Vehicle anti-theft method, device and system, electronic equipment and storage medium

Abstract

The invention provides a vehicle anti-theft method, device and system, electronic equipment and a storage medium, and the method is applied to a cloud end, and comprises the steps: receiving real-time riding data reported by a vehicle, and obtaining a user habit model associated with the vehicle; performing anti-theft detection based on the real-time riding data and the user habit model to obtain a detection result; and under the condition that the detection result indicates that the vehicle state is abnormal, alarm information is sent to an operation and maintenance terminal, so that the operation and maintenance terminal performs anti-theft intervention based on the alarm information, the defects that a traditional vehicle anti-theft means is high in false alarm rate, lack of individuation and lagged in response are overcome, an individualized behavior mode of a user is introduced into anti-theft logic, and the anti-theft effect is improved. According to the invention, the conversion from passive response to active early warning is realized, the accuracy and the real-time performance of anti-theft alarm are greatly improved, the false alarm rate and the missing alarm rate are reduced, and more active and more intelligent guarantee is provided for vehicle safety.

Description

Technical Field

[0001] This invention relates to the fields of intelligent transportation and the Internet of Things, and in particular to a vehicle anti-theft method, device, system, electronic device, and storage medium. Background Technology

[0002] With the increasing popularity of green travel, electric vehicles and other two-wheeled transportation have become an important part of people's daily commutes and lives. However, electric vehicles currently face a serious risk of theft, making the protection of vehicle assets a challenge for the industry.

[0003] Currently, common vehicle anti-theft methods include GPS (Global Positioning System) location tracking, vibration sensor alarms, and electronic fence technology. However, these methods suffer from high false alarm rates. For example, vibration alarms are difficult to distinguish between malicious damage and normal use, which can easily cause harassment. They also lack personalization, as electronic fence technology uses a fixed geographical range and cannot adapt to changes in travel plans. Furthermore, they have slow response times, as GPS tracking is often a reactive measure, giving thieves ample time to move the vehicle or damage the positioning device, resulting in high difficulty and low success rate in vehicle recovery. Summary of the Invention

[0004] This invention provides a vehicle anti-theft method, device, system, electronic device, and storage medium to solve the problem that existing anti-theft methods are significantly insufficient in terms of accuracy and real-time performance due to their inability to personalize user behavior patterns. It achieves accurate identification and rapid response to vehicle theft and optimizes vehicle anti-theft performance.

[0005] This invention provides a vehicle anti-theft method applied in the cloud, the method comprising: Receive real-time riding data reported by the vehicle and obtain the user habit model associated with the vehicle; Anti-theft detection is performed based on the real-time riding data and the user habit model to obtain the detection results; If the detection result indicates that the vehicle is in an abnormal state, an alarm message is sent to the maintenance terminal so that the maintenance terminal can take anti-theft intervention based on the alarm message.

[0006] According to a vehicle anti-theft method provided by the present invention, the user habit model is determined based on the following steps: Obtain the user's historical riding data associated with the vehicle; Based on the user's historical cycling data, multi-dimensional feature extraction is performed to obtain the user's historical cycling features; The user's historical cycling characteristics are modeled using a time series analysis model to obtain the user habit model.

[0007] According to a vehicle anti-theft method provided by the present invention, the step of modeling the user's historical riding characteristics using a time series analysis model to obtain the user habit model includes: A user cycling profile is generated based on the user's historical cycling features; the user's historical cycling features include cycling route features, time pattern features, speed and acceleration pattern features, and parking position stability features. The user cycling profile is modeled using a time series analysis model to obtain the user habit model.

[0008] According to a vehicle anti-theft method provided by the present invention, the method further includes, after performing anti-theft detection based on the real-time riding data and the user habit model to obtain the detection result, the method further includes: If the detection result indicates that the vehicle is in normal condition, the user habit model is updated with parameters through incremental learning based on the real-time riding data. The user cycling profile is revised based on the updated user habit model.

[0009] According to a vehicle anti-theft method provided by the present invention, when the detection result indicates that the vehicle status is abnormal, an alarm message is sent to the operation and maintenance terminal so that the operation and maintenance terminal can perform anti-theft intervention based on the alarm message, including: If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a first preset level, an alarm message is sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is the second preset level, the vehicle is controlled to activate an audible and visual alarm and send an alarm message to the maintenance terminal so that the maintenance terminal can generate a remote intervention command based on the alarm message and perform anti-theft intervention based on the remote intervention command. The anomaly level is determined based on the deviation between the real-time cycling data and the user's cycling habits as represented by the user habit model.

[0010] According to a vehicle anti-theft method provided by the present invention, the step of sending alarm information to the maintenance terminal includes: Based on the real-time riding data, trajectory extraction is performed to obtain continuous trajectory data when the vehicle is in an abnormal state and the abnormality level is the second preset level. Structured abnormal trajectory records are generated based on the continuous trajectory data, and the abnormal trajectory records are stored. Based on the abnormal trajectory record and the detection result, an alarm message is generated and sent to the operation and maintenance terminal.

[0011] The present invention also provides a vehicle anti-theft device applied in the cloud, the device comprising: The acquisition unit is used to receive real-time riding data reported by the vehicle and acquire the user habit model associated with the vehicle. The detection unit is used to perform anti-theft detection based on the real-time riding data and the user habit model, and obtain the detection result; An anti-theft unit is used to send an alarm message to the operation and maintenance terminal when the detection result indicates that the vehicle is in an abnormal state, so that the operation and maintenance terminal can take anti-theft intervention based on the alarm message.

[0012] The present invention also provides a vehicle anti-theft system, including a vehicle, a cloud, a user terminal, and an operation and maintenance terminal; The cloud platform is used to receive real-time riding data collected and reported by the vehicle, and to perform anti-theft detection based on the real-time riding data and user habit model to obtain detection results. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a first preset level, an alarm message is sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a second preset level, the cloud platform controls the vehicle to activate an audible and visual alarm and sends an alarm message to the maintenance terminal, so that the maintenance terminal can generate a remote intervention command based on the alarm message and perform anti-theft intervention based on the remote intervention command.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the vehicle anti-theft method as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vehicle anti-theft method as described above.

[0015] The vehicle anti-theft method, device, system, electronic device, and storage medium provided by this invention use a user habit model associated with the vehicle as a personalized benchmark for anti-theft detection and compare it with the real-time riding data reported by the vehicle. This enables accurate identification of abnormal behaviors that do not conform to the user's normal riding habits. Once an anomaly is detected, an alarm message is immediately sent to the maintenance terminal for timely anti-theft intervention. This overcomes the shortcomings of traditional vehicle anti-theft methods, such as high false alarm rate, lack of personalization, and delayed response. By introducing the user's personalized behavior pattern into the anti-theft logic, it realizes the transformation from passive response to proactive warning, greatly improving the accuracy and real-time performance of anti-theft alarms, reducing false alarm rate and missed alarm rate, and providing more proactive and intelligent protection for vehicle security. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the vehicle anti-theft method provided by the present invention; Figure 2 This is a flowchart illustrating the construction process of the user habit model provided by this invention; Figure 3 This is a flowchart illustrating the hierarchical alarm process provided by the present invention; Figure 4 This is a schematic diagram of the vehicle anti-theft device provided by the present invention; Figure 5 This is a schematic diagram of the vehicle anti-theft system provided by the present invention; Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] In recent years, with the popularization of green urban travel, two-wheeled transportation, represented by electric vehicles, has become an important part of the public's daily commuting and life. However, with the surge in their ownership, both civilian and shared electric vehicles generally face a serious risk of theft, and how to effectively protect vehicle assets has become an urgent problem to be solved by the industry.

[0020] Currently, common vehicle anti-theft methods include using GPS positioning technology to track and locate vehicles. Once a vehicle is stolen, its location can be viewed through a backend system, providing clues for recovery. Another common method is vibration alarms, which involve installing vibration sensors on the vehicle. Once abnormal shaking is detected, an alarm is triggered to deter thieves. Additionally, there are solutions using electronic fence technology. This involves pre-setting a virtual geographical security zone; when the vehicle's real-time location exceeds this zone, it is considered abnormal and an alarm is triggered.

[0021] However, the aforementioned anti-theft methods still have many shortcomings. Specifically, current anti-theft methods have a high false alarm rate. For example, vibration alarm systems cannot distinguish between malicious damage and normal use. A user's compliant vehicle movement, the vehicle's movement during transport, or even unintentional contact with the external environment can all trigger unnecessary alarms, causing annoyance and reducing the credibility of alarm information. Furthermore, existing anti-theft methods lack personalization. For instance, electronic fence technology typically uses fixed geographical areas, failing to adapt to users' temporary and reasonable changes in travel plans, and making it difficult to effectively distinguish between normal user behavior and theft. More importantly, the aforementioned anti-theft methods generally suffer from response delays. For example, GPS tracking is mostly a reactive measure. By the time a user discovers their vehicle is missing and initiates tracking, thieves often have ample time to move the vehicle or disable / damage the tracking device, making vehicle recovery difficult and with a low success rate.

[0022] In response, this invention provides a vehicle anti-theft method, which aims to establish a personalized user habit model that reflects the user's riding habits, and analyze and detect the real-time status of the vehicle based on this model, so as to make decisions in the event of abnormal vehicle status to prevent vehicle theft. This solves the problems of high false alarm rate, lack of personalization and slow response of traditional vehicle anti-theft methods, and achieves accurate identification and rapid response to vehicle theft, thereby greatly improving the anti-theft performance of vehicles.

[0023] Figure 1 This is a flowchart illustrating the vehicle anti-theft method provided by the present invention, as shown below. Figure 1 As shown, this method can be applied in the cloud, and the specific execution entity can be a cloud server or server cluster, which has data storage, data processing, and computational analysis capabilities. The method includes: Step 110: Receive real-time riding data reported by the vehicle and obtain the user habit model associated with the vehicle. Step 120: Perform anti-theft detection based on real-time cycling data and user habit models to obtain detection results; Step 130: If the detection result indicates that the vehicle status is abnormal, send an alarm message to the operation and maintenance terminal so that the operation and maintenance terminal can take anti-theft intervention based on the alarm message.

[0024] Specifically, in the vehicle anti-theft process, the cloud platform first needs to acquire the user's real-time riding data when performing anti-theft detection tasks. That is, in practical applications, after the vehicle starts moving, its data acquisition modules (such as GPS modules, speed sensors, and accelerometers) continuously collect real-time riding data and send it to the cloud via wireless communication modules (such as 4G, 5G, and NB-IoT (Narrow Band Internet of Things)). Here, the vehicle can be any mobile vehicle requiring anti-theft monitoring, such as civilian electric vehicles, shared electric bicycles, and logistics delivery vehicles.

[0025] The real-time riding data here reflects the vehicle's current status and can include vehicle position, instantaneous speed, acceleration, steering angle, vehicle start / stop status, and time (such as current time and riding time). Specifically, after collecting the real-time riding data, the vehicle can transmit this data to the cloud via its wireless communication module at a preset frequency or when the vehicle's status changes significantly. Therefore, the cloud's reception of data reported by the vehicle is a continuous and dynamic process.

[0026] After receiving real-time riding data reported by the vehicles, the cloud can obtain the user habit model associated with that vehicle or its linked user. This user habit model represents the user's riding habits, which are the normalized and regular riding behaviors formed by the user over a long period of riding. It should be understood that in practical applications, each vehicle requiring anti-theft detection, or its associated user, corresponds to a user habit model to ensure the targeting and accuracy of theft detection. Here, the vehicle, its associated user, and the user habit model can be associated through identifiers such as vehicle identification numbers (VINs) and user accounts. Furthermore, this association can be stored in a cloud database or storage.

[0027] Specifically, after receiving real-time riding data reported by a vehicle, the cloud can read or retrieve a pre-generated and stored user habit model from its internal database or storage based on the vehicle's identifier. This user habit model can be constructed based on the user's historical riding data.

[0028] Next, anti-theft detection can be performed based on the received real-time riding data and the acquired user habit model to obtain the detection results. That is, the cloud compares and analyzes the real-time riding data with the acquired user habit model to determine whether the current riding behavior deviates from the user's normal habits.

[0029] It is understood that, in this embodiment of the invention, the anti-theft detection process is essentially a process of deviation calculation and evaluation. The user habit model defines the range or pattern of normal riding behavior, while anti-theft detection calculates the probability or distance that the current riding behavior, as represented by real-time riding data, falls outside this normal range. For example, the user habit model may define that a user "often rides on route A between 8-9 am on weekdays." If real-time riding data shows that the vehicle appears in area B at "3 am," the cloud can calculate the significant deviation between the current behavior and the model in both the time and spatial dimensions by comparison. This can determine that the vehicle's status is abnormal, there is a possibility of theft, and anti-theft intervention is necessary.

[0030] The detection algorithm here can be diverse, such as statistical probability model judgment, distance-based metric calculation, or simple rule engine judgment. It will ultimately output a detection result, which can be a Boolean value (normal or abnormal), an enumerated status label, or a quantified anomaly score. The higher the score, the greater the deviation of the vehicle's state from normal behavior, and the higher the risk of theft.

[0031] Afterwards, anti-theft intervention can be implemented based on the detection results. Specifically, when the detection results from the previous step indicate an abnormal vehicle status, it means the cloud has determined that the vehicle is at risk of being stolen. At this point, the cloud will trigger an alarm process, sending alarm information to the operations and maintenance (O&M) team so that the relevant O&M personnel can quickly and accurately understand the vehicle's abnormal situation and make intervention decisions. Here, the alarm information may include the vehicle identifier, alarm time, vehicle location, alarm type or reason (e.g., unusual activity during off-peak hours, significant route deviation, etc.), and the associated user terminal. The cloud can send alarm information to the O&M team via API calls, message queue push, real-time push, etc., to ensure low-latency delivery of information.

[0032] Upon receiving an alarm, the corresponding maintenance personnel can immediately view alarm details, vehicle location, and other information. Based on this information, they can decide and execute a series of anti-theft intervention measures. For example, they can send remote intervention commands to the cloud, which then forwards them to the vehicle to perform operations such as remote locking, remote speed limiting, and remote power cut-off to prevent theft. Alternatively, intervention measures can also be offline manual intervention, such as dispatching nearby security personnel based on the alarm's trajectory to handle the situation and recover the vehicle.

[0033] The vehicle anti-theft method provided by this invention uses a user habit model associated with the vehicle as a personalized benchmark for anti-theft detection and compares it with the real-time riding data reported by the vehicle. It can accurately identify abnormal behaviors that do not conform to the user's normal riding habits. Once an anomaly is detected, an alarm message is immediately sent to the operation and maintenance terminal for timely anti-theft intervention. This overcomes the shortcomings of traditional vehicle anti-theft methods, such as high false alarm rate, lack of personalization, and delayed response. It introduces the user's personalized behavior pattern into the anti-theft logic, realizing the transformation from passive response to proactive warning. This greatly improves the accuracy and real-time performance of anti-theft alarms, reduces false alarm rate and missed alarm rate, and provides a more proactive and intelligent guarantee for vehicle security.

[0034] Based on the above embodiments, the user habit model is determined based on the following steps: Obtain user's historical riding data associated with the vehicle; Based on the user's historical cycling data, multi-dimensional feature extraction is performed to obtain the user's historical cycling features; A time series analysis model was used to model the historical cycling characteristics of users, resulting in a user habit model.

[0035] Specifically, a user habit model can be constructed through the following steps: Figure 2 This is a flowchart of the user habit model construction process provided by the present invention, such as... Figure 2 As shown, in order to generate a user habit model that accurately reflects a user's cycling habits, this embodiment of the invention requires first obtaining the user's or vehicle's cycling data over a past period, i.e., the user's historical cycling data. User historical cycling data is a collection of cycling data associated with a vehicle or its linked user, continuously received and stored by the cloud during long-term operation. For example, the cloud can store all cycling data reported by the vehicle over the past month, three months, or even longer, such as vehicle location, instantaneous speed, acceleration, steering angle, vehicle start / stop status, and time, in a database, forming the user's historical cycling data associated with that vehicle.

[0036] Then, multi-dimensional feature extraction can be performed based on the user's historical cycling data to obtain the user's historical cycling features. That is, structured information that can effectively represent the user's behavioral patterns is extracted from the user's historical cycling data, namely, the user's historical cycling features.

[0037] Multidimensional feature extraction refers to the analysis, calculation, and summarization of users' historical cycling data from different angles and dimensions to quantify users' behavioral patterns. These dimensions can include spatial, temporal, and vehicle dynamic dimensions. For example, the cloud can perform cluster analysis on historical vehicle locations to identify users' frequently visited locations (such as home and office) and frequently taken routes; it can also perform statistical analysis on historical usage time to identify users' frequently used travel time periods (such as morning and evening rush hours on weekdays); and it can also analyze historical instantaneous speed and acceleration to summarize users' typical driving styles (such as whether they prefer smooth driving or rapid acceleration / deceleration); this embodiment of the invention does not specifically limit these aspects.

[0038] Then, a time series analysis model can be used to model the user's historical cycling characteristics to obtain a user habit model. That is, using machine learning, deep learning, and other technologies, the extracted historical cycling characteristics are transformed into a user habit model capable of predicting and judging future behavior. Specifically, since user cycling behavior has obvious time dependencies—for example, the behavior patterns on Monday mornings and Saturday nights are significantly different—a time series analysis model is chosen for modeling in this embodiment. The time series analysis model here is a type of algorithm specifically designed for processing and analyzing time series data (i.e., data points arranged in chronological order). This type of model can learn the dependencies, periodic patterns, and long-term trends of data over time. The time series analysis model can be a traditional statistical model or a machine learning model, such as deep learning-based recurrent neural networks, long short-term memory networks, gated recurrent units, etc. These deep learning models are particularly adept at capturing long-term dependencies and can accurately learn complex user cycling patterns.

[0039] The modeling process involves using extracted historical cycling features of users as training data, which are then input into a time series analysis model for training. During training, the model continuously adjusts its internal parameters to best fit and predict user behavior based on the input features. Once training is complete, a model representing the corresponding user's cycling habits is obtained; this model is the user habit model associated with the vehicle or the user associated with the vehicle.

[0040] In this embodiment of the invention, a highly personalized anti-theft benchmark model that can capture changes in user behavior can be automatically learned and constructed from massive and messy historical data. This makes subsequent anti-theft detection no longer rely on rigid and universal rules, but is based on a deep understanding of each user's unique habits, thereby greatly improving the accuracy of anti-theft detection.

[0041] Based on the above embodiments, a time series analysis model is used to model the user's historical cycling characteristics to obtain a user habit model, including: A user cycling profile is generated based on the user's historical cycling characteristics. The user's historical cycling characteristics include cycling route characteristics, time pattern characteristics, speed and acceleration pattern characteristics, and parking position stability characteristics. A time series analysis model was used to model the user's cycling profile, resulting in a user habit model.

[0042] Specifically, the above-mentioned modeling process using time series analysis can include: First, a user cycling profile can be generated based on the characteristics of the user's historical cycling routes, time patterns, speed and acceleration patterns, and parking location stability. This user cycling profile can be understood as a multi-dimensional and comprehensive snapshot of user behavior patterns formed by structuring and tagging the user's historical cycling characteristics. It can systematically and clearly depict the user's cycling habits.

[0043] Cycling route features are used to depict users' spatial and geographical activity patterns. The cloud can perform cluster analysis on vehicle locations in users' historical cycling data to identify frequently occurring starting points, destinations, and common routes or activity hotspots connecting these points. The extracted cycling route features can be represented as a set of geofences, a series of key waypoints, a weighted route map, etc.

[0044] Temporal pattern characteristics are used to characterize users' bike-sharing habits over time. The cloud can perform statistical analysis on users' historical riding data to identify their frequent travel periods, stationary periods, etc.

[0045] Speed ​​and acceleration pattern features are used to quantify a user's dynamic riding style. By analyzing instantaneous speed and acceleration in a user's historical riding data, the cloud can calculate statistical indicators such as average riding speed, speed variance, acceleration / deceleration distribution range, and the frequency of rapid acceleration / deceleration events. For example, a user's normal riding pattern might be a smooth start and gentle acceleration / deceleration. This feature can effectively distinguish between a user's normal riding and abnormal dynamics such as "aggressive handling" or "severe bumps" in non-riding states.

[0046] Parking location stability features are used to assess the safety and reliability of users' frequently used parking locations. By analyzing the start-stop status of vehicles in users' historical riding data, the cloud can identify users' frequently used parking spots and the stability of each parking spot, such as parking frequency, average parking duration, and main parking times.

[0047] The cloud platform integrates the features from the four dimensions mentioned above to generate a structured user cycling profile. For example, a user cycling profile could be: "User A is accustomed to traveling from home (location A) to the company (location B) along Route 1 at 8 am on weekdays, with a smooth driving style; the vehicle is usually parked stably at location A at night."

[0048] Once a user's cycling profile is obtained, the cloud can utilize a time series analysis model for deep learning to capture the complex relationships between various features over time, thereby constructing a user habit model. Specifically, the user's cycling profile can be used as input to train the time series analysis model. Through learning, the model can understand and remember the inherent logic and temporal dependencies between the features reflected in the user's cycling profile. For example, it can learn that "under the 'time feature' of a weekday morning, there is a 95% probability that a vehicle will appear on 'route feature 1,' and its 'speed and acceleration pattern features' should conform to a smooth driving pattern." After sufficient training, the final user habit model can be obtained.

[0049] In this embodiment of the invention, a user cycling profile is constructed through four key features, and time series modeling is performed based on this profile. This greatly enriches the information dimensions of the anti-theft detection basis, making the final generated user habit model more comprehensive and three-dimensional.

[0050] Based on the above embodiments, anti-theft detection is performed based on real-time cycling data and user habit models to obtain detection results. The process then includes: If the detection results indicate that the vehicle is in normal condition, the parameters of the user habit model are updated through incremental learning based on real-time riding data. Based on the updated user habit model, the user cycling profile is revised.

[0051] Specifically, to ensure the accuracy of anti-theft detection, this embodiment of the invention also requires that the user habit model be able to evolve dynamically to adapt to the user's riding habits, avoiding false alarms caused by changes in the user's riding habits. That is, see [link to relevant documentation]. Figure 2 It is evident that the user habit model can be optimized and updated to ensure the accuracy and precision of anti-theft detection.

[0052] In detail, when the detection result indicates that the vehicle's condition is normal, the user habit model can be updated with parameters through incremental learning based on this real-time riding data. That is, after the cloud-based anti-theft detection task is completed, a detection result is obtained. When this result indicates that the vehicle's condition is normal, it means that the riding behavior represented by the current real-time riding data is the user's own normal behavior. At this point, the model update process can be initiated. That is, based on the real-time riding data, the parameters of the existing user habit model are updated through incremental learning.

[0053] Incremental learning, a machine learning method, allows a model to continuously learn and optimize using newly arriving data without complete retraining. Compared to periodically retraining using all historical data, incremental learning is more efficient and consumes fewer computational resources, making it particularly suitable for vehicle theft prevention scenarios that require continuous adaptation to changes in data flow.

[0054] After this, the user's cycling profile can be revised based on the updated user habit model. In other words, after updating the parameters of the user habit model, this updated model already contains the user's latest behavioral patterns. At this point, the cloud can use this updated user habit model to update the user profile in reverse.

[0055] In detail, this could involve using the updated user habit model to update the labeled and structured feature information in the user's cycling profile. For example, if the incremental learning process allows the model to accept a new frequently used parking spot, then the parking location stability feature dimension of the user's cycling profile needs to include a record of this new frequently used parking spot and update its related confidence and other statistical information. Similarly, if a user develops a new commuting route, the cycling route feature in the user's cycling profile should also be updated accordingly.

[0056] In this embodiment of the invention, an adaptive optimization mechanism when the vehicle is in a normal state enables continuous learning and evolution of the user habit model. This makes the anti-theft detection basis no longer a static, fixed set of rules, but a model with dynamic adaptability, thereby greatly improving the accuracy of anti-theft detection and reducing the false alarm rate.

[0057] Based on the above embodiments, step 130 includes: If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is the first preset level, an alarm message will be sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is the second preset level, the vehicle will be controlled to activate the audible and visual alarm and send alarm information to the operation and maintenance terminal so that the operation and maintenance terminal can generate remote intervention instructions based on the alarm information and perform anti-theft intervention based on the remote intervention instructions. The anomaly level is determined based on the deviation between real-time cycling data and user cycling habits represented by the user habit model.

[0058] Specifically, the aforementioned anti-theft intervention process based on detection results includes: In order to ensure theft prevention and security while taking into account user experience and avoiding excessive disturbance to users, a hierarchical alarm mechanism is proposed in this embodiment of the invention. Figure 3This is a flowchart illustrating the hierarchical alarm process provided by the present invention, as follows: Figure 3 As shown, the mechanism adopts different levels of alarm and intervention strategies based on the severity of the anomaly.

[0059] In detail, when performing anti-theft detection, the cloud platform, in addition to outputting normal or abnormal detection results, also needs to calculate the severity level of the abnormal situation, i.e., the abnormality level, when the detection result is abnormal, i.e., the vehicle's condition is abnormal. This abnormality level can be obtained based on the deviation between real-time riding data and the user's riding habits. When the deviation is large, the abnormality level is high, and vice versa. When the deviation is small, the abnormality level is low.

[0060] For different levels of anomalies, different alarm and intervention procedures will be executed in this embodiment of the invention. Specifically, when the detection result is abnormal and the anomaly level is the first preset level (i.e., a low anomaly level and a slight abnormality in the vehicle's condition), the preferred strategy to avoid unnecessary audible and visual alarms and maintenance intervention causing inconvenience to the user is to confirm with the user. That is, the cloud will send an alarm message to the user's device linked to the vehicle. This alarm message can be a push notification, in-app message, or SMS, such as "Your vehicle moved near [time] and [location]. Please confirm whether it was you operating it?" After receiving the alarm message, the user can confirm or deny it on their device. For example, if the user confirms it was them operating it, the vehicle is in normal condition and no further anti-theft intervention is needed; if the user does not confirm or reports that it was not them operating it, the risk can be escalated, and anti-theft intervention can continue, such as sending an alarm message to the maintenance department.

[0061] Correspondingly, when the detection result is abnormal, and the abnormality level is the second preset level (i.e., a high abnormality level and a severely abnormal vehicle condition), the vehicle is highly likely to be stolen, and therefore, strong anti-theft intervention measures must be taken immediately. Specifically, the cloud will control the vehicle to activate audible and visual alarms, such as emitting a high-decibel buzzer and high-frequency flashing warning lights. Simultaneously, an alarm message will be sent to the operations and maintenance (O&M) system. Unlike minor abnormalities, this alarm message will be sent directly to the O&M system. The message will indicate that this is a "serious alarm" and will include all necessary related data, such as vehicle location, trajectory, and time, to facilitate a rapid response from O&M personnel.

[0062] Upon seeing the alarm information, maintenance personnel will immediately assess the situation. Based on the vehicle location and trajectory data provided by the alarm, they can issue remote intervention commands, such as remotely locking the vehicle, remotely limiting speed, or remotely cutting off power. Specifically, this could involve maintenance personnel clicking a corresponding button on their terminal; the terminal would then capture the user's action, generate the appropriate remote intervention command, and send it to the vehicle via the cloud for execution, effectively preventing theft. Simultaneously, maintenance personnel can also use the trajectory data to deploy offline security forces to track and recover the vehicle.

[0063] In this embodiment of the invention, a tiered alarm mechanism enables targeted responses to abnormal situations. For minor anomalies with low risk, confirmation is achieved by sending a reminder to the user, avoiding unnecessary disturbance. For serious anomalies with extremely high risk, local vehicle audible and visual alarms are immediately activated, along with remote intervention from the maintenance department, forming a two-layer protection system combining user confirmation and maintenance intervention. While ensuring user experience, this significantly improves the response speed and handling capabilities for high-risk theft situations, achieving a balance between security and user experience.

[0064] Based on the above embodiments, alarm information is sent to the operation and maintenance terminal, including: Trajectory extraction is performed based on real-time cycling data to obtain continuous trajectory data under the condition that the vehicle status is abnormal and the abnormality level is the second preset level. Structured abnormal trajectory records are generated based on continuous trajectory data, and these abnormal trajectory records are stored. Based on the abnormal trajectory records and detection results, alarm information is generated and sent to the operation and maintenance terminal.

[0065] Specifically, the process of sending alarm information to the operations and maintenance department when the anomaly level is the second preset level includes: In the event of a serious anomaly, to facilitate decision-making and intervention by operations and maintenance personnel, this embodiment of the invention can generate continuous trajectory data under abnormal conditions and send it to the operations and maintenance end to assist operations and maintenance personnel in making judgments and decisions.

[0066] In detail, this could involve first extracting the trajectory based on real-time cycling data, then extracting and integrating a segment of the trajectory directly related to the current severe anomaly from the discrete trajectory points (vehicle locations) reported by the vehicle, i.e., continuous trajectory data under severe anomaly conditions.

[0067] Specifically, this means that while the cloud continuously receives real-time riding data, it typically stores it. When the anti-theft detection determines that the vehicle's status is abnormal and the abnormality level reaches the second preset level, the cloud will immediately extract the trajectory. Trajectory extraction here refers to using the moment the vehicle's abnormal status is determined as a baseline, tracing back and extracting all trajectory points from that moment up to the moment of the abnormality. These extracted trajectory points, arranged in chronological order, constitute a continuous trajectory data segment.

[0068] Next, a structured abnormal trajectory record can be generated based on this continuous trajectory data, and the abnormal trajectory record can be stored. That is, the cloud will generate a structured abnormal trajectory record based on the continuous trajectory data extracted in the previous step, and can store it for subsequent case tracing, data analysis, etc.

[0069] Here, the abnormal trajectory record can be a JSON (JavaScript Object Notation) or XML (Extensible Markup Language) file.

[0070] Then, based on this abnormal trajectory record and the detection results, an alarm message can be generated; that is, the cloud will integrate the generated structured abnormal trajectory record and the current detection results to generate an alarm message. This alarm message can then be sent to the operations and maintenance team so that the relevant operations and maintenance personnel can make informed decisions and intervene.

[0071] In this embodiment of the invention, under severe abnormal conditions, alarm information is generated based on the abnormal trajectory record and sent to the operation and maintenance terminal, so that the operation and maintenance personnel can see a clear and complete abnormal trajectory map. This allows the operation and maintenance personnel to understand the whole picture of the event in a very short time, which helps the operation and maintenance personnel to make quick and accurate decisions, thereby improving the efficiency and success rate of anti-theft response.

[0072] The vehicle anti-theft device provided by the present invention is described below. The vehicle anti-theft device described below can be referred to in correspondence with the vehicle anti-theft method described above.

[0073] Figure 4 This is a structural schematic diagram of the vehicle anti-theft device provided by the present invention, as shown below. Figure 4 As shown, the device is used in the cloud and includes: The acquisition unit 410 is used to receive real-time riding data reported by the vehicle and acquire the user habit model associated with the vehicle. Detection unit 420 is used to perform anti-theft detection based on the real-time riding data and the user habit model, and obtain the detection result; The anti-theft unit 430 is used to send alarm information to the operation and maintenance terminal when the detection result indicates that the vehicle is in an abnormal state, so that the operation and maintenance terminal can perform anti-theft intervention based on the alarm information.

[0074] The vehicle anti-theft device provided by this invention uses a user habit model associated with the vehicle as a personalized benchmark for anti-theft detection and compares it with the real-time riding data reported by the vehicle. It can accurately identify abnormal behaviors that do not conform to the user's normal riding habits. Once an anomaly is detected, an alarm message is immediately sent to the maintenance terminal for timely anti-theft intervention. This overcomes the shortcomings of traditional vehicle anti-theft methods, such as high false alarm rate, lack of personalization, and delayed response. It introduces the user's personalized behavior pattern into the anti-theft logic, realizing the transformation from passive response to proactive warning. This greatly improves the accuracy and real-time performance of anti-theft alarms, reduces false alarm rate and missed alarm rate, and provides a more proactive and intelligent guarantee for vehicle security.

[0075] Based on the above embodiments, the device further includes a modeling unit, used for: Obtain the user's historical riding data associated with the vehicle; Based on the user's historical cycling data, multi-dimensional feature extraction is performed to obtain the user's historical cycling features; The user's historical cycling characteristics are modeled using a time series analysis model to obtain the user habit model.

[0076] Based on the above embodiments, the modeling unit is used for: A user cycling profile is generated based on the user's historical cycling features; the user's historical cycling features include cycling route features, time pattern features, speed and acceleration pattern features, and parking position stability features. The user cycling profile is modeled using a time series analysis model to obtain the user habit model.

[0077] Based on the above embodiments, the modeling unit is also used for: If the detection result indicates that the vehicle is in normal condition, the user habit model is updated with parameters through incremental learning based on the real-time riding data. The user cycling profile is revised based on the updated user habit model.

[0078] Based on the above embodiments, the anti-theft unit 430 is used for: If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a first preset level, an alarm message is sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is the second preset level, the vehicle is controlled to activate an audible and visual alarm and send an alarm message to the maintenance terminal so that the maintenance terminal can generate a remote intervention command based on the alarm message and perform anti-theft intervention based on the remote intervention command. The anomaly level is determined based on the deviation between the real-time cycling data and the user's cycling habits as represented by the user habit model.

[0079] Based on the above embodiments, the anti-theft unit 430 is used for: Based on the real-time riding data, trajectory extraction is performed to obtain continuous trajectory data when the vehicle is in an abnormal state and the abnormality level is the second preset level. Structured abnormal trajectory records are generated based on the continuous trajectory data, and the abnormal trajectory records are stored. Based on the abnormal trajectory record and the detection result, an alarm message is generated and sent to the operation and maintenance terminal.

[0080] This invention also provides a vehicle anti-theft system. Figure 5 This is a structural schematic diagram of the vehicle anti-theft system provided by the present invention, as shown below. Figure 5 As shown, the system includes vehicle 510, cloud 520, user terminal 530, and operation and maintenance terminal 540; The cloud 520 is used to receive real-time riding data collected and reported by the vehicle 510, and perform anti-theft detection based on the real-time riding data and user habit model to obtain detection results; when the detection result indicates that the vehicle status is abnormal and the abnormality level is a first preset level, an alarm message is sent to the user terminal 530 bound to the vehicle 510; when the detection result indicates that the vehicle status is abnormal and the abnormality level is a second preset level, the cloud 520 controls the vehicle 510 to activate the audible and visual alarm, and sends an alarm message to the maintenance terminal 540, so that the maintenance terminal 540 generates a remote intervention command based on the alarm message, and performs anti-theft intervention based on the remote intervention command.

[0081] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a vehicle anti-theft method. This method is applied in the cloud and includes: acquiring a user habit model associated with the vehicle; receiving real-time riding data reported by the vehicle; performing anti-theft detection based on the real-time riding data and the user habit model to obtain a detection result; and, if the detection result indicates that the vehicle's status is abnormal, sending an alarm message to the maintenance terminal so that the maintenance terminal can intervene in the anti-theft process based on the alarm message.

[0082] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, wherein when the program instructions are executed by a computer, the computer is able to execute the vehicle anti-theft method provided by the above methods, the method being applied in the cloud, the method comprising: acquiring a user habit model associated with the vehicle; receiving real-time riding data reported by the vehicle; performing anti-theft detection based on the real-time riding data and the user habit model, and obtaining a detection result; and, if the detection result indicates that the vehicle status is abnormal, sending alarm information to the maintenance terminal, so that the maintenance terminal can perform anti-theft intervention based on the alarm information.

[0084] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program is implemented to perform the vehicle anti-theft methods provided by the above methods. The method is applied in the cloud and includes: acquiring a user habit model associated with the vehicle; receiving real-time riding data reported by the vehicle; performing anti-theft detection based on the real-time riding data and the user habit model to obtain a detection result; and, if the detection result indicates that the vehicle is in an abnormal state, sending an alarm message to the maintenance terminal so that the maintenance terminal can perform anti-theft intervention based on the alarm message.

[0085] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0086] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle anti-theft method, characterized in that, Applied to the cloud, the method includes: Receive real-time riding data reported by the vehicle and obtain the user habit model associated with the vehicle; Anti-theft detection is performed based on the real-time riding data and the user habit model to obtain the detection results; If the detection result indicates that the vehicle is in an abnormal state, an alarm message is sent to the maintenance terminal so that the maintenance terminal can take anti-theft intervention based on the alarm message.

2. The vehicle anti-theft method according to claim 1, characterized in that, The user habit model is determined based on the following steps: Obtain the user's historical riding data associated with the vehicle; Based on the user's historical cycling data, multi-dimensional feature extraction is performed to obtain the user's historical cycling features; The user's historical cycling characteristics are modeled using a time series analysis model to obtain the user habit model.

3. The vehicle anti-theft method according to claim 2, characterized in that, The process of modeling the user's historical cycling characteristics using a time series analysis model to obtain the user habit model includes: A user cycling profile is generated based on the user's historical cycling features; the user's historical cycling features include cycling route features, time pattern features, speed and acceleration pattern features, and parking position stability features. The user cycling profile is modeled using a time series analysis model to obtain the user habit model.

4. The vehicle anti-theft method according to claim 3, characterized in that, The process of performing anti-theft detection based on the real-time cycling data and the user habit model to obtain detection results further includes: If the detection result indicates that the vehicle is in normal condition, the user habit model is updated with parameters through incremental learning based on the real-time riding data. The user cycling profile is revised based on the updated user habit model.

5. The vehicle anti-theft method according to any one of claims 1 to 4, characterized in that, When the detection result indicates that the vehicle's status is abnormal, an alarm message is sent to the maintenance terminal so that the maintenance terminal can take anti-theft intervention based on the alarm message, including: If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a first preset level, an alarm message is sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is the second preset level, the vehicle is controlled to activate an audible and visual alarm and send an alarm message to the maintenance terminal so that the maintenance terminal can generate a remote intervention command based on the alarm message and perform anti-theft intervention based on the remote intervention command. The anomaly level is determined based on the deviation between the real-time cycling data and the user's cycling habits as represented by the user habit model.

6. The vehicle anti-theft method according to claim 5, characterized in that, Sending alarm information to the operation and maintenance terminal includes: Based on the real-time riding data, trajectory extraction is performed to obtain continuous trajectory data when the vehicle is in an abnormal state and the abnormality level is the second preset level. Structured abnormal trajectory records are generated based on the continuous trajectory data, and the abnormal trajectory records are stored. Based on the abnormal trajectory record and the detection result, an alarm message is generated and sent to the operation and maintenance terminal.

7. A vehicle anti-theft device, characterized in that, The device, applied in the cloud, includes: The acquisition unit is used to receive real-time riding data reported by the vehicle and acquire the user habit model associated with the vehicle. The detection unit is used to perform anti-theft detection based on the real-time riding data and the user habit model, and obtain the detection result; An anti-theft unit is used to send an alarm message to the operation and maintenance terminal when the detection result indicates that the vehicle is in an abnormal state, so that the operation and maintenance terminal can take anti-theft intervention based on the alarm message.

8. A vehicle anti-theft system, characterized in that, This includes vehicles, cloud, user terminals, and operations and maintenance terminals; The cloud platform is used to receive real-time riding data collected and reported by the vehicle, and to perform anti-theft detection based on the real-time riding data and user habit model to obtain detection results. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a first preset level, an alarm message is sent to the user terminal bound to the vehicle. If the detection result indicates that the vehicle is in an abnormal state and the abnormality level is a second preset level, the cloud platform controls the vehicle to activate an audible and visual alarm and sends an alarm message to the maintenance terminal, so that the maintenance terminal can generate a remote intervention command based on the alarm message and perform anti-theft intervention based on the remote intervention command.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the vehicle anti-theft method as described in any one of claims 1 to 6.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the vehicle anti-theft method as described in any one of claims 1 to 6.

Images