Method for linking a pet collar to a feeder

Abstract

This application relates to the field of IoT smart devices and provides a method for linking a pet collar and a feeder. The method includes: controlling the pet collar as a Bluetooth beacon to broadcast data packets with a preset transmission power and preset period; controlling the feeder to initiate Bluetooth passive scanning, listening to the broadcast channel according to preset scanning parameters, and performing format filtering and signal strength value threshold filtering after receiving data packets; storing the media access control address, signal strength value, and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue after the format filtering and threshold filtering are passed; controlling the feeder to perform verification at a preset period, clearing historical data exceeding a preset valid duration in the circular buffer queue; when all data corresponding to the pet collar in the circular buffer queue is cleared, determining that the pet has left the area, reporting a record containing the corresponding media access control address, current UTC, and cumulative presence time, and triggering a preset linkage action.

Description

Technical Field

[0001] This application relates to the field of Internet of Things (IoT) smart device technology, and in particular to a method for linking a pet collar with a feeder. Background Technology

[0002] With the rapid development of smart homes and the pet economy, the penetration rate of terminal devices such as smart pet collars and smart feeders continues to increase, and users' core needs for accurate pet behavior statistics and intelligent linkage of multiple devices are becoming increasingly urgent.

[0003] Currently, smart pet collars and feeders generally operate as independent single devices, and can only perform basic functions such as location tracking, movement data statistics, and timed and quantitative feeding, respectively. They cannot achieve automated and highly accurate device linkage based on the pet's real-time spatial location. Summary of the Invention

[0004] This application provides a method for linking a pet collar and a feeder, aiming to solve the problem that in the current technology, smart pet collars and feeders generally operate as independent single devices, which can only realize basic functions such as positioning and tracking, movement data statistics, and timed and quantitative feeding, and cannot achieve automated and highly accurate device linkage based on the pet's real-time spatial location.

[0005] In a first aspect, embodiments of this application provide a method for linking a pet collar and a feeder, the method comprising: Control the pet collar to act as a Bluetooth beacon, broadcasting data packets containing field information, identification information, and service information at a preset transmission power and preset period; The feeder is controlled to start Bluetooth passive scanning, listen to the broadcast channel according to the preset scanning parameters, and perform format filtering and signal strength value threshold filtering after receiving data packets. After the format filtering and threshold filtering are passed, the media access control address, signal strength value and timestamp of the corresponding pet collar are stored in the local fixed-capacity circular buffer queue. The feeder is controlled to perform verification according to a preset cycle, clearing historical data in the circular buffer queue that exceeds a preset valid duration. If the total number of samples within the valid duration is not less than the first preset threshold and the number of valid samples with a signal strength value greater than the preset threshold is not less than the second preset threshold, the pet is determined to be present and the presence duration is accumulated. When all data of the corresponding pet collar in the circular buffer queue is cleared, the pet is determined to have left the scene, a record containing the corresponding media access control address, the current commute time and the accumulated presence duration is reported, and a preset linkage action is triggered.

[0006] In some embodiments, the control of the pet collar as a Bluetooth beacon broadcasts data packets containing field information, identification information, and service information at a preset transmission power and a preset period, including: controlling the pet collar to adjust the Bluetooth transmission power to a preset low-power transmission power, and broadcasting the data packets at a preset fixed broadcast period; wherein the field information, identification information, and service information in the data packets are all set with fixed check values, and data packets that match the check values ​​are subjected to format filtering and threshold filtering.

[0007] In some embodiments, controlling the feeder to initiate Bluetooth passive scanning and listening to the broadcast channel according to preset scanning parameters includes: controlling the feeder to initiate Bluetooth passive scanning mode, continuously listening to the Bluetooth broadcast channel according to preset scanning intervals and scanning windows, not actively initiating Bluetooth connection requests during the scanning process, and only receiving data packets broadcast by surrounding devices.

[0008] In some embodiments, after receiving the data packet, the sequential execution of format filtering and signal strength value threshold filtering includes: controlling the feeder to first perform format verification on the received data packet, and verifying the verification values ​​of field information, identity information and service information in the data packet; performing threshold filtering on the signal strength value corresponding to the data packet that passes the verification, and determining that the data packet with a signal strength value greater than a preset minimum threshold is filtered through.

[0009] In some embodiments, storing the media access control address, signal strength value, and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue after format filtering and threshold filtering are passed includes: controlling the feeder to extract the media access control address of the corresponding pet collar from the filtered data packets, and tracking only the data packets corresponding to one media access control address at a time; storing the new signal strength value and timestamp into the circular buffer queue in the order of reception, and deleting the oldest historical data stored according to the timestamp when the circular buffer queue is full.

[0010] In some embodiments, controlling the feeder to perform verification at a preset period and clearing historical data in the circular buffer queue that exceeds a preset valid duration includes: controlling the feeder to perform verification operations at a fixed period, obtaining the current time for each verification, comparing the difference between the timestamp of each data in the circular buffer queue and the current time, and clearing historical data whose difference exceeds a preset valid duration.

[0011] In some embodiments, determining that the pet is present and accumulating the presence time if the total number of samples within the effective duration is not less than a first preset threshold and the number of effective samples with a signal strength value greater than a preset threshold is not less than a second preset threshold includes: controlling the feeder to count the total number of samples in the circular buffer queue within the preset effective duration and the number of effective samples with a signal strength value greater than a preset threshold; when the total number of samples is not less than the first preset threshold and the number of effective samples is not less than the second preset threshold, continuously determining that the pet is present and accumulating the pet's presence time according to the verification period.

[0012] In some embodiments, determining that the pet has left the venue when all data of the corresponding pet collar in the circular buffer queue is cleared, and reporting a record containing the corresponding media access control address, the current commute time, and the cumulative on-site duration, includes: when all data of the corresponding pet collar in the circular buffer queue is cleared due to exceeding a preset valid duration, controlling the feeder to determine that the pet has left the venue, generating a record containing the media access control address of the corresponding pet collar, the commute time at the time of departure, and the cumulative complete on-site duration, and reporting the record to the bound application server and user terminal.

[0013] In some embodiments, triggering the preset linkage action includes: controlling the feeder to count the number of times the corresponding pet eats and the duration of each feeding based on the reported presence record, and synchronously updating the pet diet data module of the preset terminal; and triggering the preset operation of the corresponding linkage device according to preset rules, which includes at least one or more of the following: drinking water data statistics of the water dispenser, toilet data statistics of the litter box device, and prompting operation of the sound and light reminder device.

[0014] In some embodiments, the method further includes: controlling the feeder to continuously collect and store the historical presence data of the corresponding pet; analyzing the pet's daily activity patterns and signal strength change characteristics through machine learning algorithms; adaptively adjusting the preset signal strength threshold, the first preset threshold, and the second preset threshold; identifying non-feeding scenarios, including passing by and lingering, based on the pet's movement characteristics; filtering invalid presence determinations; and reducing the probability of false triggering.

[0015] This solution employs a two-level data filtering approach, combining format filtering and signal strength threshold filtering with a dual threshold judgment mechanism that considers both the total number of samples and the number of valid samples within a time window. This effectively filters out false judgments and missed judgments caused by environmental interference and signal fluctuations, significantly reducing the false trigger rate of linkage. It ensures that linkage and behavior statistics are triggered only when the pet actually enters the valid area, and the judgment results are highly consistent with the pet's actual state, thus solving the core pain point of insufficient judgment accuracy in existing technologies.

[0016] This solution employs a connectionless Bluetooth broadcast and passive scanning mode. The collar does not require establishing a continuous Bluetooth connection; it only broadcasts data packets at preset intervals. The feeder uses a passive scanning mode, eliminating the need for active connection initiation. Both devices achieve extremely low power consumption, significantly extending battery life. Simultaneously, through optimized broadcast and scanning parameters and a periodic verification mechanism, it can quickly complete the judgment and trigger the linkage after the pet enters the effective area, completely resolving the technical contradiction of existing technologies that cannot simultaneously achieve low power consumption and high responsiveness. The entire linkage process requires no manual Bluetooth pairing or device binding by the user. Once powered on, the device automatically completes the entire process of broadcasting, scanning, judgment, and linkage, achieving fully automated and seamless linkage, significantly reducing the user's operational threshold and improving the user experience.

[0017] The linkage logic of this solution can not only be adapted to smart feeders, but can also be directly extended to other smart pet devices such as smart water fountains and smart litter boxes, realizing accurate statistics and device linkage of pet drinking, toileting and other behaviors in all scenarios, with strong scenario adaptability and solution scalability.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic flowchart illustrating the steps of a method for linking a pet collar and a feeder according to an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a pet collar provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of a feeder provided in one embodiment of this application; Figure 4 This is a schematic diagram illustrating a method for linking a pet collar and a feeder according to this application. Figure 5 This is a schematic block diagram of the linkage system between a pet collar and a feeder provided in one embodiment of this application; Figure 6 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application.

[0021] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0024] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0025] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0027] With the rapid development of smart homes and the pet economy, the penetration rate of terminal devices such as smart pet collars and smart feeders continues to increase, and users' core needs for accurate pet behavior statistics and intelligent linkage of multiple devices are becoming increasingly urgent.

[0028] Currently, smart pet collars and feeders generally operate as independent, standalone devices, capable of only basic functions such as location tracking, movement data statistics, and timed and measured feeding. They cannot achieve automated, high-precision device linkage based on the pet's real-time spatial location. The few technological solutions that do possess linkage capabilities suffer from the following insurmountable core flaws: Firstly, the linkage relies on the traditional Bluetooth pairing connection mode, which requires users to manually complete device pairing and establish a continuous Bluetooth data link. This is not only cumbersome and results in a poor user experience, but the continuous Bluetooth connection also causes a sharp increase in power consumption at the pet collar, significantly shortening the device's battery life. This makes it completely unsuitable for wearable device scenarios like pet collars, which have extremely high battery life requirements. At the same time, the frequent movement of pets can easily cause Bluetooth connection interruptions, and the reconnection process will further increase power consumption. It can also cause problems such as linkage trigger delay and missed triggers.

[0029] Secondly, the proximity sensing and judgment logic has fundamental flaws, exhibiting poor anti-interference capabilities and a high false trigger rate. Some solutions attempt to determine distance using the received signal strength indicator value broadcast via Bluetooth, but these generally rely on a single signal strength threshold for simple judgment, failing to consider instantaneous signal fluctuations caused by environmental obstruction, signal reflection, and multipath effects. This easily leads to false triggers when a pet briefly passes by or lingers nearby, or missed detections when a pet has entered the valid area. Furthermore, existing solutions lack effective time-series management and validity screening of sampling data, and do not set reasonable time windows or sampling validity verification mechanisms. This makes it easy to use expired or invalid sampling data for judgment, resulting in judgment results that are seriously inconsistent with the pet's actual location and status, failing to meet the high reliability requirements of pet eating behavior statistics and device linkage.

[0030] Third, the core contradiction of low power consumption and fast response remains unresolved. To improve response speed, existing solutions require high-frequency scanning and continuous broadcasting, which directly leads to a significant increase in device power consumption. To reduce power consumption, the broadcasting and scanning frequencies must be reduced, which in turn increases detection latency and results in insufficient sampling data, further exacerbating the problems of misjudgment and missed judgment. It is impossible to achieve a balance between low power consumption, high response speed, and high judgment accuracy.

[0031] In summary, no existing technology can simultaneously solve the aforementioned core defects, and therefore cannot meet the actual usage needs of the industry and users for the linkage of smart pet devices.

[0032] To solve the above problem, please refer to Figures 1 to 4 This application provides a method for linking a pet collar and a feeder, applied to a computer device. The computer device can be deployed on a single server or a server cluster. It can also be deployed on a handheld terminal, laptop, wearable device, or robot, etc. It should be noted that all information involved in the method provided in this application is extracted with the authorization of the relevant user and in accordance with relevant regulations, and will not infringe on user privacy.

[0033] The provided method for linking the pet collar and feeder includes steps S101 to S103. Details are as follows: Step S101. Control the pet collar as a Bluetooth beacon to broadcast data packets containing field information, identification information, and service information at a preset transmission power and preset period.

[0034] Specifically, by optimizing Bluetooth broadcast parameters and standardizing data packet structure, the collar can operate with low power consumption while ensuring that broadcast data can be accurately identified by the feeder and filtering out interference from irrelevant Bluetooth devices.

[0035] For example, a pet collar equipped with a low-power Bluetooth SoC has a built-in 32-bit MCU, Bluetooth RF antenna, and removable lithium battery. After powering on, the MCU completes the initialization of the Bluetooth module's clock, RF, and power management, shuts down unnecessary peripherals, and enters a low-power operating mode.

[0036] The default transmit power is configured to -20dBm, which has been calibrated through real-world testing in various environments. In unobstructed conditions, the maximum receive distance for the Bluetooth broadcast signal is approximately 50 meters, but only within a range of 50cm ± 5cm does the Received Signal Strength Indication (RSSI) exhibit a stable linear correlation with the actual physical distance, enabling highly accurate distance determination. This precise range precisely covers the feeder's feeding area, effectively filtering out invalid broadcasts and misidentifications at long distances, while also considering device power consumption.

[0037] The preset broadcast cycle is configured to be a fixed 200ms. Each broadcast event sends a data packet once on each of the three broadcast channels 37, 38, and 39 of Bluetooth, ensuring that the feeder can quickly capture the broadcast signal and reduce detection latency. During the broadcast interval, the collar enters a deep sleep mode, retaining only the Bluetooth broadcast timer for wake-up. The average operating current is less than 1mA, meeting the requirements for long battery life.

[0038] The broadcast packet adopts the BLE standard AD structure, which is fixed to contain three core fields, corresponding to field information, identity information, and service information. The structure and fixed verification rules are as follows: Field Information: ADType=0x01 (Flags field), 2 bytes long, fixed value 0x06, marking the device as a general discovery mode and not supporting classic Bluetooth BR / EDR, serving as the first-level verification identifier for the data packet. Identity Information: ADType=0x03 (Complete 16-bit service UUID list), 3 bytes long, fixed UUID value 0x2333, serving as a unique identity prefix for the pet collar, distinguishing it from other Bluetooth devices such as mobile phones and headphones. Service Information: ADType=0x16 (Service Data), 21 bytes long, the first 4 bytes are fixed verification headers of 0x23, 0x33, 0xA5, 0x0A, and the subsequent bytes contain the collar's 6-byte MAC address and the pet's unique ID number, serving as the final verification of device identity and the basis for data traceability.

[0039] Step S102. Control the feeder to start Bluetooth passive scanning, listen to the broadcast channel according to the preset scanning parameters, and perform format filtering and signal strength value threshold filtering after receiving data packets. After the format filtering and threshold filtering are passed, store the media access control address, signal strength value and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue.

[0040] Specifically, low-power channel monitoring is achieved through a connectionless passive scanning mode; invalid and illegal data packets are eliminated through a two-level filtering mechanism; and time-series management of sampled data is achieved through a circular buffer queue, providing effective data support for subsequent state determination.

[0041] This step is performed by a smart feeder equipped with a Bluetooth main control chip. The feeder has a built-in 32-bit MCU, runs the FreeRTOS real-time operating system, has an onboard Bluetooth antenna, and RAM and Flash storage modules. After the device is powered on, the MCU completes the Bluetooth module initialization. Users can remotely enable / disable the Bluetooth scanning function via an app. Once the scanning function is enabled, subsequent operations are performed.

[0042] Bluetooth passive scanning is configured in pure passive scanning mode. During the scanning process, it does not actively initiate scanning requests or Bluetooth connection requests to any broadcast devices. It only listens for data packets on Bluetooth broadcast channels 37, 38, and 39, thereby reducing power consumption from the source. At the same time, the scanning parameters are configured as follows: scanning interval 0x50 (corresponding to 80ms in the Bluetooth specification) and scanning window 0x32 (corresponding to 50ms). These parameters balance power consumption and data packet capture rate, and can fully cover broadcast packets in a 200ms cycle without the risk of missed detection.

[0043] Each time the Bluetooth module receives a broadcast data packet, it immediately triggers an interrupt callback, passing the raw data packet, RSSI value, and sender MAC address to the filtering function. The filtering strictly follows the order of format filtering first, then threshold filtering. Data packets that fail the first stage are discarded without consuming system resources: First stage: Format filtering. The AD structure of the broadcast packet is parsed byte by byte, sequentially verifying the Flags field, UUID field, and service data checksum header. The fixed checksum values ​​of these three fields must match exactly; otherwise, the data packet is discarded, retaining only valid pet collar broadcast data. Second stage: Threshold filtering. After format filtering passes, the corresponding RSSI value of the data packet is read and compared with a preset minimum threshold of -58dBm. If the RSSI value is ≤-58dBm, it indicates the collar is too far away and is discarded; only data packets with an RSSI value >-58dBm are considered to have passed the filter.

[0044] In the RAM of the feeder MCU, a circular buffer queue with a fixed capacity of 10 groups is allocated. Each group of data structures includes: a 6-byte MAC address of the collar, an int32 RSSI value, a uint32 UTC timestamp, and a corresponding write index and busy flag. The write rules are as follows: Single-device tracking: Only one MAC address collar is tracked at a time. When there is no tracking device, the MAC address of filtered data packets will be locked, and the busy flag will be set to 1; when there is a tracking device, only data packets with the exact same locked MAC address will be processed, and data packets from other devices will be discarded.

[0045] Circular write: For filtered data packets, the RSSI value and the current UTC timestamp are stored in the queue according to the write index. The write index is modulo 10, and the index is incremented after the write is completed. When the queue is full, new data automatically overwrites the oldest historical data, and always retains the latest 10 sets of sampled data.

[0046] Step S103. Control the feeder to perform verification according to a preset cycle, clear historical data in the circular buffer queue that exceeds the preset valid duration. If the total number of samples within the valid duration is not less than the first preset threshold and the number of valid samples with a signal strength value greater than the preset threshold is not less than the second preset threshold, determine that the pet is present and accumulate the presence time. When all data of the corresponding pet collar in the circular buffer queue is cleared, determine that the pet has left the scene, report a record containing the corresponding media access control address, the current commute time and the accumulated presence time, and trigger a preset linkage action.

[0047] Specifically, a periodic verification mechanism is used to filter the validity of the data, a dual threshold joint judgment is used to accurately identify the presence / absence of pets, and standardized reporting and linkage actions are used to achieve pet behavior statistics and automated equipment control.

[0048] Periodic verification task creation involves creating a 1-second periodic task in the FreeRTOS system. The task has a medium priority and runs only when the scanning function is enabled. It is automatically suspended when the scanning is disabled and does not occupy system resources. The complete verification process is executed automatically every 1 second. The process sequence is: clear timeout data → status determination → duration accumulation → departure determination → reporting linkage.

[0049] The timeout data clearing execution logic first obtains the current UTC second-level timestamp after each verification task starts, then iterates through the 10 sets of data in the circular buffer queue, calculates the difference between the timestamp of each set of data and the current time, and directly clears the data set to zero and marks it as invalid data if the difference exceeds the preset valid duration of 12 seconds, ensuring that all data involved in the judgment are real-time data within the time window.

[0050] After clearing timeout data, immediately perform data statistics and judgment: Data statistics: Count the total number of valid samples that were not cleared within the 12-second time window, and the number of valid samples with an RSSI value > -58dBm. Only when the total number of samples is ≥ the first preset threshold 7 and the number of valid samples is ≥ the second preset threshold 6, is the pet determined to be present; if either condition is not met, the pet is not determined to be present, effectively filtering out misjudgments caused by signal fluctuations and momentary passing by.

[0051] When the presence is determined, the presence duration counter is incremented by 1 (corresponding to a 1-second verification cycle); when the absence is determined, the counter remains unchanged; the counter is only reset to zero after the absence determination is triggered, in preparation for the next accumulation.

[0052] The exit determination is made immediately after traversing the circular buffer queue and all 10 sets of data are cleared (there is no sampled data within the valid time window), triggering subsequent operations.

[0053] Standardized data reporting generates a fixed-format presence record containing (collar MAC address in hexadecimal string, UTC timestamp at departure, and cumulative presence time in seconds). This record is first reported to the bound IoT cloud platform, then pushed to the user's mobile app by the platform, and simultaneously stored in the feeder's local Flash memory to ensure no data loss. The app also synchronously updates the pet's dry food data module, generating bar chart statistics for the corresponding time dimension.

[0054] Based on the reported presence records, the feeder MCU executes preset linkage operations, including counting the number of times the pet eats, updating diet data, or sending linkage instructions to the bound smart water dispenser, litter box, and other devices to complete the statistics of corresponding behavioral data.

[0055] In some embodiments, the control of the pet collar as a Bluetooth beacon broadcasts data packets containing field information, identification information, and service information at a preset transmission power and a preset period, including: controlling the pet collar to adjust the Bluetooth transmission power to a preset low-power transmission power, and broadcasting the data packets at a preset fixed broadcast period; wherein the field information, identification information, and service information in the data packets are all set with fixed check values, and data packets that match the check values ​​are subjected to format filtering and threshold filtering.

[0056] To address the requirements for battery life and broadcast recognition in the collar, the Bluetooth transmission power is further limited, the broadcast cycle is fixed, and the data packet uses a three-segment fixed verification value to ensure that only legitimate pet collar data packets can enter the subsequent processing stage, while minimizing collar power consumption.

[0057] The collar uses a low-power Bluetooth SOC, paired with a 3.7V removable lithium battery. The Bluetooth RF circuit is matched with a 50Ω onboard antenna. Through impedance matching and debugging, the transmission power is ensured to be stable and without attenuation, avoiding abnormal recognition distance caused by power fluctuations.

[0058] During collar initialization, the Bluetooth transmission power is fixed at -20dBm and the broadcast period is fixed at 200ms. Each broadcast event sends data packets only on 3 broadcast channels. During the broadcast interval, the MCU enters deep sleep mode, retaining only a sleep current of less than 10μA. In actual testing, the battery life in ultra-low power mode can reach more than 30 days, which matches the battery life requirements in the requirement document.

[0059] The data packet verification rules are fixed in the collar's firmware. The Flags field of the broadcast packet is fixed to 0x06, the 16-bit UUID is fixed to 0x2333, and the service data verification header is fixed to 0xA50A. These three verification values ​​cannot be tampered with. Only data packets with all three segments completely matching will be recognized by the feeder as legitimate collar data, thus filtering out interference from other Bluetooth devices such as mobile phones and wristbands at the source.

[0060] When the collar is low on power, the broadcast cycle is automatically reduced to 500ms to further reduce power consumption. At the same time, a low power indicator is added to the service data of the broadcast packet and pushed to the APP to remind the user to charge.

[0061] In some embodiments, controlling the feeder to initiate Bluetooth passive scanning and listening to the broadcast channel according to preset scanning parameters includes: controlling the feeder to initiate Bluetooth passive scanning mode, continuously listening to the Bluetooth broadcast channel according to preset scanning intervals and scanning windows, not actively initiating Bluetooth connection requests during the scanning process, and only receiving data packets broadcast by surrounding devices.

[0062] To address the power consumption and scanning response requirements of the feeder, the feeder is limited to a purely passive scanning mode, without actively initiating Bluetooth connections, and only listening to the broadcast channel. At the same time, the scanning parameters are optimized to balance power consumption and data packet capture rate, thus solving the problems of high power consumption and easy disconnection of traditional Bluetooth connections.

[0063] When the scanning mode is initialized via the feeder's Bluetooth module, it is forcibly configured to passive scanning mode, disabling the active scanning SCAN_REQ request sending function. During the scanning process, it only listens for ADV_IND broadcast packets on the broadcast channel and does not initiate connection requests to any device, thus avoiding the extra power consumption caused by connection establishment, disconnection and reconnection.

[0064] Precise configuration of scanning parameters: By setting the scanning interval to 80ms and the scanning window to 50ms, the scanning duty cycle is 62.5%. This ensures that 100% of broadcast packets in the 200ms period are captured without any risk of missed detection, and also allows the Bluetooth module to enter a low-power mode during the 30ms interval between scans. In actual testing, the average current during scanning is less than 2mA, which is far lower than the power consumption of traditional connection modes.

[0065] The scanning start / stop control is achieved through a reserved APP remote control interface on the feeder. Users can enable / disable the proximity recognition function in the APP. When the function is disabled, the Bluetooth scanning task is immediately suspended, the Bluetooth module enters sleep mode, and only basic network functions are retained, further reducing standby power consumption.

[0066] Channel optimization prioritizes listening to the three broadcast channels 37, 38, and 39 during the scanning process, while disabling data channel listening. The data channel is only enabled when firmware upgrades are required, further reducing unnecessary power consumption.

[0067] In some embodiments, after receiving the data packet, the sequential execution of format filtering and signal strength value threshold filtering includes: controlling the feeder to first perform format verification on the received data packet, and verifying the verification values ​​of field information, identity information and service information in the data packet; performing threshold filtering on the signal strength value corresponding to the data packet that passes the verification, and determining that the data packet with a signal strength value greater than a preset minimum threshold is filtered through.

[0068] By implementing an irreversible two-stage filtering process that prioritizes format filtering followed by threshold filtering, and refining the verification criteria and execution logic for each level of filtering, we ensure that only legitimate and nearby valid data packets enter the caching stage, thereby reducing invalid data processing and improving system operating efficiency.

[0069] The irreversible filtering process is designed so that the feeder's filtering function uses a fixed execution order. The format filtering and validation must be completed first before the threshold filtering stage can begin. Data packets that fail the previous stage of filtering are directly released from memory and discarded, and do not proceed to the next stage of processing, thus avoiding invalid data consuming MCU computing power.

[0070] The byte-by-byte verification logic for format filtering includes: Step 1: Verify the broadcast packet length. It must conform to the fixed length of 26 bytes for the three AD segments. If the length does not match, the packet is discarded. Step 2: Verify the Flags segment. The ADType must be 0x01, the length must be 2 bytes, and the value must be 0x06. If any of these do not match, the packet is discarded. Step 3: Verify the UUID segment. The ADType must be 0x03, the length must be 3 bytes, and the UUID value must be 0x2333. If any of these do not match, the packet is discarded. Step 4: Verify the service data segment. The ADType must be 0x16, the length must be 21 bytes, and the first 4 bytes of the checksum header must be 0x23, 0x33, 0xA5, or 0x0A. If any of these do not match, the packet is discarded. Only if all four verification steps pass, is the format filtering considered successful.

[0071] After the format filtering passes, the corresponding RSSI value of the data packet is read and compared with the preset minimum threshold of -58dBm. Only data packets with an RSSI value greater than -58dBm are considered to have passed the threshold filtering; data packets with an RSSI value ≤ -58dBm are discarded regardless of their format, ensuring that only valid data within close proximity is retained. Data packets that fail filtering are logged with the reason for failure and their MAC address in a local log, storing up to 1000 entries for subsequent debugging and interference source identification.

[0072] In some embodiments, storing the media access control address, signal strength value, and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue after format filtering and threshold filtering are passed includes: controlling the feeder to extract the media access control address of the corresponding pet collar from the filtered data packets, and tracking only the data packets corresponding to one media access control address at a time; storing the new signal strength value and timestamp into the circular buffer queue in the order of reception, and deleting the oldest historical data stored according to the timestamp when the circular buffer queue is full.

[0073] By limiting the feeder to tracking only the MAC address of one pet collar at a time, interference caused by multiple devices broadcasting simultaneously is avoided. At the same time, the write and overwrite mechanism of the circular buffer queue is optimized to ensure that the latest sampled data is always kept in the cache, providing reliable data support for subsequent status determination.

[0074] The single MAC address tracking state machine implementation includes: Idle state: no MAC address is locked, busy_flag=0. After receiving a filtered data packet, the MAC address is locked, busy_flag is set to 1, and the system switches to tracking state; Tracking state: only data packets that are exactly the same as the locked MAC address are processed, and data packets from other MAC addresses are directly discarded; when there are no valid data packets for the MAC address for 30 consecutive seconds, the MAC address is automatically unlocked, busy_flag is cleared to zero, and the system switches back to idle state; The circular buffer queue is optimized to prevent overflow by using a 32-bit unsigned integer for writing indexes, which increments after each write. The array index is obtained by taking the remainder of 10 during writing, thus avoiding index overflow. When the queue is full, new data automatically overwrites the oldest historical data without manual deletion, ensuring that the queue always retains the latest 10 sets of sampled data. Data mutual exclusion protection adds critical section protection to the write and read operations of the circular buffer queue to avoid data corruption caused by the interrupt callback function writing data at the same time when the verification task reads data, thus ensuring the consistency of data reading and writing; cached data is not lost when power is off by automatically storing the complete sampling data of each pet leaving the field into the local Flash, storing up to 30 days of historical data, which can be exported through the APP for subsequent data analysis and algorithm optimization.

[0075] In some embodiments, controlling the feeder to perform verification at a preset period and clearing historical data in the circular buffer queue that exceeds a preset valid duration includes: controlling the feeder to perform verification operations at a fixed period, obtaining the current time for each verification, comparing the difference between the timestamp of each data in the circular buffer queue and the current time, and clearing historical data whose difference exceeds a preset valid duration.

[0076] Define the fixed cycle and execution logic of the feeder's periodic verification, refine the rules for clearing timed-out data, and ensure that all data involved in the status determination are valid data within the preset time window to avoid distortion of the determination results caused by expired data.

[0077] Precise periodic verification is achieved through the `vTaskDelayUntil` function in FreeRTOS, ensuring a precise 1-second periodic verification and avoiding periodic deviations caused by system scheduling. This guarantees that the time interval error for each verification is ≤10ms. Accurate timestamp acquisition utilizes the feeder's onboard RTC real-time clock to obtain UTC timestamps. The RTC clock is equipped with a backup button battery, allowing it to continue running even after power failure, ensuring timestamp accuracy. Each time the device connects to the network, it automatically synchronizes UTC time via an NTP server to correct clock deviations. Optimized timeout data clearing logic involves iterating through 10 data sets in the circular buffer queue during each verification, calculating the difference between the timestamp and the current time for each set. If the difference exceeds 12 seconds, the RSSI value and timestamp of that set of data are directly cleared. After iteration, the number of remaining valid data sets is counted to provide a basis for subsequent judgments. Anomaly handling automatically pauses the verification task and sends a clock anomaly alert to the app when the RTC clock malfunctions or timestamp acquisition fails, preventing misjudgments caused by invalid data. The verification task automatically restarts after the clock returns to normal.

[0078] In some embodiments, determining that the pet is present and accumulating the presence time if the total number of samples within the effective duration is not less than a first preset threshold and the number of effective samples with a signal strength value greater than a preset threshold is not less than a second preset threshold includes: controlling the feeder to count the total number of samples in the circular buffer queue within the preset effective duration and the number of effective samples with a signal strength value greater than a preset threshold; when the total number of samples is not less than the first preset threshold and the number of effective samples is not less than the second preset threshold, continuously determining that the pet is present and accumulating the pet's presence time according to the verification period.

[0079] By refining the statistical logic and judgment criteria of the dual threshold judgment, clarifying the continuous judgment rules and duration accumulation method of the presence status, we can ensure the accuracy of the pet's presence status judgment and avoid frequent status switching caused by signal fluctuations.

[0080] The dual-threshold judgment logic is solidified in the firmware. The first preset threshold (minimum total number of samples) is fixed at 7, and the second preset threshold (minimum number of valid samples) is fixed at 6. The pet is only determined to be present when both conditions are met simultaneously; if either condition is not met, the pet is determined to be absent. The logic is an "AND" operation with no exceptions. State switching anti-shake processing is implemented by requiring two consecutive periodic checks to satisfy the dual-threshold conditions after the initial determination of presence before officially switching to the present state and starting to accumulate time. After the initial determination of absence, three consecutive periodic checks must fail to satisfy the dual-threshold conditions before the pet is officially switched to the present state and the time accumulation begins. The system officially switches to an absent state, stopping the accumulation of time to avoid frequent state switching caused by instantaneous signal fluctuations. For accurate time accumulation statistics, in the presence state, the time counter increments by 1 for each completed 1-second cycle verification and confirmation of presence, with an accumulation accuracy of 1 second. After the pet leaves, the total accumulated time is stored in local Flash memory and simultaneously reported to the cloud platform to ensure the accuracy of time statistics. Multi-dimensional data verification automatically pushes an anomaly alert to the app when the accumulated time exceeds 30 minutes, preventing abnormal statistics caused by device malfunctions and recording abnormal data for subsequent troubleshooting.

[0081] In some embodiments, determining that the pet has left the venue when all data of the corresponding pet collar in the circular buffer queue is cleared, and reporting a record containing the corresponding media access control address, the current commute time, and the cumulative on-site duration, includes: when all data of the corresponding pet collar in the circular buffer queue is cleared due to exceeding a preset valid duration, controlling the feeder to determine that the pet has left the venue, generating a record containing the media access control address of the corresponding pet collar, the commute time at the time of departure, and the cumulative complete on-site duration, and reporting the record to the bound application server and user terminal.

[0082] By clarifying the criteria for determining when a pet leaves the premises, and refining the rules for generating presence records, reporting formats, and reporting recipients, we ensure complete recording of pet departure events and synchronization of data across multiple devices, providing users with comprehensive pet behavior data.

[0083] The departure judgment standard is fixed: a pet is only judged to have left when all 10 sets of sampled data corresponding to the locked MAC address in the circular buffer queue are cleared due to exceeding the preset valid time. If only some data is cleared and valid data remains, departure judgment is not triggered to avoid false departure reports caused by the pet's brief absence. Standardized presence record generation: immediately after a pet leaves, a fixed-format presence record is generated, including: a hexadecimal string of the collar's 6-byte MAC address, a UTC timestamp at the time of departure, the total accumulated presence time (in seconds), and the sequence number of this feeding. After the record is generated, it is simultaneously stored in local Flash and the cloud platform. To ensure data integrity, a multi-terminal reporting mechanism uses the MQTT protocol to report presence records to the IoT cloud platform. Upon receiving the data, the cloud platform immediately pushes it to the user's bound mobile app, tablet, and other terminal devices, while simultaneously updating the pet's diet data under the user's account. When a device is offline, the records are first stored in local Flash memory. Once the device reconnects to the network, the records are automatically re-reported to ensure no data loss. For reporting anomalies, if three consecutive reporting failures occur, the records are automatically stored in a local offline queue, storing a maximum of 1000 offline records. After the device reconnects to the network, the records are re-reported in chronological order. Once a re-report is successful, the corresponding record is cleared to ensure data integrity.

[0084] In some embodiments, triggering the preset linkage action includes: controlling the feeder to count the number of times the corresponding pet eats and the duration of each feeding based on the reported presence record, and synchronously updating the pet diet data module of the preset terminal; and triggering the preset operation of the corresponding linkage device according to preset rules, which includes at least one or more of the following: drinking water data statistics of the water dispenser, toilet data statistics of the litter box device, and prompting operation of the sound and light reminder device.

[0085] Based on the pet's presence / absence determination results, the execution rules of the linkage actions are refined, the multi-device linkage scenarios are expanded, and the automated statistics and device linkage of pet's eating, drinking, toileting and other behaviors in all scenarios are realized, thereby improving the practicality and scalability of the solution.

[0086] After the pet leaves the feeder, the core linkage actions of the feeder automatically update the pet's daily feeding times, single feeding duration, and cumulative feeding duration according to the current presence record, generate statistical reports in daily / weekly / monthly dimensions, and synchronize them to the pet diet data module of the APP, which are displayed in the form of bar charts; when the cumulative feeding duration is lower than the preset threshold, an abnormal diet reminder is automatically pushed to the APP; multi-device linkage is achieved by the feeder sending linkage instructions to devices such as intelligent water dispensers, intelligent litter boxes, and intelligent weighing litter boxes under the same user account through the local area network MQTT protocol; after receiving the instructions, the intelligent water dispenser counts the drinking times and duration of the corresponding pet and synchronously updates them to the APP; after receiving the instructions, the intelligent litter box enables the proximity recognition function and counts the toilet use times and duration of the corresponding pet; after receiving the instructions, the sound and light reminder device can trigger LED light flashing and buzzer prompts for scenarios such as pet retrieval and behavior training; users can customize linkage rules in the APP, such as "when the pet finishes eating, the water dispenser automatically discharges water for 5 seconds" and "when the pet's daily feeding times are less than 2 times, the APP pushes a reminder". After the rule setting is completed, it is automatically sent to the feeder firmware, and the device automatically executes the linkage actions according to the rules; only the devices bound to the user account can receive linkage instructions, and all instructions are transmitted through encryption to prevent illegal device access and ensure the security of device linkage.

[0087] In some embodiments, the method further includes: controlling the feeder to continuously collect and store the historical presence data of the corresponding pet, analyzing the daily activity patterns and signal strength change characteristics of the pet through machine learning algorithms, and adaptively adjusting the preset signal strength threshold, first preset threshold, and second preset threshold; identifying at least non-feeding scenarios including passing by and wandering based on the movement characteristics of the pet, filtering invalid presence judgments, and reducing the false trigger probability.

[0088] Based on the core method, a machine learning algorithm optimization link is added. Through the learning and analysis of pet historical data, the judgment threshold is adaptively adjusted, and at the same time, non-feeding scenarios are identified, and invalid presence judgments are filtered, further reducing the false trigger rate and improving the adaptability of the solution to different environments and different pets.

[0089] Historical data collection and storage: The feeder continuously collects and stores the historical presence data of the corresponding pet, including daily activity patterns, RSSI change characteristics in different scenarios, feeding duration distribution, and signal characteristics of non-feeding scenarios such as passing by / wandering. The data is stored in the local Flash and cloud platform to provide data sources for algorithm training; The adaptive threshold adjustment algorithm includes: Cloud training: The cloud platform uses a lightweight linear regression algorithm to analyze the pet's historical data and fit the optimal RSSI threshold, total sample threshold, and effective sample threshold for the pet. For example, the RSSI threshold can be appropriately relaxed for larger pets and tightened for smaller pets; Edge deployment: The optimal threshold after training is automatically deployed to the feeder firmware via OTA. The firmware automatically updates the preset threshold without requiring manual adjustment by the user; Online learning: The model is retrained every 7 days to update the threshold and continuously adapt to the pet's growth and environmental changes. Non-feeding scene recognition and filtering includes: Feature extraction: Extracting signal features from passing scenes (RSSI value rises and falls rapidly, total number of samples < 5) and loitering scenes (RSSI value fluctuates frequently, number of effective samples insufficient) to build a feature library; Scene recognition: Using a lightweight decision tree algorithm, the scene corresponding to the current sampled data is identified in real time. When it is determined to be a non-feeding scene such as passing or loitering, even if the double threshold conditions are met, it is not considered to be present, and the duration is not accumulated, effectively filtering out invalid judgments; Lightweight algorithm adaptation: The algorithm model adopts a lightweight design and can run directly on the feeder MCU without relying on the cloud. It can still realize scene recognition and threshold adjustment normally in offline state, ensuring normal operation of the device when the network is disconnected.

[0090] In some embodiments, for multi-cat households with a single feeder or multiple feeders, the limitations of the original single MAC address tracking solution are overcome. A multi-MAC time-sharing rotation tracking mechanism is designed, allocating an independent broadcast time slot and a circular buffer queue to each pet collar. Combining the collar's built-in identity priority, RSSI spatial distance difference, and dwell time characteristics, the independent eating behavior of multiple cats is accurately distinguished. At the same time, the food-grabbing behavior is identified by the dwelling characteristics of dual devices in the same area, triggering hierarchical linkage control, which completely solves the industry pain points of chaotic feeding data statistics, inability to identify food-grabbing behavior, and lack of protection of the feeding rights of vulnerable cats in multi-cat scenarios.

[0091] By assigning a unique 1-byte timeslot number to each pet collar within the same household and embedding it in the service data segment of the broadcast packet, and simultaneously embedding the pet's identity priority, weight range, and age information into the broadcast packet, a standardized broadcast packet with identity identification and timeslot synchronization is formed. After all collars are powered on, timeslot synchronization is completed through the Bluetooth broadcast channel, broadcasting in a time-divided manner with a basic period of 200ms and a timeslot interval of 10ms, avoiding collisions and conflicts between broadcast packets from multiple collars in the same space, while maintaining a broadcast power of -20dBm to ensure that the effective recognition distance is stable at 30cm. When the G-sensor built into the collar detects that the pet is in a moving state, it automatically embeds a motion status identifier into the broadcast packet, providing a data source for subsequent food-grabbing identification.

[0092] The feeder's BLE module is configured for time-division multiplexing, with an 80ms scan interval and a 50ms scan window. Each scan window listens to the broadcast packets of the corresponding collar according to the time slot number, allowing parallel tracking of up to 8 collars with MAC addresses at the same time. The format filtering stage adds verification of the time slot number and the family UUID, ensuring that only broadcast packets from collars within the same family can pass through. The threshold filtering stage retains the basic threshold of -58dBm and adds broadcast packet time slot matching verification to avoid MAC address confusion caused by time slot misalignment. Eight independent circular buffer queues are created in SRAM, each queue corresponding to a collar with one MAC address. The queue structure is consistent with the original scheme, and each queue is equipped with an independent write index and status flag. The sampled data from different collars are completely isolated and stored without interference.

[0093] A multi-queue rotation verification task with a 1-second cycle is created using a feeder. It iterates through the circular buffer queue corresponding to each MAC address, performs timeout data clearing and dual-threshold presence determination, and independently accumulates the presence time and feeding frequency for each cat. The data is stored separately in local Flash partitions to prevent data from being mixed up. The food-stealing behavior recognition logic determines a food-stealing scenario when two or more collars with the same MAC address simultaneously meet the presence determination conditions, and the RSSI value difference is ≤3dBm and the duration is ≥5 seconds. At the same time, it combines the motion status indicators broadcast by the collars to identify food-stealing actions such as chasing and chasing.

[0094] The tiered linkage actions include: Basic linkage: generating independent feeding data reports for each cat and synchronizing them to the corresponding pet's account section in the APP, supporting independent viewing and comparison of data for multiple cats; Food-stealing warning: when food-stealing behavior is detected, an immediate food-stealing reminder is pushed to the APP, while recording the two parties involved, the duration, and the frequency of occurrence; users can set feeding protection rules for vulnerable cats in the APP. When a high-priority cat is detected stealing food, the feeder automatically closes the food outlet and triggers an audio and visual alert until the stealing cat leaves the effective area, protecting the feeding rights of vulnerable cats.

[0095] In some embodiments, addressing the industry pain points of severe RSSI value distortion and misjudgment / missed judgment in home environments caused by Bluetooth co-channel interference from devices such as refrigerators and routers, signal attenuation due to pet body obstruction, and multipath effects caused by wall / furniture reflection, a dual-antenna spatial diversity reception mechanism is designed at the feeder end. Combined with a real-time co-channel interference detection algorithm and an RSSI dynamic interference cancellation model, it breaks through the limitation of the fixed threshold of a single antenna in the original solution, achieves accurate presence judgment in complex interference environments, reduces the misjudgment rate in extreme interference environments, and does not increase the power consumption of the device.

[0096] A fixed-length 16-bit channel calibration sequence is embedded in the service data segment of the broadcast packet by means of a collar. The calibration sequence uses an anti-interference Barker code. The same calibration sequence is carried in all three broadcast channels of each broadcast packet, providing a reference benchmark for interference detection at the feeder end. Every 10 broadcast cycles, the collar sends an additional full-power (0dBm) calibration broadcast packet for baseline calibration of environmental interference at the feeder end. The calibration broadcast packet does not participate in the presence determination and is only used for updating environmental parameters.

[0097] The feeder is equipped with two orthogonally arranged BLE onboard antennas, one horizontally polarized and one vertically polarized. These two antennas are connected to the RF front-end of the main control chip via an RF switch, supporting simultaneous reception of two RSSI sample values ​​from the same broadcast packet. The scanning mechanism has been upgraded to a dual-antenna synchronous scanning mode, with both antennas activated simultaneously in each scanning window. This allows for the simultaneous acquisition of two independent RSSI values ​​and raw data from the same broadcast packet, avoiding signal loss due to single-antenna blockage. A calibration sequence verification has been added to the format filtering stage. By comparing the correlation between the received calibration sequence and the standard sequence, the interference-to-noise ratio (INR) of the current channel is calculated, and the environmental interference baseline is updated in real time. The threshold filtering stage has been upgraded to a dynamic interference cancellation threshold. The base threshold remains -58dBm, but the threshold offset is automatically adjusted based on the real-time detected INR. The stronger the interference, the tighter the threshold is, while filtering out abnormal RSSI sample values ​​with instantaneous fluctuations exceeding 10dBm. The circular buffer queue has also been upgraded: each set of sampled data simultaneously stores the RSSI values, INR, and timestamps from both antennas, providing a complete data source for subsequent spatial diversity determination.

[0098] Each verification process first clears timeout data, then performs dual-antenna spatial diversity fusion calculation. The maximum ratio combining algorithm is used to generate a fused effective RSSI value for the RSSI values ​​of the two antennas, eliminating signal distortion caused by single-antenna obstruction and multipath reflection. The total sampling threshold remains 7, and the effective sampling threshold remains 6. The criterion for effective sampling is upgraded to (fused RSSI value > dynamically adjusted threshold). An upper limit verification of the interference-to-noise ratio is added; sampling data with interference exceeding the preset upper limit is not included in the effective statistics. The feeder continuously records environmental interference characteristics at different times, automatically learning the interference patterns of the home environment. For example, router interference is low at night, while refrigerator startup interference is high during the day. Scanning parameters and threshold offsets are automatically optimized according to time periods to achieve high accuracy throughout the day. When environmental interference continuously exceeds the upper limit, causing inability to make a normal judgment, an environmental interference reminder is automatically pushed to the APP, suggesting that the user adjust the feeder's placement and enable a high anti-interference mode to reduce the scanning interval and increase the sampling frequency, ensuring the normal operation of the judgment function.

[0099] In some embodiments, by breaking through the limitations of the original solution which only relies on the duration of presence for basic behavioral statistics, the motion data, body posture data, and chewing action features of the G-sensor at the collar end are fused with the presence determination data, feeding frequency / duration data, and food weight data at the feeder end in a multi-dimensional spatiotemporal manner. A lightweight embedded behavioral anomaly recognition algorithm is designed to accurately identify abnormal behaviors such as pet refusal to eat, overeating, pica, difficulty eating, and vomiting. This triggers graded health warnings and emergency treatments, achieving a leapfrog upgrade from basic behavioral statistics to proactive pet health management, which meets the core needs of pet medical care and health management.

[0100] The collar features a built-in high-precision three-axis G-sensor with a sampling frequency of 100Hz, which collects real-time data on the pet's acceleration, body posture, and micro-vibrations from chewing movements. A lightweight feature extraction algorithm calculates three core feature values ​​in real-time: chewing frequency, posture changes, and activity level. The collar embeds these three calculated feature values, the pet's real-time activity level, and battery level information into a 200ms broadcast packet. The broadcast packet retains the original three-segment verification structure to ensure compatibility and anti-interference capabilities. When the G-sensor detects abnormal postures such as violent vomiting or collapsing, it immediately switches to a 100ms high-frequency broadcast mode, embedding an emergency anomaly indicator in the broadcast packet to prioritize triggering the feeder's emergency response.

[0101] In addition to the original three-segment verification, the length and format verification of the feature value data segment are added to ensure the integrity of motion feature data. Data packets that fail the verification are directly discarded. The RSSI basic threshold of -58dBm is retained, and the rationality verification of activity level feature values ​​is added to filter out abnormal sampling data with feature values ​​that exceed the reasonable range, so as to avoid misjudgment caused by sensor failure. Each set of sampling data synchronously stores RSSI value, timestamp, chewing frequency, posture features, and activity level to form a multi-dimensional sampling dataset, providing complete spatiotemporal correspondence data for subsequent anomaly identification.

[0102] Based on the dual-threshold presence determination and the cumulative presence time, the average chewing frequency, activity level during the feeding process, and changes in feeding posture of the pet are simultaneously counted. Combined with the weight sensor data of the feeder, the amount of food consumed per feeding and the feeding speed are calculated to form a complete profile of the feeding behavior. The lightweight abnormal behavior recognition algorithm (which can run directly on the feeder's MCU) includes: Refusal to eat recognition: If a pet eats ≤1 time and for ≤30 seconds within 24 consecutive hours, and its activity level is normal, it is considered an abnormal refusal to eat. Binge eating recognition: If a single meal lasts more than 30 minutes, or the pet eats ≥3 times within 1 hour, and its eating speed is far above the historical average, it is considered a abnormal binge eating. Difficulty eating recognition: If the frequency of chewing during eating is far below the historical average, the pet's posture changes frequently, and the eating time is abnormally prolonged, it is considered a difficulty eating (indicating oral inflammation or dental problems). Emergency abnormality recognition: If an emergency abnormality marker is received from the collar, and abnormal posture and a sudden drop in activity are detected, it is considered an emergency health abnormality. The tiered and coordinated early warning mechanism includes: Level 1 Warning (Observation Level): When a slight abnormality in eating is detected, a health observation reminder is generated in the APP, and the time, frequency and characteristics of the abnormal behavior are pushed at the same time, and users are advised to pay attention. Level 2 Warning (Attention Level): When abnormal eating is detected for more than 2 consecutive days, a dual reminder will be sent to the user via SMS and APP. At the same time, an abnormal eating report of the pet will be generated, along with the normal reference values ​​for the same breed of pet, and a veterinary examination will be recommended. Level 3 Early Warning (Emergency Level): When an emergency health abnormality is detected, the system will immediately dial the user's preset emergency contact number and simultaneously push the pet's real-time location information. It will also activate the sound and light pet-finding mode on the collar to help users quickly find their pet. The feeder continuously stores the pet's historical eating behavior data, automatically generates weekly / monthly health trend reports, and synchronizes them to the APP. It can also connect with partner pet hospitals to enable remote health consultations and dietary plan recommendations.

[0103] In some embodiments, a distributed caching mechanism for the collar, feeder, and surrounding smart pet devices is designed to address extreme scenarios such as low collar battery, home network outage, cloud platform offline, and 4G signal failure. Combined with multiple emergency downgrade working modes, this breaks through the limitations of the original solution that relies on cloud platform reporting and fixed broadcast parameters of the collar. It ensures that the linkage function does not fail and the behavior data is not lost in extreme scenarios, while maximizing the emergency battery life of the collar and solving the industry pain points of device paralysis and data loss in extreme scenarios.

[0104] The collar's firmware features four preset operating modes with built-in automatic mode switching logic, requiring no manual user intervention: Normal mode: -20dBm transmission power, 200ms broadcast cycle, corresponding to normal usage scenarios; Low battery degradation mode: When the battery level is ≤10%, it automatically switches to -25dBm transmission power, 500ms broadcast cycle, disables G-sensor high-frequency sampling, retains only the core broadcast function, and can extend battery life by more than 3 times; Network outage emergency mode: Upon receiving the feeder's network outage broadcast command, it switches to 100ms high-frequency broadcast mode, adding a data storage identifier to the broadcast packet to ensure that sampling data is not lost in network outage scenarios; Emergency retrieval mode: After the user triggers the retrieval mode in the APP, it switches to 0dBm full-power broadcast, 100ms cycle, and simultaneously activates sound and light pet locator for easy long-distance positioning; The collar has a built-in 32KB offline cache partition, which can store up to 7 days of broadcast and sampling data. When the feeder is detected to be offline, it automatically caches its own movement data and broadcast timestamps, and automatically re-reports after the device reconnects.

[0105] The feeder's local Flash memory allocates a 1MB distributed cache partition, divided into a real-time cache, an offline cache, and a relay cache, which can store up to 30 days of pet behavior data. In the event of a network outage, all judgment data and behavior statistics are first stored in the offline cache and will not be lost. Cross-device relay scanning mechanism: When the feeder detects a failure in its own Bluetooth module or the collar is outside its own scanning range, it sends a relay scanning command to devices under the same account, such as smart water fountains, litter boxes, and smart cameras, via the local area network. The surrounding devices start BLE scanning, receive the broadcast packets from the collar, perform two-level filtering, and then send the sampled data back to the feeder via the local area network to achieve whole-house tracking without blind spots. Emergency adaptation of the filtering mechanism: In low power mode, the fluctuation range of the RSSI threshold is automatically widened to avoid missed detections caused by the attenuation of transmission power due to low power. In relay scanning mode, the format filtering rules and threshold parameters are automatically synchronized to the surrounding devices to ensure consistency of multi-device scanning.

[0106] The offline determination mechanism allows the feeder to perform the entire process of clearing timed data, determining presence with dual thresholds, counting behavior, and identifying anomalies locally without relying on a cloud platform, even in the case of network outage. All logic is completed at the edge, and network outage does not affect the core linkage function. The breakpoint resume and data reporting mechanism automatically re-reports all data from the offline cache to the cloud platform in chronological order after the device regains network connectivity. After re-reporting, the data integrity is verified, and missing data is automatically re-reported to ensure 100% consistency between cloud and local data. After successful re-reporting, the offline cache space is released. When the collar's battery level is detected to be ≤5%, a low battery reminder is automatically pushed to the APP, while unnecessary linkage functions are turned off, retaining only the core presence determination and data statistics to maximize the collar's battery life. When the home network is down, the feeder automatically turns on the local area network hotspot. After the user's mobile phone connects to the hotspot, they can directly view the pet's real-time status and historical data on the local area network without relying on the cloud platform. When the feeder's own hardware fails, the core determination logic is automatically sent to nearby relay devices, which complete the presence determination and data caching. After the feeder returns to normal, the data is automatically synchronized to ensure that the core functions are not interrupted.

[0107] In some embodiments, based on all the above embodiments, this embodiment further integrates a passive infrared (PIR) sensor and a 24GHz millimeter-wave radar sensor at the feeder end to construct a three-level fusion judgment system of "Bluetooth distance determination + PIR liveness detection + radar three-dimensional spatial positioning". This solves the core problems of insufficient directionality and false triggering outside the feeding area caused by single Bluetooth omnidirectional perception, while retaining the low power consumption advantage of the original solution.

[0108] The smart feeder provided in this embodiment adds a PIR sensor module, a millimeter-wave radar module and corresponding radio frequency front-end circuit to the original hardware structure. The sensor communicates with the feeder MCU through the I2C bus.

[0109] The PIR sensor selection and layout uses a low-power digital PIR sensor, whose detection area strictly covers the rectangular area directly in front of the feeding bowl, completely shielding the interference sources on the sides, back and top of the feeder.

[0110] Millimeter-wave radar can be installed directly below the PIR sensor and arranged coaxially with it to ensure that the detection areas of the two completely overlap, achieving precise matching of spatial coverage.

[0111] The hardware shielding design uses absorbing materials and metal shielding around the PIR and radar modules to shield the electromagnetic interference from the motor and power module inside the feeder, while also reducing radio frequency interference from the external environment.

[0112] The tiered wake-up low-power operation mechanism employs a three-tiered wake-up mechanism (Bluetooth → PIR → Radar) to ensure that most of the time only the Bluetooth module is in a low-power scanning state, while the PIR and radar are only woken up when necessary, resulting in an overall power consumption increase of no more than 1mA. In standby mode, both the PIR and radar are in deep sleep mode, with only the Bluetooth module performing the original passive scanning and two-stage filtering process. First-level wake-up (Bluetooth trigger): When the Bluetooth module detects a legitimate collar entering a distance range of 30cm±5cm and meets the initial signal strength requirements, it triggers an interrupt to wake up the PIR sensor. Second-level wake-up (PIR trigger): After the PIR sensor wakes up, it samples the ambient infrared signal at a frequency of 1Hz. If a living heat source is detected within 3 seconds, it triggers an interrupt to wake up the millimeter-wave radar; if no living body is detected within 3 seconds, the PIR immediately returns to deep sleep mode, and the radar remains unawakened.

[0113] After the radar is awakened, it samples the spatial coordinates and motion status of the target at a frequency of 10Hz. When it is determined that the pet has left the feeding area or there is no effective target for 5 consecutive seconds, the radar and PIR return to deep sleep mode in turn and resume Bluetooth single-module scanning state.

[0114] Based on the existing Bluetooth dual-threshold judgment, PIR liveness detection and radar spatial positioning judgment conditions are added to form a strict "AND" logic. Only when all three conditions are met simultaneously is the pet determined to be in a feeding presence state: The original Bluetooth dual-level filtering, circular buffer queue data processing, and dual-threshold presence judgment are executed to obtain the Bluetooth judgment result. PIR detection results are verified: The PIR sensor is confirmed to have detected a live heat source within the current time window, excluding interference from inanimate objects (such as dropped toys or sunlight spots). Radar spatial positioning results are verified: The millimeter-wave radar outputs the target's distance, horizontal angle, and vertical angle information. The preset feeding zone is a three-dimensional spatial area with a distance of 0-30cm from the radar, a horizontal angle of ±15°, and a vertical angle of ±10°. The radar detection is considered successful only if three consecutive radar sampling results fall within the preset feeding zone and the target's movement speed is <0.2m / s (excluding cases of rapid passing).

[0115] Only when all three conditions are met simultaneously will the pet officially switch to the presence status and begin accumulating presence time. If any condition is not met, the pet will be judged as not being fed and will not accumulate presence time. This effectively solves the problem of accidental triggering caused by pets lingering on the side or back of the feeder or by the collar being placed alone near the feeder.

[0116] PIR anti-interference optimization incorporates an ambient temperature compensation algorithm to calibrate the PIR detection threshold in real time, eliminating the impact of ambient temperature changes on detection sensitivity; combined with radar motion detection results, it filters out interference from stationary heat sources (such as hot water bottles and heating pads).

[0117] Radar anti-jamming optimization eliminates false targets caused by reflections from metal objects and multipath effects through digital filtering algorithms; combined with Bluetooth identification, it filters out target reflection signals corresponding to non-tethered collars.

[0118] Please see Figure 5 As shown, Figure 5 This is a schematic diagram of the linkage system 200 between a pet collar and a feeder provided in this application embodiment. The linkage system 200 is used to execute the steps of the linkage method between the pet collar and the feeder shown in the above embodiments. The linkage system 200 can be a single server or a server cluster, or it can be a terminal, such as a handheld terminal, a laptop computer, a wearable device, or a robot.

[0119] like Figure 5 As shown, the linkage system 200 between the pet collar and the feeder includes: The data broadcasting unit 201 is used to control the pet collar as a Bluetooth beacon to broadcast data packets containing field information, identification information and service information at a preset transmission power and preset period. The queue buffer unit 202 is used to control the feeder to start Bluetooth passive scanning, listen to the broadcast channel according to preset scanning parameters, and perform format filtering and signal strength value threshold filtering after receiving data packets. After the format filtering and threshold filtering are passed, the media access control address, signal strength value and timestamp of the corresponding pet collar are stored in the local fixed-capacity circular buffer queue. The linkage trigger unit 203 is used to control the feeder to perform verification according to a preset cycle, clear historical data in the circular buffer queue that exceeds the preset valid duration, and if the total number of samples within the valid duration is not less than the first preset threshold and the number of valid samples with a signal strength value greater than the preset threshold is not less than the second preset threshold, it is determined that the pet is present and the presence duration is accumulated; when all data of the corresponding pet collar in the circular buffer queue is cleared, it is determined that the pet has left the scene, reports a record containing the corresponding media access control address, the current commute time and the accumulated presence duration, and triggers a preset linkage action.

[0120] In some embodiments, the control of the pet collar as a Bluetooth beacon broadcasts data packets containing field information, identification information, and service information at a preset transmission power and a preset period, including: controlling the pet collar to adjust the Bluetooth transmission power to a preset low-power transmission power, and broadcasting the data packets at a preset fixed broadcast period; wherein the field information, identification information, and service information in the data packets are all set with fixed check values, and data packets that match the check values ​​are subjected to format filtering and threshold filtering.

[0121] In some embodiments, controlling the feeder to initiate Bluetooth passive scanning and listening to the broadcast channel according to preset scanning parameters includes: controlling the feeder to initiate Bluetooth passive scanning mode, continuously listening to the Bluetooth broadcast channel according to preset scanning intervals and scanning windows, not actively initiating Bluetooth connection requests during the scanning process, and only receiving data packets broadcast by surrounding devices.

[0122] In some embodiments, after receiving the data packet, the sequential execution of format filtering and signal strength value threshold filtering includes: controlling the feeder to first perform format verification on the received data packet, and verifying the verification values ​​of field information, identity information and service information in the data packet; performing threshold filtering on the signal strength value corresponding to the data packet that passes the verification, and determining that the data packet with a signal strength value greater than a preset minimum threshold is filtered through.

[0123] In some embodiments, storing the media access control address, signal strength value, and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue after format filtering and threshold filtering are passed includes: controlling the feeder to extract the media access control address of the corresponding pet collar from the filtered data packets, and tracking only the data packets corresponding to one media access control address at a time; storing the new signal strength value and timestamp into the circular buffer queue in the order of reception, and deleting the oldest historical data stored according to the timestamp when the circular buffer queue is full.

[0124] In some embodiments, controlling the feeder to perform verification at a preset period and clearing historical data in the circular buffer queue that exceeds a preset valid duration includes: controlling the feeder to perform verification operations at a fixed period, obtaining the current time for each verification, comparing the difference between the timestamp of each data in the circular buffer queue and the current time, and clearing historical data whose difference exceeds a preset valid duration.

[0125] In some embodiments, determining that the pet is present and accumulating the presence time if the total number of samples within the effective duration is not less than a first preset threshold and the number of effective samples with a signal strength value greater than a preset threshold is not less than a second preset threshold includes: controlling the feeder to count the total number of samples in the circular buffer queue within the preset effective duration and the number of effective samples with a signal strength value greater than a preset threshold; when the total number of samples is not less than the first preset threshold and the number of effective samples is not less than the second preset threshold, continuously determining that the pet is present and accumulating the pet's presence time according to the verification period.

[0126] In some embodiments, determining that the pet has left the venue when all data of the corresponding pet collar in the circular buffer queue is cleared, and reporting a record containing the corresponding media access control address, the current commute time, and the cumulative on-site duration, includes: when all data of the corresponding pet collar in the circular buffer queue is cleared due to exceeding a preset valid duration, controlling the feeder to determine that the pet has left the venue, generating a record containing the media access control address of the corresponding pet collar, the commute time at the time of departure, and the cumulative complete on-site duration, and reporting the record to the bound application server and user terminal.

[0127] In some embodiments, triggering the preset linkage action includes: controlling the feeder to count the number of times the corresponding pet eats and the duration of each feeding based on the reported presence record, and synchronously updating the pet diet data module of the preset terminal; and triggering the preset operation of the corresponding linkage device according to preset rules, which includes at least one or more of the following: drinking water data statistics of the water dispenser, toilet data statistics of the litter box device, and prompting operation of the sound and light reminder device.

[0128] In some embodiments, the method further includes: controlling the feeder to continuously collect and store the historical presence data of the corresponding pet; analyzing the pet's daily activity patterns and signal strength change characteristics through machine learning algorithms; adaptively adjusting the preset signal strength threshold, the first preset threshold, and the second preset threshold; identifying non-feeding scenarios, including passing by and lingering, based on the pet's movement characteristics; filtering invalid presence determinations; and reducing the probability of false triggering.

[0129] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the pet collar and feeder linkage system and each module described above can be referred to the corresponding content in the various embodiments of the pet collar and feeder linkage method described above, and will not be repeated here.

[0130] The aforementioned method for linking the pet collar and feeder can be implemented as a computer program, which can, for example... Figure 5 It runs on the device shown.

[0131] Please see Figure 6 , Figure 6 This is a schematic block diagram of the structure of a computer device provided in an embodiment of this application. The computer device includes a processor, a memory, and a network interface connected via a device bus, wherein the memory may include a storage medium and internal memory.

[0132] The storage medium may store operating devices and computer programs. The computer program includes program instructions that, when executed, cause the processor to perform any method of linkage between the pet collar and the feeder.

[0133] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0134] Internal memory provides an environment for the execution of computer programs in non-volatile storage media. When the computer program is executed by the processor, it enables the processor to execute any method of linking a pet collar and a feeder.

[0135] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the terminal to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0136] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0137] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps: Control the pet collar to act as a Bluetooth beacon, broadcasting data packets containing field information, identification information, and service information at a preset transmission power and preset period; The feeder is controlled to start Bluetooth passive scanning, listen to the broadcast channel according to the preset scanning parameters, and perform format filtering and signal strength value threshold filtering after receiving data packets. After the format filtering and threshold filtering are passed, the media access control address, signal strength value and timestamp of the corresponding pet collar are stored in the local fixed-capacity circular buffer queue. The feeder is controlled to perform verification according to a preset cycle, clearing historical data in the circular buffer queue that exceeds a preset valid duration. If the total number of samples within the valid duration is not less than the first preset threshold and the number of valid samples with a signal strength value greater than the preset threshold is not less than the second preset threshold, the pet is determined to be present and the presence duration is accumulated. When all data of the corresponding pet collar in the circular buffer queue is cleared, the pet is determined to have left the scene, a record containing the corresponding media access control address, the current commute time and the accumulated presence duration is reported, and a preset linkage action is triggered.

[0138] In some embodiments, the control of the pet collar as a Bluetooth beacon broadcasts data packets containing field information, identification information, and service information at a preset transmission power and a preset period, including: controlling the pet collar to adjust the Bluetooth transmission power to a preset low-power transmission power, and broadcasting the data packets at a preset fixed broadcast period; wherein the field information, identification information, and service information in the data packets are all set with fixed check values, and data packets that match the check values ​​are subjected to format filtering and threshold filtering.

[0139] In some embodiments, controlling the feeder to initiate Bluetooth passive scanning and listening to the broadcast channel according to preset scanning parameters includes: controlling the feeder to initiate Bluetooth passive scanning mode, continuously listening to the Bluetooth broadcast channel according to preset scanning intervals and scanning windows, not actively initiating Bluetooth connection requests during the scanning process, and only receiving data packets broadcast by surrounding devices.

[0140] In some embodiments, after receiving the data packet, the sequential execution of format filtering and signal strength value threshold filtering includes: controlling the feeder to first perform format verification on the received data packet, and verifying the verification values ​​of field information, identity information and service information in the data packet; performing threshold filtering on the signal strength value corresponding to the data packet that passes the verification, and determining that the data packet with a signal strength value greater than a preset minimum threshold is filtered through.

[0141] In some embodiments, storing the media access control address, signal strength value, and timestamp of the corresponding pet collar into a local fixed-capacity circular buffer queue after format filtering and threshold filtering are passed includes: controlling the feeder to extract the media access control address of the corresponding pet collar from the filtered data packets, and tracking only the data packets corresponding to one media access control address at a time; storing the new signal strength value and timestamp into the circular buffer queue in the order of reception, and deleting the oldest historical data stored according to the timestamp when the circular buffer queue is full.

[0142] In some embodiments, controlling the feeder to perform verification at a preset period and clearing historical data in the circular buffer queue that exceeds a preset valid duration includes: controlling the feeder to perform verification operations at a fixed period, obtaining the current time for each verification, comparing the difference between the timestamp of each data in the circular buffer queue and the current time, and clearing historical data whose difference exceeds a preset valid duration.

[0143] In some embodiments, determining that the pet is present and accumulating the presence time if the total number of samples within the effective duration is not less than a first preset threshold and the number of effective samples with a signal strength value greater than a preset threshold is not less than a second preset threshold includes: controlling the feeder to count the total number of samples in the circular buffer queue within the preset effective duration and the number of effective samples with a signal strength value greater than a preset threshold; when the total number of samples is not less than the first preset threshold and the number of effective samples is not less than the second preset threshold, continuously determining that the pet is present and accumulating the pet's presence time according to the verification period.

[0144] In some embodiments, determining that the pet has left the venue when all data of the corresponding pet collar in the circular buffer queue is cleared, and reporting a record containing the corresponding media access control address, the current commute time, and the cumulative on-site duration, includes: when all data of the corresponding pet collar in the circular buffer queue is cleared due to exceeding a preset valid duration, controlling the feeder to determine that the pet has left the venue, generating a record containing the media access control address of the corresponding pet collar, the commute time at the time of departure, and the cumulative complete on-site duration, and reporting the record to the bound application server and user terminal.

[0145] In some embodiments, triggering the preset linkage action includes: controlling the feeder to count the number of times the corresponding pet eats and the duration of each feeding based on the reported presence record, and synchronously updating the pet diet data module of the preset terminal; and triggering the preset operation of the corresponding linkage device according to preset rules, which includes at least one or more of the following: drinking water data statistics of the water dispenser, toilet data statistics of the litter box device, and prompting operation of the sound and light reminder device.

[0146] In some embodiments, the method further includes: controlling the feeder to continuously collect and store the historical presence data of the corresponding pet; analyzing the pet's daily activity patterns and signal strength change characteristics through machine learning algorithms; adaptively adjusting the preset signal strength threshold, the first preset threshold, and the second preset threshold; identifying non-feeding scenarios, including passing by and lingering, based on the pet's movement characteristics; filtering invalid presence determinations; and reducing the probability of false triggering.

[0147] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to implement the steps of the pet collar and feeder linkage method provided in any embodiment of this application.

[0148] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard equipped on the computer device.

[0149] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for linking a pet collar and a feeder, characterized in that, include: Control the pet collar to act as a Bluetooth beacon, broadcasting data packets containing field information, identification information, and service information at a preset transmission power and preset period; The feeder is controlled to start Bluetooth passive scanning, listen to the broadcast channel according to the preset scanning parameters, and perform format filtering and signal strength value threshold filtering after receiving data packets. After the format filtering and threshold filtering are passed, the media access control address, signal strength value and timestamp of the corresponding pet collar are stored in the local fixed-capacity circular buffer queue. The feeder is controlled to perform verification according to a preset cycle, clearing historical data in the circular buffer queue that exceeds a preset valid duration. If the total number of samples within the valid duration is not less than the first preset threshold and the number of valid samples with a signal strength value greater than the preset threshold is not less than the second preset threshold, the pet is determined to be present and the presence duration is accumulated. When all data of the corresponding pet collar in the circular buffer queue is cleared, the pet is determined to have left the scene, a record containing the corresponding media access control address, the current commute time and the accumulated presence duration is reported, and a preset linkage action is triggered.

2. The method according to claim 1, characterized in that, The pet collar, acting as a Bluetooth beacon, broadcasts data packets containing field information, identification information, and service information at a preset transmission power and preset period, including: The pet collar is controlled to adjust its Bluetooth transmission power to a preset low-power transmission power and broadcast the data packets at a preset fixed broadcast period; The data packet contains fields, identity information, and service information, each with a fixed check value. Data packets that match the check values ​​are subject to format filtering and threshold filtering.

3. The method according to claim 1, characterized in that, The control feeder initiates Bluetooth passive scanning and listens to the broadcast channel according to preset scanning parameters, including: The feeder is controlled to start Bluetooth passive scanning mode, continuously listening to the Bluetooth broadcast channel according to the preset scanning interval and scanning window. During the scanning process, it does not actively initiate Bluetooth connection requests, but only receives data packets broadcast by surrounding devices.

4. The method according to claim 1, characterized in that, After receiving the data packet, the following steps are performed sequentially: format filtering and signal strength value threshold filtering, including: The feeder first performs format verification on the received data packets, checking the verification values ​​of the field information, identity information, and service information within the data packets; The signal strength value corresponding to the data packet that passes the verification is threshold filtered, and data packets with a signal strength value greater than the preset minimum threshold are judged to have passed the filtering.

5. The method according to claim 4, characterized in that, After format filtering and threshold filtering pass, the media access control address, signal strength value, and timestamp of the corresponding pet collar are stored in a local fixed-capacity circular buffer queue, including: The feeder controls the filtering of data packets, extracts the media access control address corresponding to the pet collar, and tracks only the data packets corresponding to one media access control address at a time. The new signal strength value and timestamp are stored in the circular buffer queue in the order of reception. When the circular buffer queue is full, the earliest stored historical data is deleted according to the timestamp.

6. The method according to claim 1, characterized in that, The control feeder performs verification at a preset cycle, clearing historical data in the circular buffer queue that has exceeded a preset valid duration, including: The feeder is controlled to perform a verification operation at a fixed period. Each verification first obtains the current time, compares the difference between the timestamp of each data in the circular buffer queue and the current time, and clears historical data whose difference exceeds the preset valid duration.

7. The method according to claim 1, characterized in that, The step of determining that the pet is present and accumulating the presence time if the total number of samples within the effective duration is not less than a first preset threshold and the number of effective samples with a signal strength value greater than a preset threshold is not less than a second preset threshold includes: The feeder is controlled to count the total number of samples in the circular buffer queue within a preset effective time period, as well as the number of valid samples with a signal strength value greater than a preset threshold. When the total number of samples is not less than the first preset threshold and the number of valid samples is not less than the second preset threshold, the pet is continuously determined to be present, and the pet's presence time is accumulated according to the verification cycle.

8. The method according to claim 1, characterized in that, When all data for the corresponding pet collar in the circular buffer queue is cleared, it is determined that the pet has left the site, and a record containing the corresponding media access control address, the current Coordinated Universal Time (UTC), and the cumulative duration of presence is reported, including: When all data for the corresponding pet collar in the circular buffer queue is cleared due to exceeding the preset valid time, the feeder determines that the pet has left the field, generates a record containing the media access control address of the corresponding pet collar, the Coordinated Universal Time (UTC) at the time of departure, and the total duration of presence, and reports the record to the bound application server and user terminal.

9. The method according to claim 1, characterized in that, The triggering of the preset linkage action includes: The feeder controls the pet's feeding frequency and duration based on the reported presence records, and updates the data to the pet diet data module on the preset terminal. According to preset rules, the preset operations of the corresponding linked devices are triggered, including at least one or more of the following: drinking water data statistics of drinking water devices, toilet use data statistics of litter box devices, and prompting operations of sound and light reminder devices.

10. The method according to claim 1, characterized in that, The method further includes: The feeder continuously collects and stores the historical presence data of the corresponding pet. Through machine learning algorithms, it analyzes the pet's daily activity patterns and signal strength change characteristics, and adaptively adjusts the preset signal strength threshold, the first preset threshold, and the second preset threshold. Based on the pet's movement characteristics, identification includes at least non-feeding scenarios such as passing by and lingering, filtering out invalid presence judgments and reducing the probability of false triggers.

Images