Data processing method and related apparatus

Abstract

A data processing method and related apparatus are provided. The method comprises: obtaining a classification result of a target object based on radar data from a radar sensor, wherein the radar data comprises movement characteristics of at least one moving object detected by the radar sensor, and the at least one moving object comprises the target object (S301); and determining an alarm strategy for the target object based on the classification result of the target object and current position information derived based on the radar data from the radar sensor (S302). Through the data processing method and related apparatus, the alarm strategy is specially customized for the position state of the target object, which ensures robust performance and reliable detection, and improves the safety of the target object.

Description

Technical Field

[0001] This disclosure relates to the field of communication technology, and in particular to a data processing method, apparatus, device, system and storage medium. Background Technology

[0002] Video / image-based surveillance and radar-based surveillance are commonly used for presence detection. Regarding radar-based surveillance, radar sensors can transmit and receive signals over short ranges, which can be used for context-aware tasks such as presence detection, vital sign monitoring, and moving object localization and tracking. The received signals include information about objects in the environment, and the signals reflected by moving objects may change over time.

[0003] The purpose of the background information is to illustrate information that the applicant believes may be relevant to this disclosure. It is neither necessary nor appropriate to acknowledge that any of the aforementioned information constitutes prior art relative to this disclosure. Summary of the Invention

[0004] In a first aspect, embodiments of this disclosure provide a data processing method, wherein the method includes:

[0005] The classification results of target objects are obtained based on radar data from radar sensors, wherein the radar data includes the movement characteristics of at least one moving object detected by the radar sensors, and at least one moving object includes the target object.

[0006] The alarm strategy for the target object is determined based on the classification results of the target object and its current location information, which is derived from radar data from radar sensors.

[0007] According to the method provided in this disclosure, the classification result of a target object included in at least one moving object is acquired based on radar data from a radar sensor, and the alarm strategy for the target object is determined based on the classification result and current location information of the target object. By using radar data, the device can accurately classify the target object from other moving objects detected by the radar sensor. This method considers the classification result and current location information of the target object to determine an appropriate alarm strategy, which ensures that the alarm strategy is specifically tailored to the location status of the target object, ensuring robust performance and reliable detection, and improving the security of the target object. Furthermore, since the device can continuously process radar data and derive the current location information of the target object in this way, real-time monitoring of the movement of the target object can be achieved, and immediate response and alarm can be provided when necessary. Moreover, this method considers the classification result and current location information of the target object to determine an appropriate alarm strategy. This adaptability supports customized and flexible responses, thereby enhancing the effectiveness of the data processing system.

[0008] In one possible implementation of the first aspect, the movement characteristics of at least one moving object include at least one range-time characteristic in the time domain or a range-Doppler characteristic in the frequency domain. By incorporating the range-time characteristic or range-Doppler characteristic into the analysis of moving objects detected by radar sensors, the system's ability to accurately classify target objects and track their movement can be further enhanced.

[0009] In one possible implementation of the first aspect, obtaining classification results for a target object based on radar data collected from radar sensors includes:

[0010] Derive the distance-time characteristics or distance-Doppler characteristics of the target object based on the movement characteristics of at least one moving object;

[0011] The classification results of the target object are obtained based on the distance-time characteristics or distance-Doppler characteristics of the target object.

[0012] In this way, the system effectively classifies target objects using range-time or range-Doppler characteristics, thereby helping to determine appropriate alarm strategies based on the movement patterns or behaviors of target objects observed through radar data analysis.

[0013] In one possible implementation of the first aspect, obtaining classification results based on the distance-time characteristics or distance-Doppler characteristics of the target object includes:

[0014] The time-domain or frequency-domain characteristics of the target object are input into the first pre-trained model to obtain the classification result. The first pre-trained model is trained using the distance-time characteristics or distance-Doppler characteristics of objects belonging to different categories.

[0015] The first pre-trained model can effectively capture complex patterns and relationships in the distance-time or distance-Doppler characteristics of objects, which can improve the accuracy of classification results. Furthermore, by using the first pre-trained model (which is trained using the distance-time or distance-Doppler characteristics of objects belonging to different categories), the data processing system can effectively utilize the diverse features extracted from radar data for classification tasks.

[0016] In one possible implementation of the first aspect, obtaining classification results based on the distance-time characteristics or distance-Doppler characteristics of the target object includes:

[0017] The classification characteristics of a target object are determined based on its range-time characteristics or range-Doppler characteristics, wherein the classification characteristics include at least one of the target object's radar cross section (RCS) or breathing rate.

[0018] The classification result is obtained based on the classification characteristics of the target object.

[0019] By using classification characteristics that include at least one of the target object's RCS or respiratory rate to obtain classification results, the classification of the target object can be more accurate and detailed, and the method can also make the system more computationally efficient.

[0020] In one possible implementation of the first aspect, the alarm strategy for determining the target object based on the classification result and current location information of the target object includes:

[0021] Determine whether the target object is within the preset area based on its current location information;

[0022] When the target object is determined to be within the preset area, the classification results are used to determine whether the security requirements corresponding to the preset area of ​​the target object are met.

[0023] An alarm is triggered when the safety requirements are determined to be met.

[0024] By considering the current location information of the target object and the security requirements corresponding to the target object's preset area, the system's adaptability to different areas (areas or zones) with specific security requirements can be improved. Furthermore, an alarm can only be triggered when the target object is within the preset area and the security requirements corresponding to that area are met. Therefore, alarm prompts can be appropriately targeted, and unnecessary alarm prompts can be reduced in different scenarios.

[0025] In one possible implementation of the first aspect, the method further includes:

[0026] The current location information of the target object is determined based on radar data from radar sensors and radar data of the target object from at least one other radar sensor.

[0027] By combining radar data from multiple radar sensors, the accuracy and overall reliability of the identified target's current location information can be improved. Furthermore, using multiple radar sensors can expand the system's sensing range.

[0028] In one possible implementation of the first aspect, the radar data from the radar sensor includes first time of arrival (ToA) information of the target object, and the radar data of the target object from at least one other radar sensor includes second ToA information of the target object.

[0029] The information used to determine the current location of the target object includes:

[0030] Obtain the second time of arrival (ToA) information of the target object from at least one other radar sensor;

[0031] The current location information of the target object is determined based on the first ToA information, the second ToA information, and the location information of at least one other radar sensor.

[0032] By using ToA information from multiple radar sensors to determine the current location of a target object, it is possible to support the comprehensive fusion of location information and improve the accuracy of the determined current location information of the target object.

[0033] In one possible implementation of the first aspect, the method further includes:

[0034] Obtain the predicted location information of the target object based on its historical location information;

[0035] Determining the current location information of the target object based on the first ToA information, the second ToA information, and the location information of at least one other radar sensor includes:

[0036] The current position information of the target object is determined based on the first ToA information, the second ToA information, the position information of at least one other radar sensor, and the predicted position information of the target object.

[0037] By considering historical location information and combining it with predicted location, potential delays can be taken into account when acquiring real-time location information. This can improve the accuracy of the current location information of the identified target object and enhance the security of the target object.

[0038] In one possible implementation of the first aspect, determining whether the target object is within the preset area based on the target object's current location information includes:

[0039] The movement pattern of the target object is determined based on its historical location information and current location information. The movement pattern includes the target object's movement speed and movement direction.

[0040] Determine whether the target object is within the preset area based on the target object's movement pattern and the range of the preset area.

[0041] By continuously updating the movement pattern based on current and historical location information, a continuous assessment of the target object's position relative to a preset area can be provided, which can issue alerts or take action in a timely manner when necessary.

[0042] In one possible implementation of the first aspect, determining whether the target object is within the preset area based on the target object's current location information includes:

[0043] The current location information of the target object and the range of the preset area are input into the second pre-trained model to obtain the result of determining whether the target object is within the preset area. The second pre-trained model is trained using predefined safety events and dangerous events related to the preset area.

[0044] In this way, the system's ability to determine whether a target object is within a preset area based on its current location information can be enhanced, and decision-making efficiency can be improved.

[0045] In one possible implementation of the first aspect, the classification result includes the category to which the target object belongs.

[0046] In one possible implementation of the first aspect, the classification result also includes a confidence level corresponding to the classification. The confidence level can indicate the accuracy and reliability of the classification result, which can help improve the accuracy of the determination.

[0047] In a second aspect, embodiments of this disclosure provide a data processing apparatus comprising various modules for performing data processing methods according to the first aspect or any possible implementation thereof.

[0048] In a third aspect, embodiments of this disclosure provide an apparatus including processing circuitry for performing a data processing method according to the first aspect or any possible implementation thereof.

[0049] In a fourth aspect, embodiments of this disclosure provide a wireless communication system, including the device according to the third aspect.

[0050] In a fifth aspect, embodiments of this disclosure provide a computer-readable medium storing computer-executable instructions, which, when executed by a processor, cause the processor to perform a data processing method according to the first aspect or any possible implementation thereof, or according to the second aspect or any possible implementation thereof.

[0051] In a sixth aspect, embodiments of this disclosure provide a computer program that, when executed by a processor, causes the processor to perform a data processing method according to the first aspect or any possible implementation thereof.

[0052] In a seventh aspect, embodiments of this disclosure provide a computer program product including computer-executable instructions that, when executed by a processor, cause the processor to perform a data processing method according to the first aspect or any possible implementation thereof.

[0053] This disclosure provides a data processing method and related apparatus. The classification result of a target object is acquired based on radar data from a radar sensor. The radar data includes the movement characteristics of at least one moving object detected by the radar sensor, wherein the at least one moving object includes the target object. The alarm strategy for the target object is determined based on the classification result and the current location information of the target object, which is derived from the radar data from the radar sensor. By using the radar data, the apparatus can accurately classify the target object from other moving objects detected by the radar sensor. This method considers the classification result and current location information of the target object to determine an appropriate alarm strategy. This ensures that the alarm strategy is specifically tailored to the location state of the target object, ensuring robust performance and reliable detection, and improving the security of the target object. Attached Figure Description

[0054] The following figures illustrate exemplary embodiments of this disclosure by way of example.

[0055] Figure 1 This is a schematic diagram illustrating an example of using radar signals to detect human movement.

[0056] Figure 2 This is a schematic diagram illustrating an example of classified target detection in a hazardous area provided by one or more embodiments of this disclosure.

[0057] Figure 3 A schematic flowchart of a data processing method provided by one or more embodiments of the present disclosure is shown.

[0058] Figure 4 This is a schematic diagram illustrating an example of signal processing steps for detection provided in one or more embodiments of this disclosure.

[0059] Figure 5a This is a schematic diagram illustrating the distance-time characteristics of at least one moving object provided in one or more embodiments of this disclosure.

[0060] Figure 5b This is a schematic diagram of the distance-Doppler characteristics of at least one moving object provided in one or more embodiments of this disclosure.

[0061] Figure 6 This is a schematic diagram illustrating an example of security monitoring using multiple radar sensors provided in one or more embodiments of this disclosure.

[0062] Figure 7 This is a flowchart of the signal processing steps for using radar sensors to ensure the safety of children in dangerous areas.

[0063] Figure 8 A schematic structural diagram of a data processing apparatus provided in one or more embodiments of the present disclosure is shown.

[0064] Figure 9 A schematic structural diagram of a data processing apparatus provided in one or more embodiments of the present disclosure is shown. Detailed Implementation

[0065] In the following description, reference is made to the accompanying drawings, which form part of this disclosure, which illustrate by way of description specific aspects of embodiments of this disclosure or aspects that may be used with embodiments of this disclosure. It should be understood that embodiments of this disclosure can be used in other aspects and include structural or logical variations not shown in the drawings. Therefore, the following detailed description is not intended to be limiting, and the scope of this disclosure is defined by the appended claims.

[0066] To aid in understanding this disclosure, examples of wireless communication systems and devices are described below.

[0067] Figure 1 This is a schematic diagram illustrating an example of using radar signals to detect human movement. Video / image-based monitoring and radar-based monitoring are commonly used for presence detection. Regarding radar-based monitoring, such as... Figure 1 As shown, radar sensors (or radar) transmit and receive signals over short distances, which can be used for context-aware tasks such as presence detection, vital sign monitoring, and location and tracking of moving objects. Figure 1 This illustrates the principle of radar transmitting and receiving pulses reflected from objects in the environment.

[0068] In related technologies, some monitoring methods are based on video / image presence detection and monitoring. Video can be used to detect the presence of human bodies or anomalies when a target approaches the boundary of a restricted area. Additionally, some related technologies employ radar-based monitoring; however, radar sensors are typically used only for detection and ranging, and further detection of human bodies or other objects remains based on video.

[0069] This type of video / image-related monitoring may have the following problems:

[0070] (1) It violates user privacy because it obtains images and videos of the environment and people. Because it uses video, it is not suitable for use in privacy-conscious applications, such as homes or offices.

[0071] (2) It may not perform well in foggy or dark environments, especially in bright or dark environments.

[0072] Furthermore, in the real world, simply knowing the presence of people may not be sufficient; public observations suggest that human surveillance can be helpful in real-life situations. Additionally, different targets may correspond to different security requirements, which can be particularly useful in certain scenarios. Therefore, while ensuring privacy protection, a method is needed for classified target detection (or classified presence detection, which may include the detection of one or more targets and further classification of the detected targets), and for setting appropriate alert strategies for the detected targets.

[0073] In view of the foregoing, this disclosure provides a data processing method in which a classification result of a target object (or target) is acquired based on radar data from a radar sensor, and an alarm strategy for the target object is determined based on the classification result and the target object's current location information. This method considers both the target object's classification result and current location information to determine an appropriate alarm strategy. This ensures that the alarm strategy is specifically tailored to the target object's location status, guaranteeing robust performance and reliable detection, and improving the target object's security.

[0074] Figure 2 This is a schematic diagram illustrating an example of classified target detection in a hazardous area provided by one or more embodiments of this disclosure. The purpose of this disclosure is to detect the presence of classified targets in a specific area for safety purposes. This disclosure uses radar sensors to detect the presence and classifies the target from other interfering objects. Movement patterns are then analyzed, and the distance between the classified target and the safety boundary of the restricted area is continuously monitored. An alarm is issued when the distance between the classified target and the boundary falls below a set threshold, indicating that the target is in a hazardous area. Therefore, this disclosure solves the problem of target safety in areas where the safety boundary is restricted, thus requiring continuous monitoring of classified targets under all weather and lighting conditions.

[0075] Before detailing the specific embodiments of this disclosure, the concept of the proposed solution will be briefly described.

[0076] The radar signal received by radar includes information about objects in the environment. If the radar signal is observed over a period of time, it may be noticed that static objects such as chairs, tables, floors, and ceilings have constant signals across multiple time frames at the receiver, and can therefore be removed using an MTI filter, etc. On the other hand, radar signals reflected by moving objects change over time. Therefore, by applying an MTI filter to the radar signal, which includes the raw data, at the receiver, moving objects can be detected. The radar signal here can have multiple radar frames; that is, the radar signal received by the radar sensor is a signal with multiple durations. Each frame can include distance information for one or more static objects and / or one or more moving objects. For example, each frame can include the distance between the radar sensor and any object within its coverage area. Therefore, an MTI filter can be used to determine which portion of the energy is associated with one or more moving objects and which portion is associated with one or more static objects, thereby removing information about one or more static objects.

[0077] Although many moving objects may exist in the environment, such as fans, moving curtains, children, adults, and other small moving objects, these objects can be classified based on characteristics such as time and spectral properties. The presence of classified targets can be identified by using these methods. Signal processing and AI methods can be used to detect human targets or pets and distinguish them from other static or moving clutter objects in the environment.

[0078] Another crucial piece of information is the distance from the target to the radar. One possible approach is to obtain this information using the ToA (or ToA data) of the radar signal. Furthermore, multiple radars within the network can be used for target localization to improve accuracy. If multiple radar sensors are used, the distance from each radar to the target can be obtained by measuring the ToA of the radar signal. Then, multiple distances can be acquired using algorithms such as multilateral localization to determine the 2D or 3D position of the (human) target.

[0079] This disclosure provides a radar sensor system for protecting classified targets. A radar-enabled device is provided for classified target detection and localization in a 2D / 3D scene. A method and process are provided for a radar sensor to detect and classify targets by comparing their characteristics with one or more predefined characteristics of specific classified targets. For example, key features extracted to classify a target as a pet, child, or adult may be the target's RCS, respiratory rate, and spectral characteristics. An algorithm is provided for sensing hazard using the target's movement patterns and the classified target's position relative to an area boundary.

[0080] Specific embodiments of this disclosure will be described in detail in the following description.

[0081] Figure 3 A schematic flowchart of a data processing method provided by one or more embodiments of this disclosure is shown. This method can be applied to wireless communication or other communication systems using radar sensors, and is not limited thereto. The data processing method can be implemented by a data processing apparatus / device. Optionally, the data processing apparatus itself may be integrated with radar (or other components with similar functions), or it may not be integrated with radar (or other components with similar functions), but may acquire or store radar signals, and is not limited thereto. Figure 3 As shown, the method may include the following steps.

[0082] S301: Obtain the classification result of the target object based on radar data from the radar sensor, wherein the radar data includes the movement characteristics of at least one moving object detected by the radar sensor, and the at least one moving object includes the target object.

[0083] In this embodiment, the device acquires the classification result of the target object based on radar data from a radar sensor. The radar data includes the movement characteristics of at least one moving object detected by the radar sensor, where the at least one moving object includes the target object. The radar data can be analyzed to distinguish and identify the target object from other moving objects detected by the radar sensor. The radar data acquired from the radar sensor can include various types of information and parameters that can be used to detect, track, and analyze targets in the surrounding environment. The radar data can be included in the radar signal; therefore, acquiring radar data can be implemented by receiving a radar signal, or the radar data can be acquired from another device that receives a radar signal. In this case, acquiring radar data can also be referred to as input radar data.

[0084] In one possible implementation, the device can acquire radar data from multiple radar sensors.

[0085] In one possible implementation, the classification result can include the category / class to which the target object belongs. Categories can include adults, children, the elderly, animals (e.g., pets), etc. The classification of the target object will show which category it belongs to.

[0086] In one possible implementation, the classification result may also include a confidence level corresponding to the classification. The confidence level can serve as an indication of how certain or reliable the classification result is, or the probability that a target object belongs to that classification / category. For example, in the case of classification by a model, the confidence level can reflect how certain or reliable the model is in assigning the target object to a particular category or class. The device can make decisions based on the confidence level, which helps improve the accuracy of subsequent determinations in step S302.

[0087] In one possible implementation, the motion characteristics of at least one moving object may include at least one range-time characteristic in the time domain or a range-Doppler characteristic in the frequency domain. The range-time characteristic can represent the relationship between the distance (range) between the object and the radar sensor and the time required for the radar signal to reach the object and return (round-trip time). The range-Doppler characteristic can represent the relationship between the distance (range) between the object and the radar sensor and the Doppler frequency shift of the detected object. The Doppler frequency shift component of the range-Doppler characteristic is related to the frequency change of the returned radar signal caused by the motion of the detected object. As the object moves toward or away from the radar sensor, the radar wave experiences a frequency shift, which is a Doppler frequency shift and is proportional to the radial velocity of the object. By incorporating the range-time characteristic or the range-Doppler characteristic into the analysis of moving objects detected by the radar sensor, the system's ability to accurately classify target objects and track their movement can be further enhanced.

[0088] In one possible implementation, the device can acquire radar signals from a radar sensor, process the radar signals using an MTI filter to obtain processed radar signals, and perform moving object detection based on the processed radar signals to identify at least one moving object. Figure 4 This is a schematic diagram illustrating an example of signal processing steps for detection provided in one or more embodiments of this disclosure. Figure 4As shown, the signal received by the radar sensor via the receiving antenna can be sampled by the radar sensor's analog-to-digital converter (ADC) to obtain raw data represented in in-phase and quadrature (I / Q) sample form. The signal received by the radar sensor can include information about the environment, including static and moving objects. The signal can then be processed using an MTI filter to remove static objects in the environment, such as signals from the floor, ceiling, chairs, and other static objects. The filtered signal can include only information about moving objects (e.g., people, fans, or other targets whose RCS changes over time). After the received signal is processed by the MTI filter to remove static objects in the environment, radar frames of the filtered signal (or clutter-removed signal) can be combined in a matrix for a certain period of time. The size of the matrix depends on the slow-time sampling frequency of the radar sensor and can be adjusted experimentally. This matrix is ​​in the time and range domains. This matrix is ​​then processed using algorithms such as FFT to obtain the range-Doppler matrix. Then, the presence of a moving object can be detected using the distance-Doppler matrix by applying algorithms such as the cell averaging constant false alarm rate (CA-CFAR). Here, the CA-CFAR algorithm (or CFAR detection) can be used solely to detect the presence of a moving object; therefore, the CFAR decision is to indicate whether a moving object exists.

[0089] In one possible implementation, the device can derive the range-time characteristics or range-Doppler characteristics of a target object based on the movement characteristics of at least one moving object, and can obtain the classification result of the target object based on the range-time characteristics or range-Doppler characteristics of the target object. The movement characteristics of at least one moving object can include the range-time characteristics or range-Doppler characteristics of all moving objects detected by the radar sensor. Figure 5a This is a schematic diagram illustrating the distance-time characteristics of at least one moving object provided in one or more embodiments of this disclosure. Figure 5b This is a schematic diagram illustrating the distance-Doppler characteristics of at least one moving object provided in one or more embodiments of this disclosure. For example... Figure 5a and Figure 5bAs shown, the range-time characteristic can be represented by a range-time (RT) plot, and the range-Doppler characteristic can be represented by a range-Doppler (RD) plot. As mentioned above, a radar signal can have multiple radar frames. Therefore, the combined clutter removal signal here refers to the combination of range information of one or more moving objects included in multiple radar frames. For each of the multiple radar frames, the included range information can be reflected in the RT plot, where the horizontal axis represents time and the vertical axis represents range. The RD plot can be obtained by applying an FFT operation to the RT plot.

[0090] Clustering algorithms can be applied based on the movement characteristics of at least one moving object to derive the range-time characteristics (time-domain characteristics) or range-Doppler characteristics (frequency-domain characteristics) of the target object. Each cluster can represent a different target object. The specific method for deriving the range-time or range-Doppler characteristics of the target object based on the movement characteristics of at least one moving object is not limited here. After acquiring the range-time or range-Doppler characteristics of the target object, the device can classify the target object based on these characteristics and obtain the classification results. In this way, the system can effectively classify one or more target objects using range-time or range-Doppler characteristics to obtain the target object, thereby helping to determine appropriate alarm strategies based on the movement patterns or behaviors of moving objects observed through radar data analysis.

[0091] To achieve classified object detection, various manually designed features and AI-extracted features can be used for classification tasks. Classification tasks can involve separating humans from non-human targets (such as pets, fans, or curtains). Then, for some applications, adult / child classification can be further considered after detecting a human. Additionally, infants can be classified based on different features. The input radar data can be represented in the time domain (RT map, which includes the range-time characteristics of at least one moving object) or the frequency domain (RD map, which includes the range-Doppler characteristics of at least one moving object).

[0092] In one possible implementation, the device can input the temporal or frequency domain characteristics of the target object into a first pre-trained model to obtain classification results. This first pre-trained model can be trained using distance-time or distance-Doppler characteristics of objects belonging to different categories / classes. For example, the temporal characteristics of one or more objects can be located on a Time-Retrieval (RT) map, and the frequency characteristics of one or more objects can be located on an RD map. The RT and RD maps can then serve as two distinct inputs to the first pre-trained model, thus allowing the temporal and / or frequency characteristics to be fed into an AI algorithm. The AI-based classification algorithm can be used to classify targets based on features from the temporal domain (RT) and / or features from the frequency domain (RD map). The first pre-trained model can be a deep learning AI model, such as a convolutional neural network (CNN) or long short-term memory (LSTM), or other machine learning models, such as logistic regression, decision trees, random forests, histogram of oriented gradients (HOG), or support vector machines (SVM). The model can be trained first using distance-time or distance-Doppler characteristics of objects belonging to different categories. For example, the model can be trained using data from each case, such as data from individuals, children, and other objects (e.g., pets, fans, vehicles, curtains), depending on the application scenario considered. After the model is trained, online data from radar sensors can be used as input to the model so that the trained model can produce classification results.

[0093] The first pre-trained model can effectively capture complex patterns and relationships in the distance-time or distance-Doppler characteristics of objects, which can improve the accuracy of classification results. Furthermore, by using the first pre-trained model (which is trained using the distance-time or distance-Doppler characteristics of objects belonging to different categories), the data processing system can effectively utilize the diverse features extracted from radar data for classification tasks.

[0094] Signal processing-based algorithms can also be used to make the final decision. These algorithms can classify targets using their radar cross section (RCS). In one possible implementation, the device can determine the target's classification characteristics based on its range-time or range-Doppler characteristics, where the classification characteristics include at least one of the target's RCS or breathing rate. The device can then obtain a classification result based on these classification characteristics.

[0095] By using signal processing algorithms, the RCS of a target can be used as a classification criterion for targets such as adults and children, because RCS is proportional to the size of the target. If the target is small, the echo signal (reflected signal) from the target object will extend along a shorter distance in the RT map compared to a larger target. A threshold can be set to separate the RCS value into different categories corresponding to different target types. When a target is detected and its RCS value is obtained, it can be compared with a predetermined threshold. If the RCS value exceeds the threshold, the target can be classified into one type (e.g., adult), and if it is below the threshold, the target can be classified into another type (e.g., child). The threshold here is used to distinguish different categories of objects and can be set according to actual needs.

[0096] Furthermore, respiratory rates differ significantly between children and adults, so spectral features can also be used for classification between adults and children. The RD map includes the spectral characteristics of the target object. For example, when determining the distance-Doppler characteristics of a moving object on an RD map, the respiratory rate can be determined as the maximum frequency component within it. If we start with a RT map, we can apply an FFT operation to the RT map to obtain the RD map, and then the respiratory rate can be determined in the same way. Features based on respiratory rate include the average or range of respiratory frequencies, which can be used to distinguish different types of target objects. Signal processing algorithms can quantify and analyze these features to capture the unique breathing patterns associated with adults, children, and pets. The respiratory rate ranges differ between adults, children, and pets. Adults have a higher respiratory rate range compared to children. Generally, pets breathe much faster than humans. A normal resting adult respiratory rate is generally considered to be 12 to 20 breaths per minute. In the elderly population, a respiratory rate higher than 28 breaths per minute is considered tachypnea (abnormally rapid breathing). Children typically have a higher respiratory rate than adults. In the first two years of life, the median respiratory rate decreases from about 44 breaths per minute at birth to about 26 breaths per minute in the second year. Figure 5a RT diagram in Figure 5b The RD plot and respiratory rate-based features can be used in signal processing algorithms to classify targets.

[0097] By using classification features, including at least one of the target object's RCS or respiration rate, to obtain classification results, the classification of the target object can be more accurate and detailed, while also improving the system's computational efficiency. For the application under consideration, the choice between classification based on manually designed features or classification based on an AI model depends on various factors, such as memory, computational requirements, and the availability of training data. In some cases, combining signal processing and AI methods can achieve better performance. Combining manually designed features with AI models leverages the advantages of both approaches. Manually designed features can provide structured input representations for AI models, which can improve performance; the interpretability of manually designed features can complement the black-box nature of AI models. Furthermore, if both signal processing and AI method results are applied, the results derived from signal processing and AI methods can be compared for calibration.

[0098] S302: Determine the alarm strategy for the target object based on the classification results and current location information of the target object. The current location information is derived from radar data from radar sensors.

[0099] In this embodiment, the device determines an alarm strategy for the target object based on the classification results and the target object's current location information, which is derived from radar data from a radar sensor. For example, the radar data includes distance information reflecting the distance between the target object and the radar sensor receiving the radar signal; therefore, the current location information can be derived based on this distance information (or, if necessary, further derived based on the location information of the radar sensor). The target object may correspond to specific security requirements; therefore, a customized alarm strategy can be determined based on the classification results and the target object's current location information.

[0100] According to the method provided in this disclosure, the classification result of a target object included in at least one moving object is acquired based on radar data from a radar sensor, and the alarm strategy for the target object is determined based on the classification result and current location information of the target object. By using radar data, the device can accurately classify the target object from other moving objects detected by the radar sensor. This method considers the classification result and current location information of the target object to determine an appropriate alarm strategy, which ensures that the alarm strategy is specifically tailored to the location status of the target object, ensuring robust performance and reliable detection, and improving the security of the target object. Furthermore, since the device can continuously process radar data and derive the current location information of the target object in this way, real-time monitoring of the movement of the target object can be achieved, and immediate response and alarm can be provided when necessary. Moreover, this method considers the classification result and current location information of the target object to determine an appropriate alarm strategy. This adaptability supports customized and flexible responses, thereby enhancing the effectiveness of the data processing system.

[0101] In one possible implementation, the device can determine whether the target object is within a preset area based on its current location information. If the target object is determined to be within the preset area, the device can determine whether the security requirements corresponding to the preset area are met based on the classification results. If the security requirements are met, the device can trigger an alarm.

[0102] When the device determines that a target object is within a preset area based on its current location information, it can also determine whether the security requirements corresponding to the preset area are met. Security requirements may be related to the preset area. Alarm prompts may include at least one of the following: audible alarms, visual alarms, mobile notifications, smart home device alarms, wearable device notifications, SMS or telephone calls, application notifications, etc. In some cases, alarm prompts can trigger actions within a broader security system, such as activating surveillance cameras, unlocking emergency access doors, or initiating automatic emergency responses. The specific form of the alarm prompt can be determined according to actual needs and is not limited here.

[0103] For example, near a balcony, where the presence of adults is generally considered safe, the focus of monitoring might primarily revolve around the distance of children from potentially dangerous edges. This could involve detecting whether a child is approaching a balcony edge or entering an unsafe area, and triggering alarms and interventions accordingly. In other words, when the preset area includes a preset area of ​​the balcony, the safety requirement could be monitoring the safety of children. When a target is identified as being within the preset area, and the target is classified as a child, the device can trigger an alarm.

[0104] In a bedside environment, the focus is likely on ensuring the safety of children while they sleep or play, with monitoring aimed at identifying any situations that may pose a risk to their health. That is, when the preset area includes a preset area of ​​the bedside environment, the safety requirement can be monitoring for the safety of children. When a target is identified as being within the preset area and the target is classified as a child, the device can trigger an alarm.

[0105] When approaching an open, elevated platform, both children and adults need to be monitored. Open, elevated platforms can include: scenic hills, observation decks and lookout points, amusement park ride platforms (such as roller coaster platforms or Ferris wheel platforms), and construction and work platforms. In this case, a pre-defined area can include a pre-defined area of ​​the open, elevated platform (which could be a safety line), and the safety requirement could be monitoring the safety of both children and adults. When a target is identified as being within the pre-defined area, and the target is classified as either a child or an adult, the device can trigger an alarm.

[0106] In some scenarios, pedestrian access is prohibited. In such cases, the focus of monitoring will be ensuring that only vehicles, not pedestrians, enter restricted areas. For example, certain roads, bridges, or tunnels may be designated for vehicle use only. Similarly, airport runways and taxiways may be strictly reserved for aircraft movement. In these situations, safety requirements could include monitoring the safety of both children and adults. When a target is identified as being within a pre-defined area and its classification as either a child or an adult, the device can trigger an alarm.

[0107] By considering the current location information of the target object and the security requirements corresponding to the target object's preset area, the system's adaptability to different areas (or zones) with specific security requirements can be improved. This method facilitates real-time response by continuously monitoring the target object's location and dynamically assessing security requirements. This enables timely alerts when action is required, ensuring that potential security issues are addressed promptly. Furthermore, alerts are only triggered when the target object is within the preset area and the security requirements corresponding to that area are met. Therefore, alerts can be appropriately targeted, and unnecessary alerts can be reduced in different scenarios.

[0108] In one possible implementation, the device can determine the current location information of a target object based on radar data from radar sensors and radar data of the target object from at least one other radar sensor. Multiple radar sensors can be installed within the monitoring area. The field of view (FOV) of these radar sensors can cover the entire region of interest. The radar signal resolution can be fine enough to detect small changes in distance or movement of the target. For example, three radars may be required for a two-dimensional (2D) localization scene, while four radars may be required for a three-dimensional (3D) localization scene. A 3D scene may be necessary when height information is important, such as when children climb over the boundary wall of a balcony in a high-rise residential building. Figure 6 In this system, the complete flowchart can be represented by four radar sensors. However, if the intended application is a 2D scene, three radars can also be used. In situations with limited space and geometry, even two radars may be sufficient for 2D positioning. The number of radar sensors depends on the application scenario.

[0109] By combining radar data from multiple radar sensors, the accuracy and overall reliability of the identified target's current location information can be improved. Furthermore, using multiple radar sensors can expand the system's sensing range.

[0110] In one possible implementation, radar data from the radar sensor may include a first time of arrival (ToA) information of the target object, and radar data from at least one other radar sensor may include a second ToA information of the target object. The device can acquire the second ToA information of the target object from at least one other radar sensor, and the device can determine the current position information of the target object based on the first ToA information, the second ToA information, and the position information from at least one other radar sensor.

[0111] In one possible implementation, the device can acquire the predicted location information of the target object based on its historical location information. The device can also determine the current location information of the target object based on first ToA information, second ToA information, location information from at least one other radar sensor, and the predicted location information. By considering historical location information and combining it with the predicted location, potential delays can be taken into account when acquiring real-time location information. This can improve the accuracy of the determined current location information of the target object and enhance the security of the target object.

[0112] In one possible implementation, the device can acquire the measurement state using a multilateral positioning algorithm or similar method based on the position information of at least one other radar sensor and the second ToA information of the target object. In another possible implementation, the device can acquire the future state using a multilateral positioning algorithm or similar method based on the historical position information of the target object. The historical position information of the target can be stored in the device or acquired from at least one other radar sensor. In yet another possible implementation, the device can determine the current position information of the target object based on at least one of the measurement state or the future state.

[0113] By using ToA information from multiple radar sensors to determine the current location of a target object, it is possible to support the comprehensive fusion of location information and improve the accuracy of the determined current location information of the target object.

[0114] Figure 6 This is a schematic diagram illustrating an example of using multiple radar sensors for security monitoring. For example... Figure 6 As shown, after data acquisition, MTI filtering, and CFAR decision-making, the distance to the target can be detected based on the ToA of the signal. Figure 6 The system has multiple radar sensors. One of these sensors can act as the master node, while the others can act as anchor nodes (specific examples of the other radar sensors mentioned above). The anchor nodes perform presence detection. If the CFAR decision is positive (meaning the radar detects the target's presence), the radar can find the target's distance to the radar based on the strongest amplitude sample in the radar frame. The ToA of the strong sample is used to calculate the target's distance. The target distance information from the anchor nodes is transmitted to the master node using wireless technologies such as Zigbee, Wireless Fidelity (Wi-Fi), and Bluetooth (BLE). It should be noted that... Figure 6 This shows a scenario where all nodes detect the presence of a target object (or simply a target, which could be a moving child or an adult, etc.), and the classification of the target is completed at the master node.

[0115] Since multiple radar sensors can be used, data from these radars will be collected and processed. One radar is designated as the master anchor point and will perform most of the data processing tasks. The remaining "N" radar sensors will be anchor nodes, primarily using algorithms such as CFAR for presence detection, followed by target ranging. The target's distance is determined based on the ToA of the signal received by the radar.

[0116] Because there are multiple radar sensors, each performs CFAR detection for the presence of targets in the same area. Since the radars are located at different positions in space, a target might be detected by one radar but missed by another. Therefore, the final target presence detection decision is made at the master node. If two or more of the anchor points and the master node will perform target presence detection, it means that a target is likely to exist in the area, so the master node will make the target detection decision and handle other steps such as localization and safety warning prompts. Localization is initiated by the master node after receiving distance information from at least two anchor points. Here, the master node can be a fixed radar sensor or the radar sensor with the strongest processing power.

[0117] For localization, ToA data from "N" anchor nodes (acting as the master node) can be used and combined through algorithms such as multilateral localization. The position of each anchor node is known to the master node. The position information of each node and the distance between the target and each anchor point are used to estimate the target's position in 2D / 3D space. Since the distance information may include noise such as measurement noise or system noise, a state estimation algorithm can be used to reduce noise. The state estimation algorithm can use the target's historical position and predict the target's future state based on the history. Then, the target's current position is estimated based on the predicted state and the measured state. One such estimation algorithm is the Kalman filter.

[0118] After locating the classified target within the boundary of the safe environment, the next step is to analyze the location data and input features to detect potential hazards. Classified target movement patterns and tracking information can be used for hazard detection. Hazard detection will have predefined criteria. These hazard conditions can be manually defined, such as the coordinates of the hazard boundary, and used in signal processing algorithms. If the current target coordinates are close to the predefined coordinates of the hazard boundary, a hazard detection warning may be generated.

[0119] In one possible implementation, the device can determine the target object's movement pattern based on its historical and current location information. The movement pattern can include the target object's speed and direction of movement. The device can then determine whether the target object is within a preset area based on its movement pattern and the range of that area. By continuously updating the movement pattern based on current and historical location information, a continuous assessment of the target object's position relative to the preset area can be provided, enabling timely alerts or action when necessary.

[0120] In one possible implementation, the device can input the current location information of the target object and the range of a preset area into a second pre-trained model to obtain a result determining whether the target object is within the preset area. The second pre-trained model can be trained using predefined safety and danger events associated with the preset area. Furthermore, the second pre-trained model can evaluate whether the target object is within the preset area using the input data and information obtained from pre-training. This approach enhances the system's ability to determine whether a target object is within a preset area based on its current location information and improves decision-making efficiency.

[0121] However, in the case of machine learning algorithms, data consisting of both hazardous and safe cases can be provided during the training phase to help the model learn. The learned model can then perform real-time hazard detection based on the results of the machine learning algorithm. The safety alert algorithm is implemented at the master node (MN).

[0122] For ease of understanding, the following description will depict exemplary scenarios of the control methods of this disclosure. It should be noted that the exemplary scenarios are for illustrative purposes only and should not be construed as limiting the solutions of this disclosure.

[0123] Figure 7 This is a flowchart of the signal processing steps for using radar sensors to ensure the safety of children in dangerous areas.

[0124] like Figure 7 As shown, assuming the target object is a child, at least one radar sensor can send a sensing signal and receive an echo signal. The number of radar sensors is N, where N is a positive integer. The N radar sensors may include a master node, and the other radar sensors may be one or more anchor nodes. The data processing device can perform signal processing on the raw received signal and then perform presence detection based on the processed signal. The radar received signal may include information about the environment, which includes static objects and moving objects. Presence detection is used to detect at least one moving object in the environment. The movement characteristics of at least one moving object may include at least one distance-time characteristic in the time domain or distance-Doppler characteristic in the frequency domain. If the presence of at least one moving object is detected, the device can perform feature extraction on the movement characteristics, or the movement characteristics can be input into a pre-trained model to perform deep learning, thereby achieving target classification. When the presence of a moving object is determined ( Figure 7When a child is shown, ToA calculations can be performed at all N radars, and ToA information can be sent from each anchor node to the master node. Using multiple radars, the 2D / 3D position of the target object can be determined, and based on its position in space, the master node can determine whether the target is in a safe area or a danger zone.

[0125] According to one or more embodiments of this disclosure, a data processing method can detect classified targets within a region, protecting privacy while performing hazard detection for classified targets. Compared to target classification using raw radar data, the identified features used for target classification can provide more accurate results. Furthermore, the method of this disclosure can be used in various scenarios, including 2D / 3D environments, and for different target classifications, such as pets / humans, adults / children, and curtains / people, depending on the application scenario. Moreover, in this way, security assessments are not affected by environmental conditions such as fog / lighting.

[0126] The data processing method provided in one or more embodiments of this disclosure can be applied to safety-critical environments where it is necessary to monitor certain classified targets within specific boundaries of a region. The boundaries of the region can be 2D or 3D space.

[0127] Figure 8 A schematic structural diagram of a data processing apparatus provided by one or more exemplary embodiments of this disclosure is shown. The data processing apparatus can be applied to the data processing device described above, or installed in or applied to a chip in a data processing device or any other equipment, module, circuit, or unit capable of implementing the steps of the data processing device in the above method embodiments. Figure 8 As shown, the data processing device 800 may include:

[0128] The acquisition module 801 is used to acquire the classification result of the target object based on radar data from the radar sensor, wherein the radar data includes the movement characteristics of at least one moving object detected by the radar sensor, and the at least one moving object includes the target object.

[0129] The processing module 802 is used to determine the alarm strategy for the target object based on the classification result of the target object and the current location information, which is derived from radar data from the radar sensor.

[0130] In one possible implementation, the movement characteristics of at least one moving object include at least one distance-time characteristic of the moving object in the time domain or a distance-Doppler characteristic in the frequency domain.

[0131] In one possible implementation, the processing module 802 is used for:

[0132] Derive the distance-time characteristics or distance-Doppler characteristics of the target object based on the movement characteristics of at least one moving object;

[0133] The classification results of the target object are obtained based on the distance-time characteristics or distance-Doppler characteristics of the target object.

[0134] In one possible implementation, the processing module 802 is used for:

[0135] The time-domain or frequency-domain characteristics of the target object are input into the first pre-trained model to obtain the classification result. The first pre-trained model is trained using the distance-time characteristics or distance-Doppler characteristics of objects belonging to different categories.

[0136] In one possible implementation, the processing module 802 is used for:

[0137] The classification characteristics of a target object are determined based on its range-time characteristics or range-Doppler characteristics, wherein the classification characteristics include at least one of the target object's radar cross section (RCS) or breathing rate.

[0138] The classification result is obtained based on the classification characteristics of the target object.

[0139] In one possible implementation, the processing module 802 is used for:

[0140] Determine whether the target object is within the preset area based on its current location information;

[0141] When the target object is determined to be within the preset area, the classification results are used to determine whether the security requirements corresponding to the preset area of ​​the target object are met.

[0142] An alarm is triggered when the safety requirements are determined to be met.

[0143] In one possible implementation, the processing module 802 is used for:

[0144] The current location information of the target object is determined based on radar data from radar sensors and radar data of the target object from at least one other radar sensor.

[0145] In one possible implementation, radar data from the radar sensor includes the target object's first time of arrival (ToA) information, and radar data from at least one other radar sensor includes the target object's second ToA information.

[0146] The acquisition module 801 is used to acquire the second time of arrival (ToA) information of the target object from at least one other radar sensor;

[0147] The processing module 802 is used to determine the current location information of the target object based on the first ToA information, the second ToA information and the location information of at least one other radar sensor.

[0148] In one possible implementation, the acquisition module 801 is used for:

[0149] Obtain the predicted location information of the target object based on its historical location information;

[0150] The processing module 802 is used for:

[0151] The current position information of the target object is determined based on the first ToA information, the second ToA information, the position information of at least one other radar sensor, and the predicted position information of the target object.

[0152] In one possible implementation, the processing module 802 is used for:

[0153] The movement pattern of the target object is determined based on its historical location information and current location information. The movement pattern includes the target object's movement speed and movement direction.

[0154] Determine whether the target object is within the preset area based on the target object's movement pattern and the range of the preset area.

[0155] In one possible implementation, the processing module 802 is used for:

[0156] The current location information of the target object and the range of the preset area are input into the second pre-trained model to obtain the result of determining whether the target object is within the preset area. The second pre-trained model is trained using predefined safety events and dangerous events related to the preset area.

[0157] In one possible implementation, the classification result includes the category to which the target object belongs.

[0158] In one possible implementation, the classification result also includes the confidence level corresponding to the classification.

[0159] The aforementioned units or modules can be implemented using hardware, one or more components or devices that execute software, or a combination thereof. For example, one or more of the units or modules can be integrated circuits, such as programmable FPGAs, GPUs, or ASICs. It should be understood that if the aforementioned modules are implemented using software for execution by a processor, etc., these modules can be retrieved by the processor, in whole or in part, individually or collectively, for processing, or in one or more instances as needed, and these modules themselves can include instructions for further deployment and instantiation.

[0160] Those skilled in the art should understand that, in conjunction with the relevant descriptions of the data processing methods in the embodiments of this disclosure, the relevant descriptions of the above modules in the embodiments of this disclosure can be understood.

[0161] This disclosure provides a data processing apparatus, including a processing circuit for executing any of the data processing methods described above, which are performed by the data processing device. It should be understood that the data processing device can also execute the steps performed by the central device in the above method embodiments, which will not be elaborated further here.

[0162] This disclosure provides a communication system, including the data processing apparatus or data processing device described above.

[0163] like Figure 9 As shown, the data processing device 900 may include a memory 901 and a processor 902, wherein a computer program is stored in the memory 901 and is executed by the processor 902 to implement the corresponding steps described in the method embodiments. The relevant descriptions of these steps can be understood with reference to the relevant descriptions of the data processing methods in the embodiments of this disclosure. Furthermore, the data processing device 900 may also include a communication interface 903 for communicating with other devices. The computer program (also referred to as a program, software, software application, or code) includes machine instructions of a programmable processor and may be implemented using high-level procedural and implicit programming languages ​​or assembly / machine language.

[0164] One embodiment of this disclosure provides a communication system including a first processing circuit. The first processing circuit is used to perform the steps executed by a data processing device in any of the above-described data processing methods.

[0165] This disclosure provides a computer-readable medium storing computer-executable instructions that, when executed by a processor, cause the processor to perform any of the above-described data processing methods.

[0166] This disclosure provides a computer program product including computer-executable instructions that, when executed by a processor, cause the processor to perform any of the above-described data processing methods.

[0167] In the above embodiments, implementation can be achieved, in whole or in part, by software, hardware, firmware, or any combination thereof. When implemented by software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, they produce, in whole or in part, the processes or functions provided in the embodiments of this disclosure. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access, or a data storage device such as a server or data center, including one or more available media for integration together. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0168] Although the embodiments of this disclosure describe methods and processes by way of steps performed in a certain order, one or more steps in the methods and processes may be omitted or modified as appropriate. Where appropriate, one or more steps may be performed in an order other than that described.

[0169] This disclosure includes various embodiments, not only method embodiments but also other embodiments, such as apparatus embodiments and embodiments involving non-transitory computer-readable storage media. Embodiments may be combined individually or in combination with the features disclosed herein.

[0170] Although this disclosure refers to illustrative embodiments, it is not intended to be interpreted in a limiting sense. Various modifications and combinations of the illustrative embodiments, as well as other embodiments of this disclosure, will be apparent to those skilled in the art upon reference to this specification.

[0171] Features disclosed herein in the context of any particular embodiment may or alternatively be implemented in other embodiments. For example, method embodiments may or alternatively be implemented in apparatus, system, and / or computer program product embodiments. Furthermore, although embodiments are described primarily in the context of methods and apparatus, other implementations are contemplated, for example, as instructions stored in one or more non-transitory computer-readable media. These media may store programs or instructions for performing any of the various methods consistent with this disclosure.

Claims

1. A data processing method, characterized in that, include: The classification result of the target object is obtained based on radar data from radar sensors, wherein the radar data includes the movement characteristics of at least one moving object detected by the radar sensors, and the at least one moving object includes the target object; The alarm strategy for the target object is determined based on the classification result and the current location information of the target object, wherein the current location information is derived from the radar data from the radar sensor.

2. The method according to claim 1, characterized in that, The mobility characteristics of the at least one moving object include the distance-time characteristics of the at least one moving object in the time domain or the distance-Doppler characteristics in the frequency domain.

3. The method according to claim 2, characterized in that, The classification results of the target object obtained based on radar data from radar sensors include: Based on the movement characteristics of the at least one moving object, derive the distance-time characteristics or distance-Doppler characteristics of the target object; The classification result of the target object is obtained based on the distance-time characteristics or distance-Doppler characteristics of the target object.

4. The method according to claim 3, characterized in that, The process of obtaining the classification result based on the distance-time characteristics or distance-Doppler characteristics of the target object includes: The time-domain or frequency-domain characteristics of the target object are input into a first pre-trained model to obtain the classification result, wherein the first pre-trained model is trained using the distance-time characteristics or distance-Doppler characteristics of objects belonging to different categories.

5. The method according to claim 3, characterized in that, The process of obtaining the classification result based on the distance-time characteristics or distance-Doppler characteristics of the target object includes: The classification characteristics of the target object are determined based on the range-time characteristics or range-Doppler characteristics of the target object, wherein the classification characteristics include at least one of the radar cross-section (RCS) or breathing rate of the target object; The classification result is obtained based on the classification characteristics of the target object.

6. The method according to any one of claims 1 to 5, characterized in that, The alarm strategy for determining the target object based on the classification result and current location information of the target object includes: Based on the current location information of the target object, determine whether the target object is within the preset area; When the target object is determined to be within the preset area, the security requirements corresponding to the preset area of ​​the target object are determined based on the classification results; An alarm is triggered when the security requirements are determined to be met.

7. The method according to claim 6, characterized in that, Also includes: The current location information of the target object is determined based on the radar data from the radar sensor and the radar data of the target object from at least one other radar sensor.

8. The method according to claim 7, characterized in that, The radar data from the radar sensor includes a first time of arrival (ToA) information of the target object, and the radar data from the at least one other radar sensor includes a second time of arrival (ToA) information of the target object; The determination of the current location information of the target object includes: The second arrival time ToA information of the target object is obtained from the at least one other radar sensor; The current location information of the target object is determined based on the first ToA information, the second ToA information, and the location information of the at least one other radar sensor.

9. The method according to claim 8, characterized in that, Also includes: Based on the historical location information of the target object, obtain the predicted location information of the target object; Determining the current location information of the target object based on the first ToA information, the second ToA information, and the location information of the at least one other radar sensor includes: The current position information of the target object is determined based on the first ToA information, the second ToA information, the position information of the at least one other radar sensor, and the predicted position information of the target object.

10. The method according to claim 9, characterized in that, The step of determining whether the target object is within the preset area based on the current location information of the target object includes: The movement pattern of the target object is determined based on the historical location information and the current location information of the target object, wherein the movement pattern includes the movement speed and movement direction of the target object; Based on the movement pattern of the target object and the range of the preset area, it is determined whether the target object is within the preset area.

11. The method according to any one of claims 6 to 9, characterized in that, The step of determining whether the target object is within the preset area based on the current location information of the target object includes: The current location information of the target object and the range of the preset area are input into the second pre-trained model to obtain the result of determining whether the target object is within the preset area. The second pre-trained model is trained using predefined security events and danger events related to the preset area.

12. The method according to any one of claims 1 to 11, characterized in that, The classification result includes the category to which the target object belongs.

13. The method according to claim 12, characterized in that, The classification result also includes the confidence level corresponding to the classification.

14. A data processing apparatus, characterized in that, include: An acquisition module is used to acquire a classification result of a target object based on radar data from a radar sensor, wherein the radar data includes the movement characteristics of at least one moving object detected by the radar sensor, and the at least one moving object includes the target object; The processing module is used to determine the alarm strategy for the target object based on the classification result and the current location information of the target object, wherein the current location information is derived from the radar data from the radar sensor.

15. The apparatus according to claim 14, characterized in that, The mobility characteristics of the at least one moving object include the distance-time characteristics of the at least one moving object in the time domain or the distance-Doppler characteristics in the frequency domain.

16. The apparatus according to claim 15, characterized in that, The processing module is used for: Based on the movement characteristics of the at least one moving object, derive the distance-time characteristics or distance-Doppler characteristics of the target object; The classification result of the target object is obtained based on the distance-time characteristics or distance-Doppler characteristics of the target object.

17. The apparatus according to claim 16, characterized in that, The processing module is used for: The time-domain or frequency-domain characteristics of the target object are input into a first pre-trained model to obtain the classification result, wherein the first pre-trained model is trained using the distance-time characteristics or distance-Doppler characteristics of objects belonging to different categories.

18. The apparatus according to claim 16, characterized in that, The processing module is used for: The classification characteristics of the target object are determined based on the range-time characteristics or range-Doppler characteristics of the target object, wherein the classification characteristics include at least one of the radar cross-section (RCS) or breathing rate of the target object; The classification result is obtained based on the classification characteristics of the target object.

19. The apparatus according to any one of claims 14 to 18, characterized in that, The processing module is used for: Based on the current location information of the target object, determine whether the target object is within the preset area; When the target object is determined to be within the preset area, the security requirements corresponding to the preset area of ​​the target object are determined based on the classification results; An alarm is triggered when the security requirements are determined to be met.

20. The apparatus according to claim 19, characterized in that, The processing module is used for: The current location information of the target object is determined based on the radar data from the radar sensor and the radar data of the target object from at least one other radar sensor.

21. The apparatus according to claim 20, characterized in that, The radar data from the radar sensor includes a first time of arrival (ToA) information of the target object, and the radar data from the at least one other radar sensor includes a second time of arrival (ToA) information of the target object; The acquisition module is used to acquire the second arrival time ToA information of the target object from the at least one other radar sensor; The processing module is used to determine the current location information of the target object based on the first ToA information, the second ToA information, and the location information of the at least one other radar sensor.

22. The apparatus according to claim 21, characterized in that, The acquisition module is used for: Based on the historical location information of the target object, obtain the predicted location information of the target object; The processing module is used for: The current position information of the target object is determined based on the first ToA information, the second ToA information, the position information of the at least one other radar sensor, and the predicted position information of the target object.

23. The apparatus according to claim 22, characterized in that, The processing module is used for: The movement pattern of the target object is determined based on the historical location information and the current location information of the target object, wherein the movement pattern includes the movement speed and movement direction of the target object; Based on the movement pattern of the target object and the range of the preset area, it is determined whether the target object is within the preset area.

24. The apparatus according to any one of claims 19 to 22, characterized in that, The processing module is used for: The current location information of the target object and the range of the preset area are input into the second pre-trained model to obtain the result of determining whether the target object is within the preset area. The second pre-trained model is trained using predefined security events and danger events related to the preset area.

25. The apparatus according to any one of claims 14 to 24, characterized in that, The classification result includes the category to which the target object belongs.

26. The apparatus according to claim 25, characterized in that, The classification result also includes the confidence level corresponding to the classification.

27. A data processing device, characterized in that, Includes processing circuitry for performing the method according to any one of claims 1 to 13.

28. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed by a processor, implement the method according to any one of claims 1 to 13.

Images