A WiFi-CSI-based environment sensing method and system, and a smart device

Abstract

The present application relates to a kind of environment perception method and system based on WiFi-CSI, and intelligent device, the method is automatically discovered and dispatches multiple intelligent devices supporting WiFi-CSI sensing function by local area network multicast service discovery mechanism, initiator equipment can dynamically select different transceiver equipment pair, establish multiple original WiFi frame transmission links of direct connection between equipment.These new direct connection links do not pass through router forwarding, increase the diversity of signal propagation path.Even if human body is in the sensing blind area of original single link, it can also be effectively captured by the newly established direct connection link, thereby significantly expanding the sensing coverage area in the whole house, improve the sensing blind area problem caused by device location fixed or complex house layout.In data acquisition stage, silent synchronization mechanism effectively reduces the data packet conflict and background noise interference caused by same frequency channel competition, so that the continuity of CSI data and effective sampling rate are greatly improved, provide higher quality input data for subsequent sensing algorithm.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to an environmental perception method and system based on WiFi-CSI, and intelligent devices. Background Technology

[0002] With the rapid development of IoT technology and smart homes, human perception technology based on WiFi channel state information (CSI) has received widespread attention due to its advantages such as not requiring additional sensors and utilizing existing communication infrastructure.

[0003] This technology monitors the presence, location, and even activity status of people in an indoor environment by analyzing the multipath effects caused by human movement or presence during the propagation of wireless signals. However, existing technical solutions, especially in real-world deployment scenarios with multiple devices and complex apartment layouts, face numerous challenges.

[0004] Currently, human detection solutions based on WiFi CSI mainly rely on specific device links for data collection, and there are two typical working modes: The first mode is link awareness between the router and the module.

[0005] In this mode, a wireless router typically acts as the signal transmitter, and a terminal device (such as a smart home appliance) with an integrated WiFi module acts as the receiver. After receiving wireless data packets from the router, the receiver extracts CSI data for analysis.

[0006] The advantage of this model is its relative simplicity, as it utilizes existing router equipment in the home network. However, its disadvantages are also quite obvious: the sensing range is entirely limited to the single physical path between the router and the module. The coverage area of ​​this link is typically a narrow strip connecting the two points. Once a person is in the dead zone of this link, or if the distance is too far or there are obstacles blocking the path, the signal attenuation and multipath changes become extremely weak, leading to a sharp decline in sensing effectiveness, or even complete failure. In complex home environments, a single path can hardly cover every corner of the living room, bedroom, kitchen, etc., creating numerous sensing blind spots.

[0007] The second mode is point-to-point sensing between two modules.

[0008] To reduce reliance on specific router hardware, some solutions employ two dedicated WiFi modules that communicate directly. One module is responsible for sending wireless data frames, while the other is responsible for receiving and extracting CSI data.

[0009] This approach increases deployment flexibility to some extent, as any two devices supporting this function can be selected to form a sensing link. However, it is typically limited to a fixed, pre-paired pair of devices, lacking a collaborative mechanism between multiple devices. To achieve seamless coverage throughout the entire house, multiple such independent device pairs need to be deployed.

[0010] These devices operate independently and cannot coordinate with each other, which not only increases hardware costs but also causes serious signal interference between them because they all operate on the same WiFi channel.

[0011] The aforementioned existing technologies exhibit the following four significant drawbacks in practical application scenarios involving multi-device networking: 1. Limited perception range and blind spots: Whether it's a "router-module" link or a "module-to-module" link, their sensing range is essentially based on a straight or near-straight wireless link. In complex home environments with many partitions, a single link simply cannot cover the entire area. If a person is not near the main propagation path of that link, their disturbances are difficult to detect effectively. This makes determining the presence of anyone in the house based on a single path highly unreliable.

[0012] 2. Severe channel interference and lack of coordination: The problem is particularly pronounced when multiple devices supporting sensing capabilities exist within the local area network. Due to the lack of a unified scheduling protocol, these devices compete for channel access using the CSMA / CA (Carrier Sense Multiple Access / Collision Avoidance) mechanism, just like ordinary WiFi terminals. This leads to numerous packet collisions and retransmissions, causing discontinuous CSI data collection in time and generating a large number of invalid or erroneous sampling points, severely impacting the input quality of the sensing algorithm and the accuracy of the final judgment.

[0013] 3. Unable to achieve collaborative silence: At critical moments of data acquisition, the communication activity of other surrounding WiFi devices constitutes significant background noise. Existing solutions lack any mechanism to force other non-sensing devices within the local area network to remain temporarily silent during sensing. This background noise masks subtle CSI signal changes caused by human movement, severely limiting the system's accuracy in sensing weak movements and static human bodies.

[0014] 4. Complex configuration and difficult to expand: To achieve multi-device networking, existing solutions typically require users to manually configure the IP addresses of each device, establish the topology between devices, or perform cumbersome pairing operations. This is too difficult for ordinary home users, resulting in a poor user experience. The system cannot achieve automatic discovery and dynamic scheduling with zero configuration. When adding or replacing sensing devices, complex reconfiguration is often required, hindering the system's flexible expansion and large-scale deployment.

[0015] In summary, existing technologies lack a collaborative scheduling method that can coordinate multiple devices within a local area network through a unified protocol, achieve separation of control and sensing functions, possess a channel silencing mechanism, and dynamically expand the sensing range. Summary of the Invention

[0016] In view of this, the purpose of the present invention is to provide an environmental sensing method and system based on WiFi-CSI, as well as a smart device, to solve the problems of limited sensing range and severe channel interference in existing WiFi-CSI sensing technologies in complex home environments.

[0017] According to a first aspect of the present invention, an environment awareness method based on WiFi-CSI is provided, applied to a smart device configured with a WiFi module, wherein the smart device acts as an initiating device, and the method includes: The local area network multicast service discovery mechanism is used to discover and register collaborative devices that support WiFi-CSI sensing function within the local area network, and generate a list of available collaborative devices. Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; After the selected transceiver pair has reached an agreement, other devices in the list of available cooperating devices are controlled to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Data transmitted from multiple transceiver devices is aggregated, and a preset fusion strategy is used to generate the final environmental perception result.

[0018] Preferably, the step of selecting a transceiver pair based on the list of available collaborative devices and issuing a negotiation command includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status, and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0019] Preferably, the return of the negotiated agreement or rejection response includes: If its own resources are in a state that allows for execution, it returns a negotiated agreement response message; If its own resource status does not meet the execution conditions, it returns a negotiation rejection response message; If the local identifier does not match the identifier of the first or second device, the negotiation request message is ignored and no response is required.

[0020] Preferably, the method further includes: The negotiation is considered successful if both the first and second devices return a consensus response message; otherwise, the negotiation is considered to have failed. If the negotiation fails, select the next pair of transceiver devices and issue a negotiation command.

[0021] Preferably, after the selected transceiver pair has successfully negotiated, controlling other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference includes: After the negotiation is successful, the execution command is sent to all devices in the list of available collaborative devices through the control plane protocol to control the first device and the second device to collect data, and at the same time control other devices except the first device and the second device to enter a silent state during the predetermined collection window period, suspending or reducing unnecessary WiFi communication.

[0022] Preferably, the method further includes: If the initiating device itself does not assume the role of data transmission or data collection, it must also enter a silent state during the collection window period; the silent state includes pausing or reducing the transmission of WiFi data frames except for necessary link layer control behaviors.

[0023] Preferably, the data acquisition includes: The first device transmits raw radio frames at a set frequency during the predetermined acquisition window period; the second device receives the raw radio frames on the same working channel as the first device and extracts the corresponding Radio Channel State Information (CSI) data. The method further includes: The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data, generates preliminary feature values, and sends the preliminary feature values ​​back to the initiating device.

[0024] Preferably, the second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data to generate preliminary feature values, including: The validity determination is based on a comprehensive judgment of one or more of the following indicators: the number of valid samples per unit time, the integrity of continuous sampling, the proportion of outliers, and the frame loss situation. If the valid data reaches the preset threshold, the data collection is considered valid; otherwise, the data collection is considered to have failed or the data quality is insufficient. After determining that the acquisition is valid, the acquired data is preprocessed and features are extracted to generate preliminary feature values ​​for subsequent fusion determination. The preliminary feature values ​​are time-domain features, frequency-domain features, statistics, or other low-dimensional representations extracted by a preset algorithm.

[0025] Preferably, the step of transmitting the preliminary feature value back to the initiating device includes: Construct a result reporting message, which includes at least the unique task identifier, the collection result status flag, and the preliminary feature value. The collection result status flag is used to identify whether the collection is valid, the collection failed, or the data quality is insufficient. If the second device and the initiating device are the same device, the preliminary feature value is handed over to the local processing module without needing to be transmitted back over the network; If the second device and the initiating device are different devices, the result reporting message is sent to the initiating device via UDP protocol unicast.

[0026] Preferably, the control plane protocol is implemented based on the UDP protocol and is used for device discovery, task scheduling, and state synchronization; the data packets of the control plane protocol are forwarded via routers within the local area network. The original wireless frame is an original 802.11 WiFi frame, implemented using a custom management frame or a vendor-specific action frame based on IEEE 802.11. It does not go through the router network layer forwarding and directly performs physical layer point-to-point transmission between the first device and the second device.

[0027] Preferably, the execution instruction includes a preparation time parameter and a predetermined acquisition window period; After receiving the execution instruction, the first and second devices first enter the preparation phase, during which they complete mode switching, parameter loading, cache clearing, receiving filter condition configuration, and resource reservation. After the preparation phase ends, they enter the acquisition window period. Other devices besides the first and second devices enter a silent state during the predetermined acquisition window period, and also enter a silent state during the preparation phase.

[0028] Preferably, the method further includes: When the collection window period ends, the first device sends a collection end notification message to the initiating device and the second device; If the initiating device receives a data collection end notification message when the first preset timeout period is reached, it broadcasts the data collection end notification to other devices in the local area network except for the first and second devices, so that each device exits the silent state. If the initiating device does not receive a data collection end notification message after the first preset timeout period, it sends a forced de-silent state command to other devices in the local area network other than the first and second devices.

[0029] Preferably, the method further includes: When the collection window period ends and the second preset timeout period is reached, if other devices in the silent state, except for the initiating device, the first device and the second device, still do not receive a valid collection end notification message, they will automatically force exit the silent state to avoid long-term silence due to message loss or device abnormality. The second preset timeout duration is greater than the first preset timeout duration.

[0030] Preferably, the step of aggregating data transmitted from multiple different transceiver devices and generating the final environmental perception result using a preset fusion strategy includes: Different transceiver pairs are scheduled from the list of available collaborative devices to establish multiple direct-connection acquisition links; The preliminary feature values ​​transmitted back from the multiple direct-connection acquisition links are aggregated, and one or more fusion strategies, such as rule-based judgment, weighted voting, feature classifier, or probability assessment, are used to generate the confidence level of whether personnel exist or not as the final environmental perception result.

[0031] Preferably, the method further includes: Based on the list of available collaborative devices, a sensing data request is sent to a device with WiFi-CSI sensing capability. The system requests the characteristic information of the wireless channel state information (CSI) data between itself and the routers in the local area network, and incorporates it into the fusion calculation when a preset fusion strategy is adopted.

[0032] Preferably, before summarizing the preliminary feature values ​​transmitted back from multiple direct-connection acquisition links, the method further includes: After receiving the result reporting message, the message format is verified. If the verification is successful, the unique task identifier is matched with the corresponding collection task, and the data of the direct-connection collection link can be used for subsequent fusion processing based on the collection result status flag. For valid task results, they are included in the current round of fusion calculation; for task results that fail to be collected or have insufficient data quality, they are either ignored, downgraded, or the corresponding link is rescheduled for collection.

[0033] Preferably, the method further includes: After generating the final environmental perception result, the initiating device performs subsequent linkage control, result storage, or reporting operations according to the system configuration. If it is necessary to notify other cooperating devices in the local area network, the initiating device constructs a perception result notification message. The perception result notification message includes at least the unique task identifier and the final environmental perception result, and is sent unicast one by one according to the list of available cooperating devices. After receiving the perception result notification message, each collaborative device performs the corresponding linkage operation according to the preset execution strategy if the verification is successful and the result is valid.

[0034] Preferably, the local area network multicast service discovery mechanism adopts mDNS and DNS-SD protocols. The initiating device queries a predefined service type, and the cooperating device publishes its device identifier, capability information, and operating status by registering and responding to the service type. The initiating device performs deduplication based on the device identifier and filters the cooperating devices according to capability information, operating status, verification fields and network reachability to generate a list of available cooperating devices; The list of available collaborative devices is maintained in a cached manner and refreshed and timed out based on the DNS-SD record TTL.

[0035] According to a second aspect of the present invention, a smart device is provided, comprising: WiFi module, used to support WiFi-CSI sensing function; Memory, used to store computer programs; A processor for executing the computer program to implement the methods described above.

[0036] Preferably, the smart device is a combination of one or more of the following: Smart air conditioners, smart refrigerators, smart lights, smart curtains, smart speakers, smart robot vacuum cleaners, and smart TVs.

[0037] According to a third aspect of the present invention, an environmental sensing system based on WiFi-CSI is provided, comprising: A router, and at least two of the aforementioned smart devices; The smart devices are connected to the router to form a local area network.

[0038] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: The system automatically discovers and schedules multiple smart devices supporting WiFi-CSI sensing through a LAN multicast service discovery mechanism. The initiating device can dynamically select different transceiver pairs to establish multiple direct WiFi frame transmission links between devices. These newly added direct links bypass router forwarding, introducing new signal propagation paths beyond the original "router-module" links, thus increasing the diversity of signal propagation paths. Even if a person is in the sensing blind spot of the original single link, they can be effectively detected by the newly established direct links, significantly expanding the sensing coverage area throughout the house and improving the sensing blind spot problem caused by fixed device locations or complex house layouts.

[0039] During the data acquisition phase, the initiating device sends an execution command to all devices on the local area network that support this method via the control plane protocol, requiring all devices except the one currently undertaking the transmission and acquisition tasks to enter a silent state and suspend or reduce unnecessary WiFi communication. This silent synchronization mechanism effectively reduces packet collisions and background noise interference caused by competition for the same frequency channel, significantly improving the continuity and effective sampling rate of CSI data, and providing higher-quality input data for subsequent sensing algorithms.

[0040] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0041] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0042] Figure 1 This is a flowchart illustrating an environmental awareness method based on WiFi-CSI according to an exemplary embodiment; Figure 2 This is a schematic diagram illustrating the extension of the WiFi-CSI human presence sensing range according to an exemplary embodiment; Figure 3 This is a flowchart illustrating an environment awareness method based on WiFi-CSI according to another exemplary embodiment; Figure 4 This is a flowchart illustrating an environmental awareness method based on WiFi-CSI according to another exemplary embodiment. Detailed Implementation

[0043] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0044] Example 1 Figure 1 This is a flowchart illustrating an environmental awareness method based on WiFi-CSI according to an exemplary embodiment, applied to a smart device configured with a WiFi module, wherein the smart device acts as the initiating device. See also... Figure 1 The method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has successfully negotiated, control other devices in the list of available cooperative devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0045] In practice, the smart device is one or more of the following: Smart air conditioners, smart refrigerators, smart lights, smart curtains, smart speakers, smart robot vacuum cleaners, and smart TVs.

[0046] It should be noted that the environmental perception results include: the confidence level of the presence of people and the confidence level of the absence of people.

[0047] It should be noted that WiFi-CSI human body perception range refers to whether the system can reliably detect the human body within its perception range when the human body is in different locations in space.

[0048] It refers to the spatial region in which, under the condition of a wireless link formed by a transmitter (such as a router) and a receiver (such as a WiFi module), the disturbance caused by a human body to the multipath propagation characteristics after entering the space can be stably detected by CSI features and reliably determined by the algorithm.

[0049] The closer a human body is to the direct path between the transmitter and receiver, the higher the detection probability; the further the human body is from the path, or the more completely it is blocked by obstacles, the lower the detection probability.

[0050] The sensing range is essentially a probability distribution of detection related to the spatial location of the human body, rather than a definite boundary. The detection probability gradually decreases as the distance increases, the obstruction becomes stronger, and the human body's movement weakens.

[0051] See Figure 2 , Figure 2 This is a schematic diagram illustrating the extended range of WiFi-CSI human presence detection provided in this embodiment. Figure 2 Devices A, B, C, and D in the diagram represent smart devices that support WiFi-CSI sensing functionality. The WiFi modules they use must be able to support WiFi-CSI data acquisition capabilities and controllable raw WiFi frame transmission capabilities.

[0052] Figure 2 In this process, device A and device D generate a new critical propagation path sensing area for the wireless link under the existing local area network conditions by sending raw WiFi frames, thereby increasing the WiFi-CSI human presence sensing range in the home environment.

[0053] It is understood that the technical solution provided in this embodiment automatically discovers and schedules multiple smart devices supporting WiFi-CSI sensing capabilities through a local area network multicast service discovery mechanism. The initiating device can dynamically select different transceiver pairs to establish multiple direct-connection original WiFi frame transmission links between devices. These newly added direct links do not go through router forwarding, introducing new signal propagation paths in addition to the original "router-module" links, increasing the diversity of signal propagation paths. Even if a human body is in the sensing blind zone of the original single link, it can be effectively captured by the newly established direct links, thereby significantly expanding the sensing coverage area throughout the house and improving the sensing blind zone problem caused by fixed device locations or complex house layouts.

[0054] During the data acquisition phase, the initiating device sends an execution command to all devices on the local area network that support this method via the control plane protocol, requiring all devices except the one currently undertaking the transmission and acquisition tasks to enter a silent state and suspend or reduce unnecessary WiFi communication. This silent synchronization mechanism effectively reduces packet collisions and background noise interference caused by competition for the same frequency channel, significantly improving the continuity and effective sampling rate of CSI data, and providing higher-quality input data for subsequent sensing algorithms.

[0055] Example 2 Figure 3This is a flowchart illustrating an environment awareness method based on WiFi-CSI according to another exemplary embodiment, applied to a smart device configured with a WiFi module, wherein the smart device acts as the initiating device. See also... Figure 3 The method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has successfully negotiated, control other devices in the list of available cooperative devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0056] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: Step S121: Select a pair of transceiver devices from the list of available collaborative devices according to the preset strategy and the sensing task. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status, and task constraints. Step S122: From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Step S123: Construct a negotiation request message, wherein the negotiation request message includes at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; Step S124: Send the negotiation request message unicast to each of the cooperative devices in the list of available cooperative devices one by one; Step S125: After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0057] Specifically, the return of a negotiated agreement or rejection response includes: If its own resources are in a state that allows for execution, it returns a negotiated agreement response message; If its own resource status does not meet the execution conditions, it returns a negotiation rejection response message; If the local identifier does not match the identifier of the first or second device, the negotiation request message is ignored and no response is required.

[0058] Specifically, if both the first device and the second device return a negotiation agreement response message, the negotiation is deemed successful; otherwise, the negotiation is deemed to have failed. If the negotiation fails, select the next pair of transceiver devices and issue a negotiation command.

[0059] In practice, one possible approach to task negotiation is as follows: After identifying the target transceiver pair, the initiating device generates a unique TaskID for this negotiation and constructs a negotiation request message.

[0060] The negotiation request message, its payload includes at least: TaskID: A unique identifier for this sensing task; TxMAC: The MAC address of the data sender; RxMAC: The MAC address of the data acquisition terminal; version ver=0; Force: Forced perception identifier.

[0061] Upon receiving a negotiation request, each collaborating device should first verify the validity of the message. If verification fails, the message should be discarded. If verification passes, the device should determine whether it is the data sender and data collector in the target transceiver pair. If the local MAC address matches TxMAC or RxMAC, further determine whether the current execution conditions are met, including but not limited to processor load, memory resources, wireless link status, whether an uninterruptible task is being executed, and the status of functional modules.

[0062] When the conditions are met, return a negotiated agreement response message; If the conditions are not met, return a negotiation rejection response message.

[0063] If the local machine is neither the data sender nor the data collector in the target transceiver pair, the negotiation request will be ignored and no response is required.

[0064] If the local MAC address does not match either the TxMAC or RxMAC, the local device will not be used as the target for this task. The device may ignore the negotiation request and not return a negotiation result.

[0065] The response message should be as concise as possible, and its payload should include at least: TaskID: Corresponding task identifier; Result: The outcome of the negotiation; 1 indicates agreement, and 0 indicates rejection. version ver=0.

[0066] It is understood that this embodiment, based on Embodiment 1, further refines the specific implementation method of "selecting a transceiver pair based on the available collaborative device list and issuing a negotiation command".

[0067] Understandably, the initiating device selects a transceiver pair from the list of available collaborative devices according to a preset strategy. This preset strategy comprehensively considers one or more factors, including device performance scores, historical signal quality, failure counts, current status, and task constraints. Compared to random selection or fixed pairing, this embodiment can dynamically evaluate the real-time capabilities and historical performance of each collaborative device, prioritizing the device with the best performance, signal quality, and current idleness to perform the sensing task. This improves the quality and success rate of sensing data acquisition and reduces acquisition failures caused by excessive device load or poor link quality.

[0068] During the negotiation process, upon receiving a negotiation request message, the collaborating device first verifies the message's validity. Then, based on its own processor load, memory resources, wireless link status, and whether it is currently executing a task that cannot be interfered with, it determines whether the conditions for execution are met and returns a negotiation agreement or rejection response. This mechanism ensures that sensing tasks do not forcibly occupy the device resources that are performing critical business operations (such as air conditioning refrigeration control, refrigerator temperature control, etc.), avoids interference between sensing functions and the original functions of the device, and achieves harmonious coexistence between sensing tasks and the normal operation of the device.

[0069] The negotiation request message includes a mandatory awareness flag. In non-mandatory mode, negotiation is considered successful only if both the first and second devices return a negotiation agreement response message; if either device refuses, negotiation fails and the next group of devices is automatically selected. In mandatory mode, even if a device fails to return a negotiation agreement response message in a timely manner, the initiating device can still forcibly initiate the data collection task as long as the first or second device does not return a negotiation rejection response message.

[0070] When negotiation fails, the initiating device automatically selects the next pair of transceiver devices and reissues the negotiation command without manual intervention. This mechanism effectively addresses negotiation failures caused by temporary device overload, offline status, or link fluctuations, ensuring that sensing tasks can be executed smoothly in most cases and improving the stability and availability of the entire system.

[0071] The negotiation request message clearly identifies the first device (data sender) and the second device (data acquisition device). Collaborating devices determine their roles by comparing their own identifiers, while non-target devices ignore the request, thus avoiding false responses and invalid communication. Simultaneously, the message validity verification mechanism prevents illegal or corrupted messages from interfering with the system, enhancing communication security.

[0072] In summary, this embodiment further optimizes the scheduling efficiency and reliability of sensing tasks based on Embodiment 1 by using intelligent device selection, flexible negotiation mechanisms, mandatory perception identification, automatic fault-tolerant retry, and precise role positioning, providing a more robust and user-friendly solution for actual deployment.

[0073] Example 3 A flowchart illustrating an environment awareness method based on WiFi-CSI, according to another exemplary embodiment, is applied to a smart device configured with a WiFi module. The smart device acts as the initiating device. The method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has successfully negotiated, control other devices in the list of available cooperative devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0074] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0075] Specifically, step S13, "after the selected transceiver pair has successfully negotiated, controlling other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference," includes: After the negotiation is successful, the execution command is sent to all devices in the list of available collaborative devices through the control plane protocol to control the first device and the second device to collect data, and at the same time control other devices except the first device and the second device to enter a silent state during the predetermined collection window period, suspending or reducing unnecessary WiFi communication.

[0076] Preferably, if the initiating device itself does not assume the role of data transmission or data collection, it also needs to enter a silent state during the collection window period; the silent state includes pausing or reducing the transmission of WiFi data frames except for necessary link layer control behaviors.

[0077] It is understood that this embodiment explicitly stipulates that, after negotiation is successful, the initiating device sends an execution command to all devices in the list of available cooperating devices via the control plane protocol. This not only requires all devices except the first and second devices to enter a silent state, but also requires the initiating device itself to enter a silent state if it is not responsible for data transmission or data acquisition. This design achieves full coverage silence for all non-participating devices within the local area network, minimizing co-channel contention and background noise during the acquisition window. This allows the direct probe link between the first and second devices to obtain a near-clean wireless channel environment, significantly improving the signal-to-noise ratio and acquisition quality of CSI data.

[0078] This embodiment precisely defines the silent state as "pausing or reducing the transmission of WiFi data frames except for necessary link-layer control behaviors." This means that the device does not completely cut off wireless communication during the silent period, but retains necessary link-layer control behaviors (such as Beacon frame responses, ACK acknowledgments, ARP replies, etc.), thereby maintaining the basic connection between the device and the router and avoiding problems such as device disconnection or network topology changes due to prolonged silence. This fine-grained control achieves both interference reduction and ensures the basic communication stability of the local area network.

[0079] Through comprehensive silent control, unnecessary WiFi data frames within the local area network are significantly reduced during the acquisition window, and the probability of collisions between the original WiFi frames sent by the first device and those received by the second device is significantly reduced. The continuous arrival of data packets ensures the temporal integrity of CSI sampling, greatly improving the effective sampling rate and providing a high-quality data foundation for subsequent time-domain and frequency-domain feature extraction, thereby enhancing the accuracy and sensitivity of human presence perception.

[0080] In a typical home environment, the daily communication activities of various smart devices (such as mobile phones, tablets, smart speakers, etc.) constitute continuous radio frequency background noise. This embodiment, by forcing non-participating devices to be silent, alleviates or eliminates to some extent the masking effect of this background noise on weak human body disturbance signals, enabling the system to detect more distant and weaker human movements (such as still breathing, slight turning over, etc.), thereby extending the effective sensing distance and sensing sensitivity.

[0081] The initiating device can silently control all devices with a single command, eliminating the need for individual configuration on each device. Upon receiving the command, each device determines its own role and executes the corresponding silent strategy, achieving logical consistency and simplified deployment. This centralized control, distributed execution architecture reduces implementation complexity on the device side, facilitating widespread implementation across different brands and models of smart devices.

[0082] In summary, this embodiment, through comprehensive silent coverage, refined silent definition, and unified synchronization mechanism, further optimizes the channel environment during the acquisition window based on Embodiment 2, significantly improving the quality of CSI data and the overall performance of the sensing system, and providing a strong guarantee for achieving high-precision, wide-coverage whole-house human presence sensing.

[0083] Example 4 A flowchart illustrating an environment awareness method based on WiFi-CSI, according to another exemplary embodiment, is applied to a smart device configured with a WiFi module. The smart device acts as the initiating device. The method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has successfully negotiated, control other devices in the list of available cooperative devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0084] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0085] The "data acquisition" in step S13 includes: The first device transmits raw radio frames at a set frequency during the predetermined acquisition window period; the second device receives the raw radio frames on the same working channel as the first device and extracts the corresponding Radio Channel State Information (CSI) data. The method further includes: The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data, generates preliminary feature values, and sends the preliminary feature values ​​back to the initiating device.

[0086] Specifically, the control plane protocol is implemented based on the UDP protocol and is used for device discovery, task scheduling, and state synchronization; the data packets of the control plane protocol are forwarded via routers within the local area network. The original wireless frame is an original 802.11 WiFi frame, implemented using a custom management frame or a vendor-specific action frame based on IEEE 802.11. It does not go through the router network layer forwarding and directly performs physical layer point-to-point transmission between the first device and the second device.

[0087] Understandably, the first device transmits raw wireless frames (raw 802.11 WiFi frames) at a set frequency within a predetermined acquisition window. The second device receives these frames and extracts CSI data on the same operating channel as the first device. Because the raw wireless frames are implemented using custom management frames or vendor-specific action frames based on IEEE 802.11, they bypass router network layer forwarding and are directly transmitted point-to-point at the physical layer between devices, thus offering the following advantages: Low latency: Eliminating the router processing and forwarding steps results in lower frame transmission latency, which helps maintain the stability of the frame sending rhythm.

[0088] Path diversity: Direct links complement traditional "router-module" links in space, increasing the diversity of signal propagation paths and helping to cover existing sensing blind spots.

[0089] High controllability: The device can autonomously control the frame transmission frequency, frame format, and channel selection, providing customized detection signals for sensing tasks.

[0090] Example 5 A flowchart illustrating an environment awareness method based on WiFi-CSI, according to another exemplary embodiment, is applied to a smart device configured with a WiFi module. The smart device acts as the initiating device. The method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has successfully negotiated, control other devices in the list of available cooperative devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0091] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0092] The "data acquisition" in step S13 includes: The first device transmits raw radio frames at a set frequency during the predetermined acquisition window period; the second device receives the raw radio frames on the same working channel as the first device and extracts the corresponding Radio Channel State Information (CSI) data. The method further includes: The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data, generates preliminary feature values, and sends the preliminary feature values ​​back to the initiating device.

[0093] Specifically, the control plane protocol is implemented based on the UDP protocol and is used for device discovery, task scheduling, and state synchronization; the data packets of the control plane protocol are forwarded via routers within the local area network. The original wireless frame is an original 802.11 WiFi frame, implemented using a custom management frame or a vendor-specific action frame based on IEEE 802.11. It does not go through the router network layer forwarding and directly performs physical layer point-to-point transmission between the first device and the second device.

[0094] Specifically, the second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data to generate preliminary feature values, including: The validity determination is based on a comprehensive judgment of one or more of the following indicators: the number of valid samples per unit time, the integrity of continuous sampling, the proportion of outliers, and the frame loss situation. If the valid data reaches the preset threshold, the data collection is considered valid; otherwise, the data collection is considered to have failed or the data quality is insufficient. After determining that the acquisition is valid, the acquired data is preprocessed and features are extracted to generate preliminary feature values ​​for subsequent fusion determination. The preliminary feature values ​​are time-domain features, frequency-domain features, statistics, or other low-dimensional representations extracted by a preset algorithm.

[0095] Specifically, the step of sending the preliminary feature value back to the initiating device includes: Construct a result reporting message, which includes at least the unique task identifier, the collection result status flag, and the preliminary feature value. The collection result status flag is used to identify whether the collection is valid, the collection failed, or the data quality is insufficient. If the second device and the initiating device are the same device, the preliminary feature value is handed over to the local processing module without needing to be transmitted back over the network; If the second device and the initiating device are different devices, the result reporting message is sent to the initiating device via UDP protocol unicast.

[0096] It should be noted that the payload of the result reporting message should include at least the following: TaskID: Task identifier; Status: The status flag of the collection result is used to indicate whether the collection is valid, the collection failed, or the data quality is insufficient; Payload: Preliminary feature values ​​and other information related to this data collection; version ver=0.

[0097] Understandably, the second device determines the validity of the collected CSI data by comprehensively evaluating indicators such as the number of valid samples per unit time, the integrity of continuous sampling, the proportion of outliers, and the number of dropped frames. This mechanism ensures that only data meeting quality requirements enters the subsequent processing flow, effectively avoiding the contamination of feature extraction and fusion decisions by inferior data caused by sudden channel interference or temporary equipment failures, thus improving the reliability and robustness of the entire sensing system.

[0098] After confirming the validity of the acquisition, the second device preprocesses and extracts features from the raw CSI data to generate preliminary feature values ​​(such as time-domain features, frequency-domain features, statistics, or other low-dimensional representations), instead of directly transmitting back the massive amount of raw CSI data. This design has the following advantages: Reduced network bandwidth usage: The amount of data in the preliminary feature values ​​is much smaller than that of the original CSI data, making it particularly suitable for transmission on resource-constrained embedded platforms such as smart home devices.

[0099] Reduce the computational burden on the initiating device: Feature extraction is completed locally at the data acquisition end. The initiating device only needs to receive lightweight feature values ​​and make fusion decisions, avoiding the computational pressure caused by centrally processing a large amount of raw data.

[0100] Improved response speed: The return of small amounts of data shortens network transmission time and accelerates the execution cycle of the entire perception closed loop.

[0101] This embodiment supports two return methods: if the second device and the initiating device are the same device, the initial feature value is directly handed over to the local processing module without network return, eliminating the risk of network latency and packet loss; if they are different devices, it is sent via UDP protocol unicast, and if necessary, an acknowledgment and retransmission mechanism can be added at the application layer to ensure reliability. This flexible design allows the system to be deployed in scenarios where a single device works independently, or to be extended to complex scenarios where multiple devices collaborate, balancing efficiency and reliability.

[0102] Through the aforementioned data acquisition, validity determination, feature extraction, and feedback process, each direct-connection acquisition link provides the initiating device with preliminary feature values ​​that have undergone quality screening and dimensionality compression. These high-quality, low-redundancy feature values ​​provide reliable input for subsequent multi-source data fusion, enabling the initiating device to perform weighted voting, feature classification, or probability evaluation based on the feature values ​​from multiple links, thereby generating more accurate and stable final perception results.

[0103] In summary, this embodiment, through the organic combination of direct link acquisition, validity determination, local feature extraction, and flexible backhaul mechanism, further improves the quality and efficiency of CSI data acquisition based on Embodiment 4, reduces system communication and computing overhead, and provides a practical technical solution for large-scale, high-precision whole-house human presence sensing.

[0104] Example 6 A flowchart illustrating an environment awareness method based on WiFi-CSI, according to another exemplary embodiment, is applied to a smart device configured with a WiFi module, wherein the smart device acts as an initiating device, and the method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has reached an agreement, control other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0105] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status, and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0106] Specifically, step S13, "after the selected transceiver pair has successfully negotiated, controlling other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference," includes: After the negotiation is successful, the execution command is sent to all devices in the list of available collaborative devices through the control plane protocol to control the first device and the second device to collect data, and at the same time control other devices except the first device and the second device to enter a silent state during the predetermined collection window period, suspending or reducing unnecessary WiFi communication.

[0107] The execution instruction includes a preparation time parameter and a predetermined acquisition window period; After receiving the execution instruction, the first and second devices first enter the preparation phase, during which they complete mode switching, parameter loading, cache clearing, receiving filter condition configuration, and resource reservation. After the preparation phase ends, they enter the acquisition window period. Other devices besides the first and second devices enter a silent state during the predetermined acquisition window period, and also enter a silent state during the preparation phase.

[0108] It should be noted that the message payload carried in the execution instruction includes at least: TaskID: Task identifier; TxMAC: MAC address of the data sender; RxMAC: MAC address of the data acquisition terminal; TxRate: Target frame transmission frequency, for example, 100 Hz; Duration: The duration of data collection, for example, 10000 ms; PrepareTime: Preparation time, for example, 2000ms; version ver=0.

[0109] Upon receiving an execution command, each device should first verify the message format. If verification fails, the message should be discarded; if verification passes, the device should enter the corresponding state based on the TaskID and its local role.

[0110] Preferably, when the collection window period ends, the first device sends a collection end notification message to the initiating device and the second device; If the initiating device receives a data collection end notification message when the first preset timeout period is reached, it broadcasts the data collection end notification to other devices in the local area network except for the first and second devices, so that each device exits the silent state. If the initiating device does not receive a data collection end notification message after the first preset timeout period, it sends a forced de-silent state command to other devices in the local area network other than the first and second devices.

[0111] It should be noted that when the collection duration reaches the end of the collection window, the first device stops sending raw WiFi data frames, the second device stops target frame filtering and CSI extraction, and the first device unicasts a collection end notification message to the initiating device and the second device.

[0112] The message payload includes at least: TaskID: Task identifier; version ver=0.

[0113] After receiving the data collection completion notification message, the initiating device should first verify the message format. If the verification passes and the TaskID matches the current task, the initiating device should unicast the message to other devices.

[0114] After receiving the data acquisition completion notification message, each device will exit the silent control state and resume normal network communication if the verification is successful and the TaskID matches the current task.

[0115] Preferably, when the collection window period ends and the second preset timeout period is reached, if other devices in the silent state, except for the initiating device, the first device and the second device, still have not received a valid collection end notification message, they will automatically force exit the silent state to avoid long-term silence due to message loss or device abnormality. The second preset timeout duration is greater than the first preset timeout duration.

[0116] It should be noted that the second preset timeout duration and the first preset timeout duration are set based on historical experience values ​​or experimental data. For example, the second preset timeout duration is set to 5 seconds and the first preset timeout duration is set to 2 seconds.

[0117] It is understandable that the execution command explicitly includes a preparation time parameter. After receiving the execution command, both the first and second devices first enter the preparation phase. During this phase, they complete initialization operations such as mode switching, parameter loading, cache clearing, receiver filter configuration, and resource reservation. After the preparation phase ends, they synchronously enter the acquisition window. This mechanism ensures that the data sending end and the data acquisition end start sending and receiving at the same time, avoiding frame loss or inconsistent acquisition start times caused by asynchronous device startup delays, and improving the time alignment accuracy of CSI data.

[0118] All devices other than the first and second devices enter a silent state not only during the acquisition window but also during the preparation phase. This means that unnecessary communication within the local area network is suppressed before the data transmitter begins sending frames, creating a clean channel environment for the upcoming acquisition task. The silence during the preparation phase effectively eliminates sudden communication interference that may occur during device initialization, ensuring channel quality at the start of acquisition.

[0119] When the data collection window ends, the first device proactively sends a data collection completion notification message to both the initiating and second devices. Upon receiving this message, the initiating device immediately broadcasts the notification to other devices on the local area network, causing all devices to exit their silent state and resume normal communication. This mechanism ensures that the silent state does not persist indefinitely, preventing disruption to users' normal network usage due to devices remaining in a silent state after the data collection task is completed.

[0120] If the initiating device does not receive the data collection completion notification message after the first preset timeout period, it will proactively send a command to other devices to forcibly end the silent state. This mechanism effectively addresses the problem of the silent state not being able to be properly ended due to the loss of the data collection completion notification message during network transmission or the abnormal failure of the first device, ensuring the system's self-healing capability.

[0121] Besides the initiating device, the first device, and the second device, other devices in a silent state will automatically exit the silent state if they do not receive a valid data collection completion notification message within the second preset timeout period. The second preset timeout period is longer than the first preset timeout period, forming a hierarchical protection: first, the initiating device attempts to actively end the silent state; if the initiating device also malfunctions (such as power failure or system crash), then each device will automatically exit the silent state as a fallback. This dual timeout mechanism minimizes the risk of prolonged device silence due to single-point failures, ensuring that the basic communication functions of devices within the local area network are not affected.

[0122] By organically combining preparation phase synchronization, data acquisition completion notification, and a two-level timeout exit mechanism, this embodiment significantly improves the robustness and reliability of the system while ensuring data acquisition quality. Even under non-ideal conditions such as packet loss or device malfunction, the system can automatically recover to normal communication status, avoiding the unpleasant experience of users having to manually restart the device to restore network connectivity.

[0123] In summary, this embodiment, through refined timing control and a comprehensive exception handling mechanism, further improves the reliability and ease of use of the sensing system based on previous embodiments, providing a solid technical guarantee for the large-scale deployment of actual products.

[0124] Example 7 A flowchart illustrating an environment awareness method based on WiFi-CSI, according to another exemplary embodiment, is applied to a smart device configured with a WiFi module, wherein the smart device acts as an initiating device, and the method includes: Step S11: Discover and register collaborative devices that support WiFi-CSI sensing function within the local area network through the local area network multicast service discovery mechanism, and generate a list of available collaborative devices; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has reached an agreement, control other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0125] Specifically, step S12, "based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command," includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status, and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

[0126] The "data acquisition" in step S13 includes: The first device transmits raw radio frames at a set frequency during the predetermined acquisition window period; the second device receives the raw radio frames on the same working channel as the first device and extracts the corresponding Radio Channel State Information (CSI) data. The method further includes: The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data, generates preliminary feature values, and sends the preliminary feature values ​​back to the initiating device.

[0127] In practice, step S14, "summarizing the data transmitted back from multiple groups of different transceiver devices and generating the final environmental perception result using a preset fusion strategy," includes: Different transceiver pairs are scheduled from the list of available collaborative devices to establish multiple direct-connection acquisition links; The preliminary feature values ​​transmitted back from the multiple direct-connection acquisition links are aggregated, and one or more fusion strategies, such as rule-based judgment, weighted voting, feature classifier, or probability assessment, are used to generate the confidence level of whether personnel exist or not as the final environmental perception result.

[0128] Preferably, the method further includes: Based on the list of available collaborative devices, a sensing data request is sent to a device with WiFi-CSI sensing capability. The system requests the characteristic information of the wireless channel state information (CSI) data between itself and the routers in the local area network, and incorporates it into the fusion calculation when a preset fusion strategy is adopted.

[0129] Preferably, before summarizing the preliminary feature values ​​transmitted back from multiple direct-connection acquisition links, the method further includes: After receiving the result reporting message, the message format is verified. If the verification is successful, the unique task identifier is matched with the corresponding collection task, and the data of the direct-connection collection link can be used for subsequent fusion processing based on the collection result status flag. For valid task results, they are included in the current round of fusion calculation; for task results that fail to be collected or have insufficient data quality, they are either ignored, downgraded, or the corresponding link is rescheduled for collection.

[0130] Preferably, the method further includes: After generating the final environmental perception result, the initiating device performs subsequent linkage control, result storage, or reporting operations according to the system configuration. If it is necessary to notify other cooperating devices in the local area network, the initiating device constructs a perception result notification message. The perception result notification message includes at least the unique task identifier and the final environmental perception result, and is sent unicast one by one according to the list of available cooperating devices. After receiving the perception result notification message, each collaborative device performs the corresponding linkage operation according to the preset execution strategy if the verification is successful and the result is valid.

[0131] It should be noted that the perception result notification message includes at least: TaskID: Task identifier or result association identifier; Decision: The final perceived outcome; version ver=0.

[0132] After receiving the sensing result notification message, each target device can execute the corresponding linkage operation according to the preset strategy if the verification is successful and the result is valid. At this point, the data transmission and result decision-making phase of this task is complete.

[0133] Understandably, through the above process, this embodiment utilizes multiple sensing-capable devices within a wireless local area network to coordinate and establish multiple direct-connection data acquisition links between devices via the UDP protocol. Without altering the existing home network structure, it effectively increases the coverage path of sensing signals, improves the sensing blind spot problem existing in a single link, and ensures the channel quality of data acquisition through the silent mechanism controlled by the protocol.

[0134] Understandably, the initiating device can schedule different transceiver pairs from the list of available collaborative devices, establish multiple direct-connection acquisition links, and aggregate the preliminary feature values ​​returned from each link. By employing one or more fusion strategies, such as rule-based judgment, weighted voting, feature classifiers, or probability assessment, a confidence level for the presence or absence of personnel is generated as the final sensing result. This design fully leverages the spatial complementarity of multiple links; even if a link has a low detection probability due to occlusion or excessive distance, other links may still provide valid information, thereby significantly improving the coverage of whole-house sensing and the stability and robustness of the final judgment.

[0135] This embodiment also sends sensing data requests to devices with CSI data reception capabilities to obtain CSI data feature information between the device and the router, and incorporates it into the fusion calculation to contribute valuable link information. This mechanism maximizes the utilization of existing devices within the local area network, expands sensing coverage, and reduces upgrade costs for users.

[0136] Before aggregating the preliminary feature values from multiple direct acquisition links, the initiating device performs format verification on the received result report messages, matches the unique task identifier with the corresponding acquisition task, and determines whether the set of data can be used for subsequent fusion processing according to the acquisition result status flag. For the task results with valid acquisition, they are included in the fusion calculation of the current round; for the results with failed acquisition or insufficient data quality, corresponding links are selected to be ignored, subjected to weight reduction processing, or re-scheduled to perform acquisition. This quality control mechanism effectively prevents inferior data from contaminating the fusion results, and ensures the accuracy of the final sensing decision.

[0137] After generating the final environment sensing result, the initiating device performs subsequent linkage control, result storage or reporting operations according to the system configuration. For example, when it is detected that no one is at home, devices such as air conditioners and lights can be automatically turned off to save energy; when the presence of a person is detected, the security system can be linked or environmental parameters can be adjusted. If it is necessary to notify other cooperative devices in the local area network, the initiating device constructs a sensing result notification message, which at least includes the unique task identifier and the final sensing result, and sends it one by one via unicast according to the available cooperative device list. After receiving the notification, each cooperative device performs the corresponding linkage operation according to the preset execution strategy when the verification is passed and the result is valid. This mechanism closes the loop between sensing and control, making WiFi CSI sensing no longer an isolated function, but a core component integrated into the overall smart home ecosystem.

[0138] The fusion strategy of this embodiment supports multiple modes such as rule judgment, weighted voting, feature classifier or probability evaluation, and the user or the system can select the most appropriate fusion algorithm according to specific application scenarios (such as security monitoring, energy-saving management, health monitoring, etc.). Meanwhile, the execution strategy of the linkage operation can also be flexibly configured, for example, it can be set to trigger an alarm with high sensitivity in a security scenario, and set to delay control with low sensitivity in an energy-saving scenario. This flexibility enables the present method to adapt to the needs and preferences of different users.

[0139] In conclusion, through multi-link fusion decision-making, compatibility with non-full-function devices, strict result quality control, and sensing-driven linkage control, this embodiment further improves the complete closed loop from sensing to decision-making and then to execution on the basis of the foregoing embodiments, and provides a comprehensive, reliable and scalable whole-house human presence sensing solution for the smart home field.

[0140] Eighth Embodiment A flowchart of a WiFi-CSI based environment sensing method according to another exemplary embodiment, which is applied to a smart device configured with a WiFi module, the smart device serves as an initiating device, and the method includes: Step S11: Discover and register cooperative devices supporting the WiFi-CSI sensing function in the local area network through a local area network multicast service discovery mechanism, and generate an available cooperative device list; Step S12: Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; Step S13: After the selected transceiver pair has reached an agreement, control other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Step S14: Summarize the data transmitted back from multiple different transceiver devices and generate the final environmental perception result using a preset fusion strategy.

[0141] The local area network multicast service discovery mechanism adopts mDNS and DNS-SD protocols. The initiating device queries predefined service types, and the cooperating device publishes its device identifier, capability information and operating status by registering and responding to the service types. The initiating device performs deduplication based on the device identifier and filters the cooperating devices according to capability information, operating status, verification fields and network reachability to generate a list of available cooperating devices; The list of available collaborative devices is maintained in a cached manner and refreshed and timed out based on the DNS-SD record TTL.

[0142] Understandably, after the initiating device starts up, it performs service discovery within the same subnet using a mechanism based on mDNS (Multicast DNS) and DNS-SD (DNS-Based Service Discovery), and queries the predefined service type "g-eps._udp.local.", for example, by querying the PTR record to obtain registered service instances. mDNS works based on link-local multicast and is suitable for LAN scenarios that do not cross network segments.

[0143] Collaborative devices within the local area network that have registered the aforementioned service type and are in an available state publish or respond to their service information, and the device resolution process yields the following information: Service instance name: <instance>._g-eps._udp.local.(PTR) Server endpoint: hostname and port number (SRV), where the port number is the private UDP protocol listening port of the cooperating device; Address record: A / AAAA record corresponding to the hostname, used for IPv4 / IPv6 communication; Extended attributes (TXT): Includes device identifier, capability information, operating status, and verification fields, such as: device_id=xxx mac=AA:BB:CC:DD:EE:FF state=idle / busy score=10 (Performance Rating) mark=FFFFFF (encoding flag, whether it has WiFi, Raw WiFi transmission capability, CSI capability, etc.) version ver=0 After the initiating device receives and parses the response, it deduplicates the responses based on the device identifier and filters cooperating devices according to the information in state, score, mark, and network reachability, generating a list of available cooperating devices. This list is maintained in a cache and is refreshed and cleaned up based on the TTL of the DNS-SD records.

[0144] It is understood that this embodiment uses mDNS (Multicast Domain Name System) and DNS-SD (Domain Name Service Discovery) protocols to achieve device discovery. The initiating device can automatically discover all cooperative devices supporting WiFi-CSI awareness within the local area network by querying predefined service types, without requiring users to manually enter IP addresses, MAC addresses, or perform any pairing operations. Users only need to connect smart devices supporting this solution to the same local area network, and the system can automatically complete discovery and registration, truly achieving "plug and play," greatly reducing the deployment threshold and improving the user experience.

[0145] mDNS and DNS-SD are standardized protocols widely used for service discovery in local area networks. This embodiment adopts these standard protocols, meaning that smart devices from different manufacturers can seamlessly connect to this system as long as they follow the same service type registration specifications, without requiring customized development for specific manufacturers. This lays a solid foundation for building an open and compatible smart home sensing ecosystem and facilitates the large-scale application of the technology.

[0146] When registering and responding, collaborative devices not only publish device identifiers (such as device ID and MAC address), but also capability information (whether they have the capability to send raw wireless frames, collect CSI data, etc.), operating status (idle / busy), and extended attributes such as performance rating. The initiating device can use this information to comprehensively understand the capabilities and current status of each collaborative device, providing a sufficient basis for subsequent intelligent scheduling (such as prioritizing devices with high performance ratings and currently idle status), avoiding data collection failures or performance bottlenecks caused by blind selection.

[0147] Upon receiving a response, the initiating device first deduplicates the device based on its identifier to prevent duplicate registrations of the same device, thus avoiding redundancy in the list. Then, it filters the list based on multiple criteria, including capability information, operating status, verification fields, and network reachability, including only devices that meet the requirements in the list of available collaborative devices. This mechanism effectively excludes devices that do not support sensing functions, are currently busy, have failed verification, or are unreachable via the network, ensuring that every device in the list is genuinely available and meets the task requirements, thereby improving the success rate of subsequent task scheduling.

[0148] The list of available collaborative devices is maintained in a cached manner and refreshed and cleaned up based on the TTL (Time To Live) of the DNS-SD records. When a device goes offline or its service stops, its corresponding DNS-SD record will be automatically cleared after the TTL expires, without the initiating device needing to remove it manually. Simultaneously, when a device comes online or its status changes, this can be reflected in real time by updating the DNS-SD records. This mechanism ensures that the device list always remains consistent with the actual device status within the local area network, avoiding scheduling failures due to device offline or outdated information due to cache expiration.

[0149] In summary, this embodiment, through a standardized automatic discovery mechanism based on mDNS and DNS-SD and a scientific list maintenance strategy, achieves zero-configuration deployment, cross-vendor interoperability, accurate device selection, and real-time status updates. This provides a reliable infrastructure guarantee for subsequent efficient collaborative scheduling and is an important prerequisite for the convenient implementation of the entire sensing solution.

[0150] Example 9 An exemplary embodiment of a smart device is shown, comprising: WiFi module, used to support WiFi-CSI sensing function; Memory, used to store computer programs; A processor for executing the computer program to implement the methods described above.

[0151] One possible implementation of the method is as follows: Figure 4 As shown.

[0152] Preferably, the smart device is a combination of one or more of the following: Smart air conditioners, smart refrigerators, smart lights, smart curtains, smart speakers, smart robot vacuum cleaners, and smart TVs.

[0153] It is understood that the technical solution provided in this embodiment automatically discovers and schedules multiple smart devices supporting WiFi-CSI sensing capabilities through a local area network multicast service discovery mechanism. The initiating device can dynamically select different transceiver pairs to establish multiple direct-connection original WiFi frame transmission links between devices. These newly added direct links do not go through router forwarding, introducing new signal propagation paths outside the original "router-module" links, increasing the diversity of signal propagation paths. Even if a human body is in the sensing blind zone of the original single link, it can be effectively captured by the newly established direct links, thereby significantly expanding the sensing coverage area throughout the house and improving the sensing blind zone problem caused by fixed device locations or complex house layouts.

[0154] During the data acquisition phase, the initiating device sends an execution command to all devices on the local area network that support this method via the control plane protocol, requiring all devices except the one currently undertaking the transmission and acquisition tasks to enter a silent state and suspend or reduce unnecessary WiFi communication. This silent synchronization mechanism effectively reduces packet collisions and background noise interference caused by competition for the same frequency channel, significantly improving the continuity and effective sampling rate of CSI data, and providing higher-quality input data for subsequent sensing algorithms.

[0155] Example 10 An exemplary embodiment of an environmental sensing system based on WiFi-CSI is shown, comprising: A router, and at least two of the aforementioned smart devices; The smart devices are connected to the router to form a local area network.

[0156] It is understood that the technical solution provided in this embodiment automatically discovers and schedules multiple smart devices supporting WiFi-CSI sensing capabilities through a local area network multicast service discovery mechanism. The initiating device can dynamically select different transceiver pairs to establish multiple direct-connection original WiFi frame transmission links between devices. These newly added direct links do not go through router forwarding, introducing new signal propagation paths outside the original "router-module" links, increasing the diversity of signal propagation paths. Even if a human body is in the sensing blind zone of the original single link, it can be effectively captured by the newly established direct links, thereby significantly expanding the sensing coverage area throughout the house and improving the sensing blind zone problem caused by fixed device locations or complex house layouts.

[0157] During the data acquisition phase, the initiating device sends an execution command to all devices on the local area network that support this method via the control plane protocol, requiring all devices except the one currently undertaking the transmission and acquisition tasks to enter a silent state and suspend or reduce unnecessary WiFi communication. This silent synchronization mechanism effectively reduces packet collisions and background noise interference caused by competition for the same frequency channel, significantly improving the continuity and effective sampling rate of CSI data, and providing higher-quality input data for subsequent sensing algorithms.

[0158] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0159] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.

[0160] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0161] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0162] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0163] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0164] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0165] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0166] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.< / instance>

Claims

1. An environment perception method based on WiFi-CSI, applied to a smart device configured with a WiFi module, wherein the smart device acts as the initiating device, characterized in that, include: The local area network multicast service discovery mechanism is used to discover and register collaborative devices that support WiFi-CSI sensing function within the local area network, and generate a list of available collaborative devices. Based on the list of available collaborative devices, select a transceiver pair and issue a negotiation command; After the selected transceiver pair has reached an agreement, other devices in the list of available cooperating devices are controlled to enter a silent synchronization period to reduce controllable interference until the selected transceiver pair completes data acquisition. Data transmitted from multiple transceiver devices is aggregated, and a preset fusion strategy is used to generate the final environmental perception result.

2. The method according to claim 1, characterized in that, The step of selecting a transceiver pair based on the list of available collaborative devices and issuing a negotiation command includes: According to the preset strategy and the sensing task, a pair of transceivers is selected from the list of available collaborative devices. The preset strategy is generated based on one or more of the following: device performance score, historical signal quality, failure count, current status and task constraints. From the selected transceiver pair, designate the first device as the data sender and the second device as the data acquisition device, and generate a unique task identifier for this negotiation; Construct a negotiation request message, the negotiation request message including at least the unique task identifier, the identifier of the first device, the identifier of the second device, and the forced perception identifier; The negotiation request message is unicast to each of the cooperative devices in the list of available cooperative devices; After receiving the negotiation request message, the collaborative device verifies the validity of the message. If the verification passes and the local identifier matches the identifier of the first device or the second device, it determines whether the execution conditions are met based on its own resource status and returns a negotiation agreement or rejection response.

3. The method according to claim 2, characterized in that, The return of a negotiated agreement or rejection response includes: If its own resources are in a state that allows for execution, it returns a negotiated agreement response message; If its own resource status does not meet the execution conditions, it returns a negotiation rejection response message; If the local identifier does not match the identifier of the first or second device, the negotiation request message is ignored and no response is required.

4. The method according to claim 3, characterized in that, Also includes: The negotiation is considered successful if both the first and second devices return a consensus response message; otherwise, the negotiation is considered to have failed. If the negotiation fails, select the next pair of transceiver devices and issue a negotiation command.

5. The method according to claim 2, characterized in that, After the selected transceiver pair has reached an agreement, controlling other devices in the list of available cooperating devices to enter a silent synchronization period to reduce controllable interference includes: After the negotiation is successful, the execution command is sent to all devices in the list of available collaborative devices through the control plane protocol to control the first device and the second device to collect data, and at the same time control other devices except the first device and the second device to enter a silent state during the predetermined collection window period, suspending or reducing unnecessary WiFi communication.

6. The method according to claim 5, characterized in that, Also includes: If the initiating device itself does not play the role of data transmission or data collection, it must also enter a silent state during the collection window period. The silent state includes pausing or reducing the transmission of WiFi data frames except for necessary link layer control actions.

7. The method according to claim 5, characterized in that, The data collection includes: The first device transmits raw radio frames at a set frequency during the predetermined acquisition window period; the second device receives the raw radio frames on the same working channel as the first device and extracts the corresponding Radio Channel State Information (CSI) data. The method further includes: The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data, generates preliminary feature values, and sends the preliminary feature values ​​back to the initiating device.

8. The method according to claim 7, characterized in that, The second device performs validity determination and feature extraction on the collected wireless channel state information (CSI) data to generate preliminary feature values, including: The validity determination is based on a comprehensive judgment of one or more of the following indicators: the number of valid samples per unit time, the integrity of continuous sampling, the proportion of outliers, and the frame loss situation. If the valid data reaches the preset threshold, the data collection is considered valid; otherwise, the data collection is considered to have failed or the data quality is insufficient. After determining that the acquisition is valid, the acquired data is preprocessed and features are extracted to generate preliminary feature values ​​for subsequent fusion determination. The preliminary feature values ​​are time-domain features, frequency-domain features, statistics, or other low-dimensional representations extracted by a preset algorithm.

9. The method according to claim 8, characterized in that, The step of sending the preliminary feature value back to the initiating device includes: Construct a result reporting message, which includes at least the unique task identifier, the collection result status flag, and the preliminary feature value. The collection result status flag is used to identify whether the collection is valid, the collection failed, or the data quality is insufficient. If the second device and the initiating device are the same device, the preliminary feature value is handed over to the local processing module without needing to be transmitted back over the network; If the second device and the initiating device are different devices, the result reporting message is sent to the initiating device via UDP protocol unicast.

10. The method according to claim 7, characterized in that, The control plane protocol is implemented based on the UDP protocol and is used for device discovery, task scheduling, and state synchronization; the data packets of the control plane protocol are forwarded via routers within the local area network. The original wireless frame is an original 802.11 WiFi frame, implemented using a custom management frame or a vendor-specific action frame based on IEEE 802.

11. It does not go through the router network layer forwarding and directly performs physical layer point-to-point transmission between the first device and the second device.

11. The method according to claim 5, characterized in that, The execution instruction includes a preparation time parameter and a predetermined acquisition window period; After receiving the execution instruction, the first and second devices first enter the preparation phase, during which they complete mode switching, parameter loading, cache clearing, receiving filter condition configuration, and resource reservation. After the preparation phase ends, they enter the acquisition window period. Other devices besides the first and second devices enter a silent state during the predetermined acquisition window period, and also enter a silent state during the preparation phase.

12. The method according to claim 5, characterized in that, Also includes: When the collection window period ends, the first device sends a collection end notification message to the initiating device and the second device; If the initiating device receives a data collection end notification message when the first preset timeout period is reached, it broadcasts the data collection end notification to other devices in the local area network except for the first and second devices, so that each device exits the silent state. If the initiating device does not receive a data collection end notification message after the first preset timeout period, it sends a forced de-silent state command to other devices in the local area network other than the first and second devices.

13. The method according to claim 12, characterized in that, Also includes: When the collection window period ends and the second preset timeout period is reached, if other devices in the silent state, except for the initiating device, the first device and the second device, still do not receive a valid collection end notification message, they will automatically force exit the silent state to avoid long-term silence due to message loss or device abnormality. The second preset timeout duration is greater than the first preset timeout duration.

14. The method according to claim 9, characterized in that, The aggregated data from multiple transceiver devices is used to generate the final environmental perception result using a preset fusion strategy, including: Different transceiver pairs are scheduled from the list of available collaborative devices to establish multiple direct acquisition links; The preliminary feature values ​​transmitted back from the multiple direct-connection acquisition links are aggregated, and one or more fusion strategies, such as rule-based judgment, weighted voting, feature classifier, or probability assessment, are used to generate the confidence level of whether personnel exist or not as the final environmental perception result.

15. The method according to claim 14, characterized in that, Also includes: Based on the list of available collaborative devices, a sensing data request is sent to a device with WiFi-CSI sensing capability. The system requests the characteristic information of the wireless channel state information (CSI) data between itself and the routers in the local area network, and incorporates it into the fusion calculation when a preset fusion strategy is adopted.

16. The method according to claim 14, characterized in that, Before summarizing the preliminary feature values ​​transmitted back from multiple direct-connection acquisition links, the process also includes: After receiving the result reporting message, the message format is verified. If the verification is successful, the unique task identifier is matched with the corresponding collection task, and the data of the direct-connection collection link can be used for subsequent fusion processing based on the collection result status flag. For valid task results, they are included in the current round of fusion calculation; for task results that fail to be collected or have insufficient data quality, they are either ignored, downgraded, or the corresponding link is rescheduled for collection.

17. The method according to claim 14, characterized in that, Also includes: After generating the final environmental perception result, the initiating device performs subsequent linkage control, result storage, or reporting operations according to the system configuration. If it is necessary to notify other cooperating devices in the local area network, the initiating device constructs a perception result notification message. The perception result notification message includes at least the unique task identifier and the final environmental perception result, and is sent unicast one by one according to the list of available cooperating devices. After receiving the perception result notification message, each collaborative device performs the corresponding linkage operation according to the preset execution strategy if the verification is successful and the result is valid.

18. The method according to any one of claims 1 to 17, characterized in that, The local area network multicast service discovery mechanism adopts mDNS and DNS-SD protocols. The initiating device queries predefined service types, and the cooperating device publishes its device identifier, capability information and operating status by registering and responding to the service types. The initiating device performs deduplication based on the device identifier and filters the cooperating devices according to capability information, operating status, verification fields and network reachability to generate a list of available cooperating devices; The list of available collaborative devices is maintained in a cached manner and refreshed and timed out based on the DNS-SD record TTL.

19. A smart device, characterized in that, include: WiFi module, used to support WiFi-CSI sensing function; Memory, used to store computer programs; A processor for executing the computer program to implement the method as described in any one of claims 1 to 18.

20. The intelligent device according to claim 19, characterized in that, The smart device is one or more of the following: Smart air conditioners, smart refrigerators, smart lights, smart curtains, smart speakers, smart robot vacuum cleaners, and smart TVs.

21. An environmental sensing system based on WiFi-CSI, characterized in that, include: A router, and at least two smart devices as described in claim 19 or 20; The smart devices are connected to the router to form a local area network.

Images