Campus attendance system based on wireless transmission
By using convolutional neural networks and radial basis function networks to process Bluetooth signal data in campus attendance system, high-precision indoor positioning is solved, and the problems of high cost, difficulty in deployment and low positioning accuracy in existing systems are solved, and efficient attendance management without additional students' operations is achieved.
Patent Information
- Application Number
- CN202411761609.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-03
- Publication Date
- 2025-03-18
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing campus attendance system has problems such as high hardware costs, difficulty in deploying, low positioning accuracy and requiring additional operations by students, making it difficult to meet the diverse needs of smart campuses.
The campus attendance system based on wireless transmission is adopted to collect the Bluetooth signal strength and MAC address of the student equipment in real time, and combine the convolutional neural network and the radial basis function network to achieve high-precision indoor positioning without additional students' operations and update student position changes in real time.
It realizes low-cost, easy deployment, and high-precision positioning, can efficiently manage and record student attendance in real time, and adapt to the diverse needs of smart campuses.
Smart Images
Figure CN119649482A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of attendance management, and more specifically, to a campus attendance system based on wireless transmission. Background Art
[0002] In modern campus management, student attendance management is an important part of ensuring teaching quality and maintaining discipline. However, traditional attendance methods such as roll call, paper sign-in or manual record by teachers are not only time-consuming, but also prone to omissions or errors, which seriously affect classroom efficiency. In order to improve the level of automation of attendance, intelligent attendance systems based on technologies such as RFID, NFC, and biometrics have been introduced on the market. However, these technologies usually require high hardware costs, and additional special equipment needs to be deployed during implementation, resulting in greater difficulty in promotion and maintenance. At the same time, some systems rely on students' active operations, such as swiping cards or signing in, and still cannot completely get rid of manual intervention. In addition, in the field of indoor positioning, traditional positioning methods based on Wi-Fi or Bluetooth signal strength are limited by signal interference in complex environments, and the accuracy is difficult to meet actual needs. Based on this, it is urgent to develop a low-cost, easy-to-deploy, high-precision intelligent attendance system that does not require additional student operations, so as to achieve efficient management and real-time recording of student attendance and adapt to the diverse needs of smart campus development. Summary of the invention
[0003] In order to overcome the above-mentioned defects of the prior art, the present invention provides a campus attendance system based on wireless transmission, which realizes high-precision indoor positioning by real-time collection of Bluetooth signal strength and MAC address of student devices, combining neural network and radial basis function network, without the need for additional operation by students, and real-time update of student position changes and timely feedback, thereby efficiently managing and real-time recording of student attendance, so as to solve the problems raised in the above-mentioned background technology.
[0004] To achieve the above object, the present invention provides the following technical solutions: A campus attendance system based on wireless transmission includes a Bluetooth signal acquisition module, an indoor positioning module, a data processing module and an attendance management module; the Bluetooth signal acquisition module collects the Bluetooth signal strength (RSSI) of student devices and the MAC address of the device in real time; the indoor positioning module uses a convolutional neural network to convert the captured RSSI signal into a feature vector, and uses a radial basis function network (RBF) to map the feature vector to an indoor two-dimensional coordinate system; the indoor positioning module processes, stores and manages the acquired data; the attendance management module integrates and processes the attendance data generated by the positioning module; The specific steps of mapping the feature vector to the indoor two-dimensional coordinate system using the radial basis function network (RBF) include: Step Y1, the feature vector extracted by the convolutional neural network As input to the radial basis function network; Step Y2, the radial basis function network calculates the distance between the input feature and each center point through a set of radial basis functions, and outputs an activation value. The calculation formula of the activation value is: ,in, For the The activation values of the radial basis functions; For the The center point of the radial basis function network represents the reference point in the feature space learned during the training phase; is the input feature vector With center point The Euclidean distance of For the The width of the radial basis function controls the activation range; the radial basis function network calculates the activation values of all center points and generates an activation vector: ,in, is the number of center points of the radial basis function network; Step Y3: Compare the activation value with the corresponding weight and Perform linear combinations to generate two-dimensional coordinates and , the calculation formula is: , ,in, and The output two-dimensional coordinates represent the actual location of the student device in the room; To correspond to The weight of To correspond to The weight of and is a bias term; the radial basis function network maps the input feature vector to the actual indoor coordinate system by learning weights and bias terms.
[0005] As a further solution of the present invention, the RSSI signal extracts local features through a two-dimensional convolution operation, and the convolution operation uses a set of trainable convolution kernels w. The upper sliding window calculates the weighted sum to generate the feature map , the calculation formula is: ,in, It is The convolution kernels are located at The generated output is, is the size of the convolution kernel, is a nonlinear activation function; the convolutional feature map is then reduced in dimension and compressed through pooling operations, retaining significant features and reducing the size of the feature map, thereby removing noise and unimportant signal fluctuations, and expanding the feature map generated by multi-layer convolution and pooling into a one-dimensional vector, expressed as a feature vector: ,in, is a high-dimensional feature vector, is the dimension of the feature vector.
[0006] As a further solution of the present invention, the Bluetooth signal acquisition module collects the Bluetooth signal strength (RSSI) of the student device and the MAC address of the device in real time, including the following specific contents: the core hardware of the Bluetooth signal acquisition module includes multiple Bluetooth signal collectors that support Bluetooth low energy (BLE) technology, and the Bluetooth signal collector is a Raspberry Pi equipped with a Bluetooth USB adapter. The Bluetooth signal collectors are installed in different locations in the classroom to ensure full coverage of the signal. Each Bluetooth signal collector works in Bluetooth discovery mode and can scan the surrounding Bluetooth devices without establishing a connection with the device. The signal acquisition range and efficiency of the Bluetooth signal acquisition module depend on the power, sensitivity and layout of the Bluetooth hardware. The propagation of Bluetooth signals will be interfered by tables, chairs, human bodies and other obstacles. Therefore, the Bluetooth signal collectors are arranged in the four corners of the classroom and installed at a height of more than 2 meters above the ground to avoid interference with the signal strength caused by ground reflection and low-level occlusion problems.
[0007] In order to enhance the robustness of data collection, a multi-point redundancy strategy is adopted, that is, each student device will be captured by multiple Bluetooth signal collectors at the same time. In addition, in order to avoid excessive consumption of the battery of the Bluetooth signal collector, the Bluetooth signal collection module uses the low power consumption characteristics of BLE technology. The scanning frequency and interval of the Bluetooth signal collector can be dynamically adjusted as needed, such as increasing the collection frequency at the beginning and end of the class, and reducing the scanning density at other times to save energy.
[0008] As a further solution of the present invention, the indoor positioning module is used to process Bluetooth signal data and calculate the indoor position of the student, including the following specific contents: the collection of RSSI signals is completed by the Bluetooth signal collection module, including data of multiple Bluetooth signal source points, the characteristics of the data are that they are easily affected by obstacles, signal interference and Bluetooth device performance in the classroom environment, therefore, the signal needs to be preprocessed, and the stability of the data is enhanced by a weighted moving average filtering algorithm, and the formula of the weighted moving average filtering algorithm is: ; Where: is the signal value after filtering, indicating that at time The smoothed value obtained after filtering the signal; Indicates at time The collected signal value; is the window size, which indicates the number of signals used to calculate the weighted average; is the weighting coefficient, which indicates the weight of a given signal value.
[0009] As a deep learning model, convolutional neural network can capture the changing pattern of RSSI signals at different spatial locations. RSSI data is mapped into a one-dimensional feature vector and input into a multi-layer convolutional network. The convolution kernel in the network extracts local features of the signal through a sliding window and generates a feature map. RSSI signals extract local features through a two-dimensional convolution operation. The convolution operation uses a set of trainable convolution kernels w. The upper sliding window calculates the weighted sum to generate the feature map , the calculation formula is: ,in, It is The convolution kernels are located at The generated output is, is the size of the convolution kernel, is a nonlinear activation function; the convolutional feature map is then reduced in dimension and compressed through pooling operations, retaining significant features and reducing the size of the feature map, thereby removing noise and unimportant signal fluctuations, and expanding the feature map generated by multi-layer convolution and pooling into a one-dimensional vector, expressed as a feature vector: ,in, is a high-dimensional feature vector, is the dimension of the feature vector.
[0010] As a further solution of the present invention, the data processing module processes, stores and manages the data obtained by the Bluetooth signal acquisition module and the indoor positioning module, including the following specific contents: the data processing module receives the original data from the Bluetooth signal acquisition module, the original data includes the MAC address, RSSI signal, timestamp and device ID of the student device, and associates the RSSI signal with the feature extraction result of the indoor positioning module to generate a data structure containing the positioning coordinates. , signal characteristics and timestamp, and then store the data packet in a background database, which uses a relational database MySQL.
[0011] The data tables in the database include: device information table, signal data table, positioning result table and attendance record table. The device information table stores the device MAC address and device type; the signal data table stores the original signal data such as RSSI value, Bluetooth beacon ID and timestamp; the positioning result table stores the attendance record of the student device. The data processing module dynamically generates attendance records based on real-time data according to the needs of the teacher. For example, when a student enters a classroom and the device is successfully located, it will automatically determine whether it is in the predefined attendance area. If the conditions are met, the data processing module will add a new record to the attendance record table, indicating the student's arrival time. Similarly, when a student leaves the classroom, the module will update his or her departure time.
[0012] The data processing module adopts a standardized interface design to support the integration and expansion of multiple functions, including real-time query of student attendance of a course, query and export of historical attendance records, reminder of abnormal situations (such as lateness, early departure or absence), etc. For example, a teacher can send a query request through the Web interface, and the API interface will retrieve relevant records from the database according to conditions such as time and course number, and return them to the teacher in a table or graphical form.
[0013] The data processing module also supports real-time update function, which can quickly respond to the dynamic location changes of students. After the indoor positioning module outputs new positioning coordinates each time, the data processing module will compare the new coordinates with the previous position. If the current position of the student changes significantly (such as leaving the attendance area), the data processing module will trigger the corresponding update operation to modify the student's attendance status. For example, when a student is marked as "out of class" due to leaving the classroom, the system can immediately send a notification to the teacher to prompt the abnormal situation.
[0014] As a further solution of the present invention, the attendance management module integrates and processes the attendance data generated by the positioning module and converts it into an attendance report to provide a convenient management tool for teachers, including the following specific contents: the attendance management module receives and updates the student device location data transmitted from the data processing module in real time, and determines whether the student is in the valid area and the time of arrival and departure of the class by setting the attendance area. Based on these data, a detailed classroom attendance record is generated, including basic information such as the arrival time, departure time, and length of stay of each student. For ease of management, the attendance records are dynamically stored in the database and indexed by student ID or MAC address.
[0015] The attendance management module compares the student's attendance rate with the preset threshold, and sends a reminder to the teacher when the attendance rate is less than 80%. In addition, it also supports multi-dimensional statistical functions, including generating comprehensive attendance reports by week, month or semester, to help teachers analyze students' attendance trends from a more macro perspective. For example, the weekly report can show the total number of class attendances, late arrivals, early departures and absences for each student, while the semester report will provide more comprehensive statistical results, such as total attendance rate and abnormal behavior summary.
[0016] The invention discloses a campus attendance system based on wireless transmission, and has the following technical effects and advantages: the invention collects the Bluetooth signal strength and MAC address of student equipment in real time through a Bluetooth signal acquisition module, processes data using a weighted moving average filtering algorithm and a convolutional neural network, and realizes high-precision indoor positioning through a radial basis function network; the data processing module can effectively process, store and manage data, support integrated expansion of multiple functions, update student position changes in real time and provide timely feedback; the attendance management module can generate detailed attendance records and reports, perform multi-dimensional statistical analysis of attendance, compare attendance rates and remind teachers, and the system realizes low cost, easy deployment and high-precision positioning, does not require additional operations by students, and can efficiently manage and record student attendance in real time. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 This is a schematic diagram of indoor positioning of a Bluetooth signal collector supporting Bluetooth low energy technology according to the present invention.
[0018] Figure 2 The present invention is a schematic structural diagram of a campus attendance system based on wireless transmission.
[0019] Figure 3 A schematic diagram of student attendance data generated by the attendance management module of the present invention. DETAILED DESCRIPTION
[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0021] Example 1 See also Figure 1-3As shown in the schematic diagram, an embodiment of the present invention provides a campus attendance system based on wireless transmission, which includes a Bluetooth signal acquisition module, an indoor positioning module, a data processing module and an attendance management module; the Bluetooth signal acquisition module collects the Bluetooth signal strength (RSSI) and device MAC address of the student's device in real time; the indoor positioning module is used to process the Bluetooth signal data and calculate the indoor position of the student; the data processing module processes, stores and manages the data obtained by the Bluetooth signal acquisition module and the indoor positioning module; the attendance management module integrates and processes the attendance data generated by the positioning module, and converts it into an attendance report, providing a convenient management tool for teachers.
[0022] Furthermore, the Bluetooth signal acquisition module collects the Bluetooth signal strength (RSSI) of the student device and the MAC address of the device in real time, including the following specific contents: the core hardware of the Bluetooth signal acquisition module includes multiple Bluetooth signal collectors that support Bluetooth low energy (BLE) technology, and the Bluetooth signal collector is a Raspberry Pi equipped with a Bluetooth USB adapter. The Bluetooth signal collectors are installed in different locations in the classroom to ensure full coverage of the signal. Each Bluetooth signal collector operates in Bluetooth discovery mode and can scan surrounding Bluetooth devices without establishing a connection with the device. The signal acquisition range and efficiency of the Bluetooth signal acquisition module depend on the power, sensitivity and device layout of the Bluetooth hardware. The propagation of Bluetooth signals will be interfered with by tables, chairs, human bodies and other obstacles. See Figure 1 As shown in the schematic diagram, the Bluetooth signal collectors are arranged in the four corners of the classroom and installed at a height of more than 2 meters above the ground to avoid interference with signal strength caused by ground reflection and low-level shielding. In addition, the Bluetooth signal acquisition module runs a dedicated software program and develops a Bluetooth scanning script based on the Python language. The Bluetooth scanning script can continuously poll the MAC address and RSSI value of nearby devices, and transmit the collected signal data in JSON format via Wi-Fi or LAN to the server module for storage and processing. The following is a Python language code example: import bluetooth import json import requests import time SERVER_URL = "http: / / <your-server-address> / api / bluetooth-data" def scan_bluetooth_devices(): print("Scanning for Bluetooth devices...") devices = bluetooth.discover_devices(duration=8, lookup_names=True,lookup_class=True, device_id=-1, flush_cache=True, lookup_oui=True) device_data_list = [] for addr, name, device_class in devices: rssi = get_rssi(addr) device_data = { "mac_address": addr, "device_name": name, "device_class": device_class, "rssi": rssi, "timestamp": int(time.time()) } device_data_list.append(device_data) return device_data_list def get_rssi(mac_address): import random return random.randint(-100, -30) def send_data_to_server(data): try: headers = {'Content-Type': 'application / json'} response = requests.post(SERVER_URL, json=data, headers=headers) if response.status_code == 200: print("Data successfully sent to server.") else: print(f"Failed to send data. Status code: {response.status_code}") except Exception as e: print(f"Error sending data to server: {e}") def main(): while True: try: device_data_list = scan_bluetooth_devices() if device_data_list: print(f"Discovered {len(device_data_list)} devices.") send_data_to_server({"devices": device_data_list}) else: print("No devices found.") time.sleep(10) except KeyboardInterrupt: print("Stopping Bluetooth scanner.") break except Exception as e: print(f"An error occurred: {e}") time.sleep(5) if __name__ == "__main__": main() In order to enhance the robustness of data collection, a multi-point redundancy strategy is adopted, that is, each student device will be captured by multiple Bluetooth signal collectors at the same time. In addition, to avoid excessive consumption of the battery of the Bluetooth signal collector, the energy management unit in the Bluetooth signal collection module uses the low power consumption characteristics of BLE technology. The scanning frequency and interval of the Bluetooth signal collector can be dynamically adjusted as needed, such as increasing the collection frequency at the beginning and end of the class, and reducing the scanning density at other times to save energy.
[0023] Furthermore, the indoor positioning module is used to process Bluetooth signal data and calculate the indoor position of the student, including the following specific contents: the collection of RSSI signals is completed by the Bluetooth signal collection module, including data of multiple Bluetooth signal source points, the characteristics of which are susceptible to obstacles, signal interference and Bluetooth device performance in the classroom environment. Therefore, the data preprocessing unit in the Bluetooth signal collection module preprocesses the signal and enhances the stability of the data through a weighted moving average filtering algorithm. The formula of the weighted moving average filtering algorithm is: ; Where: is the signal value after filtering, indicating that at time The smoothed value obtained after filtering the signal; Indicates at time The collected signal value; is the window size, which indicates the number of signals used to calculate the weighted average; is the weighting coefficient, which indicates the weight of a given signal value.
[0024] As a deep learning model, convolutional neural network can capture the changing pattern of RSSI signals at different spatial locations. RSSI data is mapped into a one-dimensional feature vector and input into a multi-layer convolutional network. The convolution kernel in the network extracts local features of the signal through a sliding window and generates a feature map. RSSI signals extract local features through a two-dimensional convolution operation. The convolution operation uses a set of trainable convolution kernels w. The upper sliding window calculates the weighted sum to generate the feature map , the calculation formula is: ,in, It is The convolution kernels are located at The generated output is, is the size of the convolution kernel, is a nonlinear activation function; the convolutional feature map is then reduced in dimension and compressed through pooling operations, retaining significant features and reducing the size of the feature map, thereby removing noise and unimportant signal fluctuations, and expanding the feature map generated by multiple layers of convolution and pooling layers into a one-dimensional vector, expressed as a feature vector: ,in, is a high-dimensional feature vector, is the dimension of the feature vector.
[0025] After the convolutional neural network extracts features, the positioning module further uses a radial basis function network to map the feature vector to an indoor two-dimensional coordinate system. The specific steps include: Step Y1, the feature vector extracted by the convolutional neural network As input to the radial basis function network; Step Y2, the radial basis function network calculates the distance between the input feature and each center point through a set of radial basis functions, and outputs an activation value. The calculation formula of the activation value is: ,in, For the The activation values of the radial basis functions; For the The center point of the radial basis function network represents the reference point in the feature space learned during the training phase; is the input feature vector With center point The Euclidean distance of For the The width of the radial basis function controls the activation range; the radial basis function network calculates the activation values of all center points and generates an activation vector: ,in, is the number of center points of the radial basis function network; Step Y3: Compare the activation value with the corresponding weight and Perform linear combinations to generate two-dimensional coordinates and , the calculation formula is: , ,in, and The output two-dimensional coordinates represent the actual location of the student device in the room; To correspond to The weight of To correspond to The weight of and is a bias term; the radial basis function network maps the input feature vector to the actual indoor coordinate system by learning weights and bias terms.
[0026] Furthermore, the data processing module processes, stores and manages the data obtained by the Bluetooth signal acquisition module and the indoor positioning module, including the following specific contents: the data processing module receives the original data from the Bluetooth signal acquisition module, the original data includes the MAC address, RSSI signal, timestamp and device ID of the student device, and associates the RSSI signal with the feature extraction result of the indoor positioning module to generate a data structure containing the positioning coordinates. , signal characteristics and timestamp, and then store the data packet in a background database, which uses a relational database MySQL.
[0027] The data tables in the database include: device information table, signal data table, positioning result table and attendance record table. The device information table stores the device MAC address and device type; the signal data table stores the original signal data such as RSSI value, Bluetooth beacon ID and timestamp; the positioning result table stores the attendance record of the student device. The attendance record table contains information about the students’ attendance, including the time of arrival, time of departure and course number. In addition, the attendance record unit in the data processing module will dynamically generate attendance records based on real-time data according to the needs of the teacher. For example, when a student enters the classroom and the device is successfully located, it will automatically determine whether it is in the predefined attendance area. If the conditions are met, the data processing module will add a new record to the attendance record table, indicating the student’s arrival time. Similarly, when a student leaves the classroom, the attendance record unit will update the time of departure.
[0028] The data processing module adopts a standardized interface design to support the integration and expansion of multiple functions, including real-time query of student attendance of a course, query and export of historical attendance records, reminder of abnormal situations (such as lateness, early departure or absence), etc. For example, a teacher can send a query request through the Web interface, and the API interface will retrieve relevant records from the database according to conditions such as time and course number, and return them to the teacher in a table or graphical form.
[0029] The data processing module also supports real-time update function, which can quickly respond to the dynamic location changes of students. After the indoor positioning module outputs new positioning coordinates each time, the data processing module will compare the new coordinates with the previous position. If the current position of the student changes significantly (such as leaving the attendance area), the data processing module will trigger the corresponding update operation to modify the student's attendance status. For example, when a student is marked as "out of class" due to leaving the classroom, the system can immediately send a notification to the teacher to prompt the abnormal situation.
[0030] Furthermore, the attendance management module integrates and processes the attendance data generated by the positioning module and converts it into an attendance report, providing a convenient management tool for teachers, including the following specific contents: the attendance management module receives and updates the student device location data transmitted from the data processing module in real time, and determines whether the student is in the valid area and the time of arrival and departure of the class by setting the attendance area. Based on these data, the data visualization unit in the attendance management module generates a detailed class attendance record, see Figure 3 The schematic diagram shown includes basic information of each student such as arrival time, departure time, length of stay, etc. For ease of management, attendance records are dynamically stored in the database and indexed by student ID or MAC address.
[0031] The reminder notification unit in the attendance management module compares the student's attendance rate with the preset threshold. When the attendance rate is lower than 80%, a reminder is sent to the teacher. In addition, it also supports multi-dimensional statistical functions, including generating comprehensive attendance reports by week, month or semester, to help teachers analyze students' attendance trends from a more macro perspective. For example, the weekly report can show the total number of class attendances, late arrivals, early departures and absences for each student, while the semester report will provide more comprehensive statistical results, such as total attendance rate and abnormal behavior summary.
[0032] The present invention uses a Bluetooth signal acquisition module to collect the Bluetooth signal strength and MAC address of student devices in real time, uses a weighted moving average filtering algorithm and a convolutional neural network to process data, and uses a radial basis function network to achieve high-precision indoor positioning; the data processing module can effectively process, store and manage data, support multiple functional integration and expansion, update student position changes in real time and provide timely feedback; the attendance management module can generate detailed attendance records and reports, perform multi-dimensional statistical analysis of attendance, compare attendance rates and remind teachers. The system achieves low cost, easy deployment, and high-precision positioning, does not require additional operations by students, and can efficiently manage and record student attendance in real time.
[0033] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
[0034] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A campus attendance system based on wireless transmission, characterized in that: It includes a Bluetooth signal acquisition module, an indoor positioning module, a data processing module and an attendance management module; the Bluetooth signal acquisition module collects the RSSI of the student's device and the MAC address of the device in real time; the indoor positioning module uses a convolutional neural network to convert the captured RSSI signal into a feature vector, and uses a radial basis function network to map the feature vector to an indoor two-dimensional coordinate system; The specific steps of mapping the feature vector to the indoor two-dimensional coordinate system using the radial basis function network include: Step Y1, the feature vector extracted by the convolutional neural network As input to the radial basis function network; Step Y2, the radial basis function network calculates the distance between the input feature and each center point through a set of radial basis functions, and outputs an activation value. The calculation formula of the activation value is: ,in, For the The activation values of the radial basis functions; For the The center point of a radial basis function network; is the input feature vector With center point The Euclidean distance of For the The width of the radial basis function; the radial basis function network calculates the activation values of all center points and generates an activation vector: ,in, is the number of center points of the radial basis function network; Step Y3: Compare the activation value with the corresponding weight and Perform linear combination to generate two-dimensional coordinates and , the calculation formula is: , ,in, and The two-dimensional coordinates of the student device in the room; To correspond to The weight of To correspond to The weight of and is the bias term.
2. A campus attendance system based on wireless transmission according to claim 1, characterized in that ,The RSSI signal is extracted with local features through a two-dimensional convolution operation, which uses a set of trainable convolution kernels w, by The upper sliding window calculates the weighted sum to generate the feature map , the calculation formula is: ,in, It is The convolution kernels are located at The generated output is, is the size of the convolution kernel, is a nonlinear activation function; the convolutional feature map is then reduced in dimension and compressed through pooling operations, retaining significant features and reducing the size of the feature map, thereby removing noise and unimportant signal fluctuations, and expanding the feature map generated by multiple layers of convolution and pooling layers into a one-dimensional vector, expressed as a feature vector: ,in, is a high-dimensional feature vector, is the dimension of the feature vector.
3. A campus attendance system based on wireless transmission according to claim 1, characterized in that The indoor positioning module pre-processes the data obtained by the Bluetooth signal acquisition module and enhances the stability of the data through a weighted moving average filtering algorithm. The formula of the weighted moving average filtering algorithm is: ; Where: is the signal value after filtering, indicating that at time The smoothed value obtained after filtering the signal; Indicates at time The collected signal value; is the window size, which indicates the number of signals used to calculate the weighted average; is the weighting coefficient, which indicates the weight of a given signal value.
4. A campus attendance system based on wireless transmission according to claim 1, characterized in that: The core hardware of the Bluetooth signal acquisition module includes a Bluetooth signal collector based on BLE technology. The Bluetooth signal collector is a Raspberry Pi equipped with a Bluetooth USB adapter. The Bluetooth signal collector is arranged in the four corners of the classroom and installed at a height of more than 2 meters from the ground.
5. The campus attendance system based on wireless transmission according to claim 1 is characterized in that ,The Bluetooth signal acquisition module adopts a multi-point redundancy strategy, and each student device will ,capture signals by multiple Bluetooth signal collectors at the same time.
6. A campus attendance system based on wireless transmission according to claim 1, characterized in that ,The data processing module receives the raw data from the Bluetooth signal ,acquisition module, associates the RSSI signal with the feature extraction ,results of the indoor positioning module, generates a data packet and stores it in the ,backend database, which uses the relational database MySQL.
7. A campus attendance system based on wireless transmission according to claim 1, characterized in that ,The data tables in the database include equipment information table, signal data table, ,positioning result table and attendance record table.
8. The campus attendance system based on wireless transmission according to claim 1 is characterized in that ,The data processing module adopts a standardized interface design and ,supports multiple functional integration and expansion, including real-time ,query, historical attendance record query and export and ,abnormal situation reminder.
9. A campus attendance system based on wireless transmission according to claim 1, characterized in that ,The data processing module supports real-time update function, and can respond to ,students’ dynamic location changes, modify students’ attendance ,status and notify the teacher end.
10. The campus attendance system based on wireless transmission according to claim 1 is characterized in that ,The attendance management module generates detailed classroom attendance records, ,comparing attendance rates, and generating comprehensive attendance reports by setting attendance ,areas.