously identifying the on / off state and switching rate of the vehicle's hazard warning lights to determine whether the vehicle is in a hazard warning state. Based on the accuracy of the identification, it calculates the visual probability of the vehicle's hazard warning lights flashing. Then, it analyzes the radar's motion perception data, extracting motion parameters such as the vehicle's speed and acceleration detected by the radar. Based on the vehicle's motion parameters, it determines whether the vehicle is in a stationary charging state, and calculates the radar motion probability of the vehicle in an abnormal state based on the accuracy of the judgment. Next, it substitutes the calculated visual probability and radar motion probability into a preset fusion formula, and obtains the vehicle's comprehensive alarm confidence score through weighted calculation. Finally, it compares the comprehensive alarm confidence score with a preset alarm threshold. If the comprehensive alarm confidence score exceeds the preset alarm threshold, it accurately determines that the vehicle is a vehicle with a full battery charge alarm.

[0045] By fusing visual data from cameras and motion data from radar to determine vehicle status, and combining the vehicle's hazard lights (double flashing) status with its motion status for comprehensive analysis, the system avoids misjudgments caused by environmental interference with data from a single sensor. Visual probabilities are calculated through headlight timing analysis to accurately identify the hazard lights' flashing status, and radar motion probabilities are calculated from radar speed data to accurately determine the vehicle's stationary charging status. Finally, a comprehensive alarm confidence score is calculated using a fusion formula to quantitatively determine whether a vehicle is in a full-charge alarm state, making the judgment more objective and accurate. This effectively distinguishes fully-charge alarm vehicles from other stationary vehicles within the charging station area, significantly reducing the misjudgment rate of alarm vehicle identification.

[0046] In the embodiments of this application, the visual probability weight and radar motion probability weight are dynamically adjusted according to the environment of the airport charging pile area. Specifically, this includes: increasing the proportion of visual probability weight when in daylight or well-lit environment; and increasing the proportion of radar motion probability weight when in night, rain, fog, or other poor lighting or severe weather conditions.

[0047] In another possible embodiment, before performing weighted calculations for fusion decision-making, the edge computing unit first determines the current environmental scene of the airport charging pile area. If it is determined that the current environment is daytime, without snow, rain, fog, or good lighting, the visual data collected by the camera is less affected by environmental interference and has high accuracy. In this case, the weight of visual probability is randomly increased, while the weight of radar motion probability is decreased, allowing the camera's visual data to play a dominant role in the fusion decision-making. If it is determined that the current environment is nighttime, with poor lighting or severe weather such as rain, snow, fog, or sandstorms, the motion data collected by the radar is less affected by environmental interference and has high accuracy. In this case, the weight of visual probability is decreased, while the weight of radar motion probability is increased, allowing the radar's motion data to play a dominant role in the fusion decision-making. The sum of the two weights is always 1. After completing the dynamic adjustment of the weights, the adjusted weights are then substituted into the fusion formula to calculate the comprehensive alarm confidence.

[0048] By dynamically adjusting the weight ratio of visual probability and radar motion probability according to the actual environmental scene of the airport charging pile area, the fusion decision can adapt to different environmental conditions. In well-lit scenes, the weight of visual probability is increased to give full play to the high accuracy of camera visual data. In poor lighting or bad weather scenes, the weight of radar motion probability is increased to give full play to the advantage of radar motion data being less affected by environmental interference. This allows the fusion decision to maintain high accuracy in various environments and effectively avoid misjudgments caused by environmental changes.

[0049] In the embodiments of this application, the perception system uses a redundant perception mode consisting of a camera and radar. Under harsh environmental conditions, the data collected by the camera and radar are mutually verified to ensure the accuracy of the perception data.

[0050] In another possible embodiment, the sensor perception system consists of both a camera and radar, achieving redundant perception in the airport charging station area. Under normal environmental conditions, the camera and radar simultaneously collect vehicle perception data. After preprocessing and spatiotemporal alignment, the two types of data are fused for decision-making to achieve accurate vehicle status identification. In adverse environmental conditions, if the camera fails to collect visual data or its accuracy decreases due to factors such as insufficient lighting, rain, snow, or fog, and it cannot effectively extract the red warning light area of ​​the vehicle, the radar can still collect the vehicle's motion parameters normally. By leveraging the high weighting of the radar's motion probability, accurate vehicle status determination is achieved. If the radar fails to collect motion data or its accuracy decreases due to factors such as environmental obstruction or electromagnetic interference, the camera can still collect the vehicle's visual data normally. By leveraging the high weighting of the visual probability, accurate vehicle status determination is achieved. If one type of sensor experiences a hardware failure, the other type of sensor can independently complete the perception and determination of the vehicle status, ensuring the continuity and effectiveness of the entire perception system and achieving complementary verification and redundancy protection of perception data.

[0051] By employing a redundant sensing mode that combines cameras and radar to form a sensor perception system, the two types of sensors are sensing devices based on different principles, each possessing different sensing advantages and being affected by different environmental interference points. This enables complementary verification of sensing data. When one type of sensor fails to collect data or its accuracy decreases due to malfunction or environmental interference, the other type of sensor can continue to provide effective sensing data, ensuring the continuity and effectiveness of vehicle status perception. Even in extremely harsh environments, it can still accurately identify vehicles with full charge warnings, meeting the high security and high reliability requirements of airports operating 24 hours a day. At the same time, the redundant sensing design can also mutually verify the data collected by the two types of sensors, further improving the accuracy of the sensing data.

[0052] In the embodiments of this application, the alarm information includes the vehicle's location coordinates, vehicle indication, full charging time, and charging pile number; the cloud alarm system receives the alarm signal, displays and stores it, and supports the push and query of alarm information.

[0053] In another possible embodiment, after the edge computing unit identifies a vehicle with a full battery charge alarm, it immediately organizes the relevant key information of the vehicle to form a complete alarm information package. The alarm information package contains core information such as the vehicle's specific location coordinates in the charging pile area, the vehicle's unique identifier, the specific time the battery is fully charged, and the charging pile number to which the vehicle is connected. Subsequently, the alarm information package is uploaded to the cloud alarm system via the network. After receiving the alarm information package, the cloud alarm system first displays the alarm information in real time on the system's visual operation interface, associating the vehicle's location, identifier, and other information with the electronic map of the airport charging pile area, allowing staff to intuitively grasp the alarm status. At the same time, the alarm information package is encrypted and securely stored in the system's database, forming traceable historical data. Then, according to the airport's preset push rules, the alarm information is accurately pushed to the designated airport dispatch terminal and operation management terminal, allowing relevant staff to obtain the alarm information in a timely manner. In addition, the cloud alarm system also supports staff to search and query the stored historical alarm information through multiple dimensions such as vehicle identifier, charging pile number, and alarm time, and also supports the statistics and export of historical data, providing data support for airport operation management.

[0054] By designing a complete alarm information system that includes core vehicle characteristics and charging information, the airport dispatch center can accurately grasp the specific location, vehicle identification, charging completion time, and charging pile number of vehicles with full-charge alarms. This provides comprehensive data support for the dispatch system's precise scheduling and route planning, making dispatch operations more targeted. Simultaneously, the cloud alarm system enables visualized display, secure storage, precise push notifications, and multi-dimensional historical query functions for alarm information. Visualized display allows staff to intuitively grasp the alarm status of the charging pile area; precise push notifications ensure timely transmission of alarm information to designated dispatch and operation terminals; and secure storage and historical query functions allow the airport to statistically analyze and trace historical data on vehicle charging and full-charge alarms, providing data support for the operation and management of autonomous vehicles at the airport and improving the informatization level of airport charging management.

[0055] Example 2:

[0056] Reference Figure 2 This application also provides a battery full charge alarm system for airport autonomous vehicles in a power outage charging state, comprising: a battery management system, installed on the vehicle, for detecting the battery full charge status and sending a full charge signal after determining that it is full charge; a vehicle controller, connected to the battery management system, for receiving the full charge signal and controlling the vehicle to execute local alarm actions; a sensor perception system, deployed in the charging area, including cameras and radar, for collecting perception data of vehicles in the area; an edge computing unit, connected to the sensor perception system, for performing spatiotemporal alignment and fusion decision-making on the perception data to determine whether there is a vehicle with a full charge alarm in the area; and a cloud alarm system, communicatively connected to the edge computing unit, for receiving alarm information and realizing remote early warning.

[0057] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above embodiments. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention.

[0058] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not describe the various possible combinations separately.

[0059] Furthermore, various different implementations of the present invention can be combined arbitrarily, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed in the present invention.

Claims

1. A method for providing a full charge alarm for an airport autonomous vehicle during a power outage charging state, characterized in that, Includes the following steps: S1. The battery management system performs a full charge detection on the vehicle battery in a power-off charging state. When it is determined that the battery has reached a full charge state, it triggers the vehicle to execute a local alarm action and disconnects the charging circuit. S2. A sensing system located in the charging area collects real-time sensing data of vehicles in the area and preprocesses the sensing data. The sensing system includes a camera and radar. S3. The edge computing unit receives the perception data, performs spatiotemporal alignment and fusion decision-making on camera data and radar data, and determines whether there are fully charged vehicles in the area that trigger local alarms. S4. If a vehicle is determined to be under full charge alarm, the edge computing unit will upload the alarm information to the alarm system to achieve remote early warning.

2. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 1, characterized in that, The battery management system performs a full charge test on the vehicle battery when it is in a power-off charging state, specifically including the following steps: S11. After the vehicle is connected to the charging pile, the battery management system is woken up and performs a self-test, collecting the initial parameters of the battery. If the state of charge is <100%, charging is started and real-time monitoring is activated. The initial parameters include voltage, state of charge, and temperature. S12. The battery management system continuously monitors the battery's voltage, current, temperature, and state of charge parameters and communicates with the charging pile in real time to adjust the charging power. S13. When the battery simultaneously meets the following conditions: state of charge ≥ 99%, voltage change rate < 0.01V / min, charging current < 0.01C, and battery temperature is normal, the battery management system determines that the battery has reached full charge state, where C is the rated capacity of the battery. S14. The battery management system disconnects the charging relay to stop charging, sends a full charge signal to the vehicle controller, and controls the battery to enter the equalization mode.

3. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 2, characterized in that, The triggering of the vehicle to execute a local alarm action specifically includes: after the vehicle controller receives the full charge signal sent by the battery management system, it controls the vehicle to turn on the hazard warning lights to flash and send a buzzer alarm sound, while automatically disconnecting the charging contactor.

4. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 1, characterized in that, The preprocessing of the sensing data includes coordinate transformation and time synchronization of the radar data. The coordinate transformation involves converting the polar coordinate data acquired by the radar into Cartesian coordinates, then obtaining the coordinates in the camera coordinate system through rotation and translation, and finally projecting the data onto the image plane using camera intrinsic parameters. The projection relationship is expressed as follows: ,in Cartesian coordinates, Let be a rotation matrix. It is a translation vector. For the camera intrinsic parameter matrix, As a scale factor, The coordinates of the projected image; The time synchronization employs a hardware-triggered or software timestamp alignment algorithm to ensure that radar data frames and camera image frames correspond to the same moment.

5. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 1, characterized in that, The preprocessing of the perceived data also includes color space conversion, red region extraction, and adaptive threshold optimization of the camera data. The color space conversion includes converting the RGB image captured by the camera into an HSV image, and normalizing R, G, and B during the conversion process. The normalized values ​​range from [0,1]. The calculation formulas for chroma H, saturation S, and luminance V are as follows: If the calculated H < 0°, then H is corrected by adding 360°, and the final value range of H is [0°, 360°], and the value range of S and V is [0, 1]. The red region extraction involves setting a threshold based on the HSV color space, extracting red pixel regions from the image to generate a binary image, and since red is distributed in two discontinuous intervals in the chroma channel, the two intervals are extracted separately and then subjected to union processing. The extraction conditions are as follows: ,in The saturation threshold, This is the brightness threshold. The adaptive threshold optimization involves dividing the region of interest (ROI) of the HSV image into blocks and dynamically adjusting the threshold using either a global thresholding method or a local adaptive thresholding method. and If a global thresholding method is used, the annular interval of the H channel is first linearized before calculating the global threshold, while also considering the periodicity of H. If a local adaptive thresholding method is used, the threshold is dynamically adjusted based on the pixel mean and standard deviation within each block. and To adapt to scenes with dynamically changing lighting.

6. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 1, characterized in that, The fusion decision includes: By analyzing the timing of vehicle lights in camera images, the probability of hazard lights flashing is calculated. ; The probability of a vehicle being in an abnormal state is calculated using vehicle speed data collected by radar. And when the vehicle speed is less than 1 km / h and continues for a preset duration, Take the highest value; Alarm confidence is calculated using a fusion formula. The fusion formula is expressed as: ,in For visual probability weights, Let be the radar motion probability weight, and ; like If the value exceeds the preset alarm threshold, the vehicle is determined to be a fully charged vehicle.

7. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 6, characterized in that, The visual probability weight and radar motion probability weight are dynamically adjusted according to the environment of the airport charging station area, specifically including: Increase the proportion of visual probability weight when in daylight or well-lit environments; When in poor lighting conditions such as nighttime, rain, fog, or severe weather, increase the weighting of radar motion probability.

8. The method for providing a full charge alarm for an airport autonomous vehicle during a power outage and charging state, as described in claim 1, is characterized in that... The perception system employs a redundant perception mode consisting of cameras and radar. Under harsh environmental conditions, the data collected by the cameras and radar are mutually verified to ensure the accuracy of the perception data.

9. The battery full charge alarm method for airport autonomous vehicles in a power outage charging state according to claim 1, characterized in that, The alarm information includes the vehicle's location coordinates, vehicle information, full charge time, and charging pile number; the cloud alarm system receives the alarm signal, displays and stores it, and supports the push and query of alarm information.

10. A battery full charge alarm system for an airport autonomous vehicle in a power outage charging state, applied to the battery full charge alarm method for an airport autonomous vehicle in a power outage charging state as described in any one of claims 1-9, characterized in that, include: The battery management system, located on the vehicle side, is used to detect the battery's full charge status and send a full charge signal after determining that it is fully charged. The vehicle controller, connected to the battery management system, is used to receive a full charge signal and control the vehicle to perform local alarm actions. A sensor perception system, including cameras and radar, is deployed in the charging area to collect perception data of vehicles in the area. An edge computing unit, connected to the sensor perception system, is used to perform spatiotemporal alignment and fusion decision-making on the perceived data to determine whether there are vehicles with full charging alarms in the area. The cloud alarm system is communicatively connected to the edge computing unit and is used to receive alarm information and realize remote early warning.

Images