An AC combined with AP architecture-based people flow counting method
By generating terminal IDs in the AC+AP architecture and using active time windows and statistical areas for joint determination, the accuracy and privacy issues of pedestrian flow statistics in existing technologies are solved, and accurate statistics of pedestrian flow and cross-AP trajectory fusion are realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU V-SOLUTION TELECOMM TECH CO LTD
- Filing Date
- 2026-01-09
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies for pedestrian flow statistics suffer from low accuracy, inability to distinguish user behavior, insufficient privacy protection, inability to identify non-associated terminals, and insufficient cross-AP trajectory fusion capabilities. In particular, in AC+AP centralized Wi-Fi networks, non-associated terminal data cannot be effectively utilized.
Terminal data is acquired through the AP, and the AC collects and preprocesses the data according to a preset cycle to generate terminal IDs. The active time window and statistical area are used for joint judgment, and user behavior is identified by combining signal strength and behavioral characteristics to achieve accurate statistics on pedestrian traffic.
It enables accurate counting of people in densely populated areas, improves statistical integrity, meets privacy protection requirements, can identify non-associated terminals and integrate cross-AP trajectories, and reduces false counts.
Smart Images
Figure CN122173755A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of pedestrian flow analysis, and in particular to a pedestrian flow statistics method based on an AC combined with AP architecture. Background Technology
[0002] In densely populated places such as large shopping malls, transportation hubs, exhibition halls, and scenic spots, accurate acquisition of pedestrian flow data is of great significance for safety management, business operations, emergency dispatch, and space optimization. Currently, mainstream pedestrian flow counting technologies include: video analysis, which uses cameras combined with AI algorithms to identify and count people, but suffers from problems such as changes in lighting, occlusion, and blind spots, and involves the collection of biometric information such as faces, which can easily lead to user privacy disputes; infrared / laser beam counters, which install sensors at entrances and exits to estimate the number of people entering and exiting by the number of times they are blocked, but cannot count the number of people staying in the area, nor can they distinguish between multiple people walking in parallel or back and forth, resulting in low accuracy; Bluetooth beacons or RFID tags, which require users to carry specific devices or turn on Bluetooth on their mobile phones, resulting in limited coverage and high deployment costs; and traditional Wi-Fi probe technology, which uses APs to passively listen to the Probe Request frames sent by the terminals and extracts MAC addresses for deduplication counting, but this method has shortcomings such as being unable to distinguish between real users and operating system background scanning (such as iOS / Android periodically sending virtual MACs), unstable signals from unassociated terminals, easy to cause missed detections or misjudgments, duplicate counting of the same user when roaming between multiple APs, inability to identify dynamic behaviors such as "entering," "leaving," and "staying," and lack of filtering mechanisms for IoT devices (such as smart cameras and POS machines), leading to inflated statistics. In existing AC+AP centralized Wi-Fi networks, although the AC can obtain complete information about the terminals associated with each AP (such as MAC, IP, RSSI, and association time), this data only covers connected users, ignoring a large number of potential users who only have Wi-Fi enabled but are not connected. Furthermore, the AC lacks the ability to effectively utilize unassociated terminals and perform cross-AP trajectory fusion. Therefore, there is an urgent need for a method for counting pedestrian traffic that requires no new hardware, is compatible with the existing AC+AP architecture, balances accuracy and privacy, and can dynamically identify human behavior.
[0003] A search of existing technical literature revealed a patent application with application number CN201910392946.X, entitled "A Method for Analyzing Pedestrian Traffic, Storage Medium, and Processor." This patent reads location information data within a specified time range from a server; performs statistical analysis on the read location information data; and predicts pedestrian traffic changes over a future period based on the statistical analysis results. However, this patent suffers from several drawbacks, including a relatively rudimentary method that fails to effectively exclude non-pedestrian terminals, does not consider detection request message data, and does not account for the movement of pedestrian terminals in different areas, thus making it impossible to accurately count pedestrian traffic. Summary of the Invention
[0004] Therefore, it is necessary to provide a people flow statistics method based on AC combined with AP architecture to address the above-mentioned technical problems and achieve accurate people flow statistics in densely populated places.
[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for counting pedestrian traffic based on an AC combined with an AP architecture, the method comprising: S1: The AP obtains terminal data by scanning or connecting to the terminal. The AC collects terminal data from all APs according to a preset period. The terminal data includes associated terminal data and non-associated terminal data. S2: Preprocess the terminal data to generate a terminal ID corresponding to each terminal, and preset an active time window according to each terminal ID. The active time window is used to detect the active status of the terminal based on the terminal data. S3: Define the coverage area of all APs and divide it into different statistical areas. The different statistical areas are used to detect the location of the terminal by scanning or connecting the terminal based on the AP. S4: Based on the active status detected by the active time window and the location detected by the different statistical areas, the terminal is jointly determined to realize the flow of people statistics.
[0006] Preferably, the preprocessing includes: The original MAC address in the terminal data is subjected to one-way hash de-identification processing to generate a terminal ID corresponding to each terminal.
[0007] Preferably, the SHA-256 algorithm combined with dynamic salt value is used to perform one-way hash desensitization processing on the original MAC address in the terminal data.
[0008] Preferably, the coverage area of all APs is defined according to the physical location or SSID configuration of the AP, and different statistical areas are obtained.
[0009] Preferably, the active time window is preset with a timeout period. The active time window detects terminal data in real time according to the current time. If terminal data is received within the timeout period, the active time window is detected as active. If no terminal data is received within the timeout period, the active time window is detected as inactive.
[0010] Preferably, if terminal data is received within the time limit, the active time window is reset.
[0011] Preferably, step S4 includes: If a new terminal is detected in all statistical areas, and the terminal's active time window is detected as active, then the pedestrian flow is incremented by one. If the disappearance of an old terminal is detected in all statistical areas, and the active time window of that terminal is detected as inactive, then the number of people is reduced by one.
[0012] Preferably, the AC identifies that the terminal is moving and switching APs in different statistical areas based on the time interval of the same terminal data reported by different APs and the trend of signal strength change in the terminal data, and determines that the flow of people remains unchanged.
[0013] Preferably, the non-associative terminal data is the data in the Probe Request frame sent by the terminal scanned by the AP, and step S4 further includes: If a terminal never initiates a DHCP request, ARP broadcast, or HTTP / HTTPS traffic during the process of sending terminal data to the AP, the AC will identify the terminal as an IoT device and determine that the number of people remains unchanged. If the AP detects that the interval between the transmission of Probe Request frames sent by the terminal is less than a preset fixed period, the AC will identify it as a background system scan and determine that the flow of people remains unchanged. If the signal strength in the terminal data fluctuates less than the preset value over a long period of time, and the location of the terminal detected in the statistical area does not change, the AC identifies the terminal as a fixed device and determines that the flow of people remains unchanged.
[0014] Preferably, after step S4, the method further includes: Based on the pedestrian flow statistics in all statistical areas, we can calculate the changes in pedestrian flow and the average user dwell time per unit time. Heatmaps are used to visualize the pedestrian flow, changes in pedestrian flow per unit time, and average user dwell time in all statistical areas.
[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a method for counting pedestrian traffic based on an AC combined with AP architecture. It adopts a centralized Wi-Fi network with AC combined with AP architecture. The AC obtains complete information of the terminals associated with each AP (such as MAC, IP, RSSI, and association time), and also obtains complete information of unassociated terminals (a large number of potential pedestrians who have only turned on Wi-Fi but are not connected). It effectively utilizes unassociated terminals, improves the completeness of statistics, and performs joint judgment on each terminal by setting active time windows and different statistical areas, so as to realize the planning of dense places and accurate counting of pedestrian traffic. Attached Figure Description
[0016] Figure 1This is a schematic diagram of a pedestrian flow statistics method based on an AC combined with AP architecture in one embodiment; Figure 2 This is a schematic diagram of the system architecture of a pedestrian flow statistics method based on an AC combined with an AP architecture in one embodiment. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0018] Example 1 like Figure 1 As shown, this embodiment proposes a method for counting pedestrian traffic based on an AC combined with an AP architecture. The method includes: S1: The AP obtains terminal data by scanning or connecting to the terminal. The AC collects terminal data from all APs according to a preset period. The terminal data includes associated terminal data and non-associated terminal data. In practice, the AC periodically (e.g., every 5-10 seconds) collects two types of data from all APs: Associated terminal data includes: de-identified terminal MAC address, IP address, signal strength (RSSI), association start time, and AP identifier; Non-associated terminal data: source MAC, timestamp, RSSI, and channel information in the Probe Request frame monitored by the AP.
[0019] S2: Preprocess the terminal data to generate a terminal ID corresponding to each terminal, and preset an active time window according to each terminal ID. The active time window is used to detect the active status of the terminal based on the terminal data. In practice, after preprocessing, an active time window is maintained for each de-identified terminal ID. The active time window detects the active status of the terminal based on new data (whether associated or probed).
[0020] S3: Define the coverage area of all APs and divide it into different statistical areas. The different statistical areas are used to detect the location of the terminal by scanning or connecting the terminal based on the AP. In practice, the statistical area is defined as a logical set of AP coverage areas. For example, 1F-3F is defined as the statistical area, and the corresponding AP group ID is "FLOOR1-3".
[0021] S4: Based on the active status detected by the active time window and the location detected by the different statistical areas, the terminal is jointly determined to realize the flow of people statistics.
[0022] In practice, based on the results of joint judgment, the current number of people in the area is output in real time to realize the statistics of people flow.
[0023] Example 2 This embodiment further explains the pedestrian flow statistics method based on AC combined with AP architecture proposed in Embodiment 1.
[0024] Preprocessing includes: The SHA-256 algorithm combined with dynamic salt value is used to perform one-way hash desensitization processing on the original MAC address in the terminal data.
[0025] In practice, all original MAC addresses are immediately subjected to one-way hash de-identification processing on the AC side. The SHA-256 algorithm combined with dynamic salt value is used to generate irreversible anonymous IDs. The end-to-end de-identification of MAC addresses ensures that the original device identifier cannot be restored, thus meeting privacy compliance requirements.
[0026] The coverage area of all APs is defined based on their physical location or SSID configuration, resulting in different statistical zones.
[0027] The active time window is preset with a timeout period. The active time window detects terminal data in real time according to the current time. If terminal data is received within the timeout period, the active time window is detected as active. If no terminal data is received within the timeout period, the active time window is detected as inactive.
[0028] In specific implementation, this embodiment presets the active time window to a default time of 5 minutes. When new data is received from the terminal (whether associated or probed), the active time window is detected as active, the terminal is marked as "entering", and the active time window is reset. If no new data is received after the window expires, the active time window is detected as inactive, and the terminal is marked as "leaving".
[0029] Step S4 includes: If a new terminal is detected in all statistical areas, and the terminal's active time window is detected as active, then the pedestrian flow is incremented by one. If the disappearance of an old terminal is detected in all statistical areas, and the active time window of that terminal is detected as inactive, then the number of people is reduced by one.
[0030] In practice, when a de-identified terminal first appears in the data of any regional AP and its active window is in an "active" state, it is determined as an "entry event" and the current number of people is incremented by 1; when the terminal disappears from all regional APs and the active window times out, it is determined as an "exit event" and the current number of people is decremented by 1.
[0031] Based on the time intervals between the same terminal data reported by different APs and the trend of signal strength changes in the terminal data, the AC identifies that the terminal is moving in different statistical areas and switching APs, and determines that the flow of people remains unchanged.
[0032] In practice, if a terminal switches between multiple APs in an area, the AC makes a judgment based on the continuity of time (adjacent reporting interval) and the trend of signal strength (RSSI) change (signal transition trend). When the adjacent reporting interval is less than 30 seconds and the signal transitions smoothly, it is judged that the same user is moving and merged into a single trajectory to avoid duplicate counting. That is, by combining the active window with cross-AP trajectory fusion, the duplicate counting problem in roaming scenarios can be effectively solved.
[0033] Example 3 This embodiment further explains the pedestrian flow statistics method based on AC combined with AP architecture proposed in Embodiments 1 and 2.
[0034] The non-associative terminal data is the data in the Probe Request frame sent by the terminal and scanned by the AP. Step S4 also includes: If a terminal never initiates a DHCP request, ARP broadcast, or HTTP / HTTPS traffic during the process of sending terminal data to the AP, the AC will identify the terminal as an IoT device and determine that the number of people remains unchanged. If the AP detects that the interval between the transmission of Probe Request frames sent by the terminal is less than a preset fixed period, the AC will identify it as a background system scan and determine that the flow of people remains unchanged. If the signal strength in the terminal data fluctuates less than the preset value over a long period of time, and the location of the terminal detected in the statistical area does not change, the AC identifies the terminal as a fixed device and determines that the flow of people remains unchanged.
[0035] In practice, step S4 also introduces multi-dimensional behavioral feature rules to exclude non-human-flow terminals. The rules include: Network interaction missing: If the terminal never initiates DHCP requests, ARP broadcasts, or HTTP / HTTPS traffic, then the terminal is determined to be an IoT device; Probe behavior regularity: If the interval between Probe Requests is fixed (standard deviation < 1 second), which matches the characteristics of the system's background scanning, then filter it; Abnormal signal stability: If the signal strength (RSSI) fluctuates less than 3dBm over a long period of time and there is no change in location, it is determined to be a fixed device.
[0036] Based on the above rules, non-human-flow terminals are filtered out, and only terminals that meet the characteristics of "randomness of human movement" are retained for statistical purposes, which significantly reduces the inflated statistical figures.
[0037] After step S4, the following also includes: Based on the pedestrian flow statistics in all statistical areas, we can calculate the changes in pedestrian flow and the average user dwell time per unit time. Heatmaps are used to visualize the pedestrian flow, changes in pedestrian flow per unit time, and average user dwell time in all statistical areas.
[0038] In practice, based on the results of the joint determination, the system also outputs the number of people entering and leaving the area per unit time (e.g., per hour); the average dwell time of users (average dwell time = Σ(departure time - entry time) / total number of people entering); and finally aggregates terminal density according to AP geographical location and uses a heat map to visualize the data.
[0039] System architecture such as Figure 2 As shown, this architecture is used to implement foot traffic statistics in a large shopping mall. In terms of system deployment, this embodiment deploys 86 Wi-Fi 6 APs in the mall, which are managed by one enterprise-level AC. The 1F-3F area is defined as the statistical area, and the corresponding AP group ID is "FLOOR1-3". The AC is configured with a data collection cycle of 8 seconds and an active time window of 5 minutes.
[0040] In terms of data collection and anonymization, the AC polls each AP every 8 seconds to obtain the terminal STA list and Probe logs; and performs sta_id = SHA256(MAC + "salt") on all MAC addresses to generate a 64-bit anonymous ID.
[0041] In terms of active time window management, a 5-minute active time window is maintained for each sta_id; the terminal STAN sends a Probe every 30 seconds, and the window is continuously refreshed; after STAN leaves the mall, the window times out, triggering a "departure event".
[0042] Regarding cross-AP trajectory merging, when a user moves from 1F (AP101) to 2F (AP201), the AC detects a time interval of 12 seconds. The signal strength (RSSI) of AP101 drops from -65dBm to -85dBm, while the signal strength (RSSI) of AP201 rises from -80dBm to -60dBm. This is determined to be the same user moving, and no new count is added.
[0043] Regarding the filtering of non-human devices, a POS machine that sends a probe every 60 seconds with a standard deviation of 0.1 seconds is filtered out; a smart camera that does not make a DHCP request is also filtered out.
[0044] Finally, based on the active time window management and cross-AP trajectory merging, the traffic flow was calculated. The AC summarized the data to the upper-level application and displayed on a large screen that the current traffic flow was 1248 people, the cumulative number of visitors today was 28765 people, and the average stay time was 42 minutes. The heat map showed that the children's area (defined as AP301-304) had the highest density.
Claims
1. A method for counting pedestrian traffic based on an AC combined with an AP architecture, characterized in that, include: S1: The AP obtains terminal data by scanning or connecting to the terminal. The AC collects terminal data from all APs according to a preset period. The terminal data includes associated terminal data and non-associated terminal data. S2: Preprocess the terminal data to generate a terminal ID corresponding to each terminal, and preset an active time window according to each terminal ID. The active time window is used to detect the active status of the terminal based on the terminal data. S3: Define the coverage area of all APs and divide it into different statistical areas. The different statistical areas are used to detect the location of the terminal by scanning or connecting the terminal based on the AP. S4: Based on the active status detected by the active time window and the location detected by the different statistical areas, the terminal is jointly determined to realize the flow of people statistics.
2. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 1, characterized in that, The preprocessing includes: The original MAC address in the terminal data is subjected to one-way hash de-identification processing to generate a terminal ID corresponding to each terminal.
3. The pedestrian flow statistics method based on an AC combined with AP architecture according to claim 2, characterized in that, The SHA-256 algorithm combined with dynamic salt value is used to perform one-way hash desensitization processing on the original MAC address in the terminal data.
4. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 1, characterized in that, The coverage area of all APs is defined based on their physical location or SSID configuration, resulting in different statistical zones.
5. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 1, characterized in that, The active time window is preset with a timeout period. The active time window detects terminal data in real time according to the current time. If terminal data is received within the timeout period, the active time window is detected as active. If no terminal data is received within the timeout period, the active time window is detected as inactive.
6. The pedestrian flow statistics method based on an AC combined with AP architecture according to claim 5, characterized in that, If terminal data is received within the time limit, the active time window is reset.
7. The pedestrian flow statistics method based on an AC combined with AP architecture according to claim 6, characterized in that, Step S4 includes: If a new terminal is detected in all statistical areas, and the terminal's active time window is detected as active, then the pedestrian flow is incremented by one. If the disappearance of an old terminal is detected in all statistical areas, and the active time window of that terminal is detected as inactive, then the number of people is reduced by one.
8. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 7, characterized in that, Based on the time intervals between the same terminal data reported by different APs and the trend of signal strength changes in the terminal data, the AC identifies that the terminal is moving in different statistical areas and switching APs, and determines that the flow of people remains unchanged.
9. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 1, characterized in that, The non-associative terminal data is the data in the Probe Request frame sent by the terminal and scanned by the AP. Step S4 also includes: If a terminal never initiates a DHCP request, ARP broadcast, or HTTP / HTTPS traffic during the process of sending terminal data to the AP, the AC will identify the terminal as an IoT device and determine that the number of people remains unchanged. If the AP detects that the interval between the transmission of Probe Request frames sent by the terminal is less than a preset fixed period, the AC will identify it as a background system scan and determine that the flow of people remains unchanged. If the signal strength in the terminal data fluctuates less than the preset value over a long period of time, and the location of the terminal detected in the statistical area does not change, the AC identifies the terminal as a fixed device and determines that the flow of people remains unchanged.
10. The pedestrian flow statistics method based on AC combined with AP architecture according to claim 1, characterized in that, After step S4, the following also includes: Based on the pedestrian flow statistics in all statistical areas, we can calculate the changes in pedestrian flow and the average user dwell time per unit time. Heatmaps are used to visualize the pedestrian flow, changes in pedestrian flow per unit time, and average user dwell time in all statistical areas.