A multi-device low-power fast networking method based on star flash technology

Abstract

This invention belongs to the field of wireless communication network technology, and particularly relates to a low-power, fast multi-device networking method based on StarFlash technology. The steps are as follows: After the gateway sends a network frame, the device to be networked calculates the optimal joining level based on its own load, signal strength with the gateway, and remaining battery power through an innovative optimal joining level strategy; the gateway, combined with the asynchronous time slot allocation mechanism of StarFlash technology, allocates independent time windows for each level; the device to be networked completes network entry within its assigned level's time slot; the network-entering devices form parent-child relationships, and child nodes accurately calculate wake-up times to reduce idle listening power consumption; all devices adjust relay durations in real time to reduce redundant relays; automatic reconfiguration is performed when the network state changes abruptly to maintain optimal performance. This invention enables efficient, low-power multi-device networking.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication network technology, and particularly relates to a low-power, fast networking method for multiple devices based on star flash technology. Background Technology

[0002] With the rapid development of IoT technology, the demand for multi-device collaborative networking in scenarios such as smart homes, industrial monitoring, and wearable devices is becoming increasingly prominent. Existing wireless networking technologies such as Bluetooth, Wi-Fi, and ZigBee generally have many bottlenecks when facing a large number of devices accessing the network simultaneously: low-power devices suffer from insufficient battery life due to frequent idle listening to maintain the connection; time slot conflicts easily occur when multiple devices join the network concurrently, causing network access delays; it is difficult to dynamically adjust relay strategies when the network topology or device status changes, resulting in redundant forwarding; and resource allocation often adopts a static approach, which cannot adapt to the real-time load and signal strength differences of devices.

[0003] As an emerging short-range wireless communication technology, Starflash technology has advantages such as low latency, high reliability, and flexible time slot scheduling. However, existing networking solutions based on Starflash technology have not yet formed a systematic optimization strategy for multi-device scenarios. They have shortcomings in device level selection, dynamic time slot allocation, time synchronization accuracy, and relay timing convergence. They are unable to meet the requirements of rapid networking and low power consumption, and cannot meet the efficiency and stability requirements of large-scale device collaborative interconnection. Summary of the Invention

[0004] To address the technical problems existing in the above-mentioned background technology, this invention proposes a low-power, fast networking method for multiple devices based on star-flash technology.

[0005] To achieve the above objectives, the technical solution adopted by the present invention includes the following steps:

[0006] S1. The gateway sends network frames. The device to be networked calculates the optimal joining level based on its own load, signal strength with the gateway, and remaining battery power collected in real time, through an innovative optimal joining level strategy.

[0007] S2. The gateway allocates an independent time window for each level based on the hierarchical distribution reported by each device and the asynchronous time slot allocation mechanism of StarFlash technology.

[0008] S3. The device to be networked wakes up in its own time slot and sends a join request. After the gateway sends back confirmation, the network access of the device is completed.

[0009] S4. After joining the network, devices within one hop form a parent-child relationship. The child node uses the next cycle scheduling information dynamically broadcast by the parent node and the two-way delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and reducing idle power consumption.

[0010] S5. All network-connected devices adjust the relay duration in real time based on their level and subnet level during each network cycle, so that the relay timing converges with load changes and reduces redundant relays.

[0011] S6. When a network topology, signal mutation, or significant change in device status is detected that exceeds the threshold, steps S1 to S5 are automatically re-executed to maintain the optimal network state.

[0012] Preferably, the specific implementation of calculating the optimal joining level through the innovative optimal joining level strategy in step S1 is as follows:

[0013] S11. When the network starts, the gateway broadcasts an auction invitation frame, which specifies the maximum number of layers, the set of bidding parameters, the auction round limit, the network period and the penalty increment, and notifies all devices to be networked to enter the bidding preparation state.

[0014] S12. Upon receiving the invitation frame, each device in each network to be formed immediately collects and standardizes the data, including load. Signal strength Battery For each level Calculate the initial price for each piece of equipment. ,in As a set of bidding parameters, the device broadcasts its bids at all levels within the allocated time slots and collects the bid lists of its neighbors at each level in parallel.

[0015] S13. After each round of bidding, the gateway will process each level. Parallel execution of the second-bid auction: From the list of received bids, select the lowest bid and the second lowest bid, designate the equipment with the lowest bid as the winning bidder, and announce the bidding results online;

[0016] S14. If a device wins a bid in a multi-level auction, then for each winning bid level... The net utility is calculated as follows: ,in, Represents the net effect. To preset the benchmark return, Net effect of equipment selection for winning bid price The highest level is the final level to join; all others are considered to have been automatically forfeited.

[0017] S15. For equipment that is not confirmed within the specified time slot after winning the bid, the gateway will add a penalty increment to its original bid in the next round of bidding; at the same time, for the vacant level caused by it, the gateway will immediately initiate a supplementary bidding among the second lowest bidders.

[0018] Preferably, step S2, where the gateway allocates an independent time window for each level based on the hierarchical distribution reported by each device and in conjunction with the asynchronous time slot allocation mechanism of the StarFlash technology, is specifically performed as follows:

[0019] S21. The gateway sorts all levels according to the number of nodes to be networked reported by each device. If the number of nodes in any level is lower than the preset merging threshold, the level is merged with the adjacent level into a joint level to reduce idle time slots.

[0020] S22. First, allocate a shortest guaranteed time slot to each level, and then allocate the remaining time slots according to the proportion of the number of nodes in each level.

[0021] S23. Before each time slot begins, the gateway sends a wake-up preparation message to the top k devices with the highest net utility value in that level. Only these k devices are given priority to enter the answering state, while the remaining devices continue to sleep until the time slot officially begins.

[0022] S24. If a certain level completes all network access confirmations within its time slot, the gateway immediately issues a time slot release, dynamically allocating the remaining time slot to the next level for emergency re-shooting or accelerated network access.

[0023] S25. The gateway calculates the network access completion rate of each level and compares it with the target completion rate. If a level does not meet expectations, it automatically increases its guarantee time slot or merge threshold in the next networking cycle until the network access rate of all levels converges to the target.

[0024] Preferably, before the parent-child relationship is formed between devices within one hop after network entry as described in step S4, it is advisable to determine the parent node. Specifically, this is implemented as follows:

[0025] The child node sends the parent node request signal within the brief query slot of its own level. The sending slot is randomly selected by the child node ID using a hash algorithm to avoid simultaneous transmission by all nodes.

[0026] All potential parent nodes, within a fixed waiting time slot after receiving the request, send out serviceable responses in ascending order of their respective device hash values, with the response time slots having the same length.

[0027] After waiting for a response, the child node temporarily adds the first, second, and third responding nodes to the candidate parent node pool to achieve fast local aggregation.

[0028] The child node is based on the round-robin priority logic and shakes hands with the candidate nodes in the order of its local historical rotation. The first node to successfully establish a handshake is confirmed as the parent node.

[0029] Preferably, in step S4, the child node utilizes the next cycle scheduling information dynamically broadcast by the parent node and the bidirectional delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and reducing idle power consumption. The specific implementation of this is as follows:

[0030] S41. After completing the first network entry, the child node initializes the error sequence and stores the time difference between the actual arrival time of the parent node's network frame and its own preset wake-up time in the most recent N periods in a circular queue.

[0031] S42. When the parent node sends a network access confirmation frame, it appends its frame transmission plan for the next cycle to the frame payload in the form of a time offset table. The table lists the sequence of scheduled frame transmission times relative to the end of the current cycle.

[0032] S43. After receiving and parsing the frame transmission plan, the child node immediately initiates the bidirectional delay detection process and sends a detection request frame with a local transmission timestamp to the parent node. After receiving the data, the parent node sends back a message containing... and its own received timestamp Response frames with local reply timestamps One-way delay of child nodes , ;

[0033] S44, the child nodes are based on the parsed frame transmission schedule. Measured one-way delay and the average deviation of the error sequence Generate a baseline wake-up point , Then, based on the current network load indicators and its remaining power, a compensation buffer time is set to form a candidate time set.

[0034] S45. Select the moment corresponding to the smallest deviation in the error sequence from the candidate moment set as the initial wake-up time, and divide the compensation buffer time into a main monitoring segment and an auxiliary monitoring segment to form a complete wake-up execution window.

[0035] S46. The sub-node calculates the confidence interval based on the initial wake-up time and the standard deviation of the error series. If the standard deviation is less than or equal to the threshold, the initial wake-up time is used directly. If the standard deviation is greater than the threshold, it is corrected based on the benchmark wake-up point and used as the final wake-up time.

[0036] S47. If the parent node frame is not captured in either the main listening segment or the auxiliary listening segment, the difference between the actual listening end time and the parent node frame transmission time is recorded again in the error sequence. When S42 to S46 are re-executed in the next cycle, the error is included in the average deviation of the error sequence. The calculation is performed, and deviation correction is applied.

[0037] Preferably, the error sequence management mechanism in step S41 is specifically implemented as follows:

[0038] S411. The sub-node sets the dynamic length N of the error sequence and adjusts it in real time according to the signal stability. When the error standard deviation is less than or equal to the set threshold for three consecutive periods, the length N changes to 1.25N to accumulate more historical data and improve the accuracy of trend prediction. When the error standard deviation is greater than the set threshold for three consecutive periods, the length N changes to 0.75N to avoid old errors interfering with the current calculation.

[0039] S412. Perform sliding window filtering on the error sequence to remove outliers. If a certain period of error... satisfy ,in If the error is the standard deviation, it is marked as an outlier, the outlier is removed, and the error value of the previous period is used to replace it;

[0040] S413. After each network cycle ends, the child node calculates the first-order difference of the error sequence. The first-order difference is obtained by extracting the current cycle error and the previous cycle error from the error sequence after each network cycle ends, and subtracting them. If the first-order difference is positive twice consecutively, then in step S44, the reference wake-up point is... Make corrections.

[0041] Preferably, in step S5, all network-connected devices adjust their relay duration in real time based on their current layer and subnet layer number during each networking cycle, so that the relay timing converges with load changes, reducing redundant relays. The specific implementation is as follows: First, the device obtains its own layer L and the maximum subnet layer depth S, and calculates the double logarithmic slowly increasing terms, which are respectively... , The adjusted relay duration is calculated and updated based on the preset base relay duration. ,in To preset the basic relay duration, This is the adjusted relay duration.

[0042] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0043] 1. Adopt an innovative optimal access level strategy. Calculate the optimal access level by combining equipment load, signal strength, and power consumption through an auction mechanism. This achieves reasonable equipment layering, reduces network access conflicts, and improves network efficiency.

[0044] 2. By combining the asynchronous time slot allocation of Starflash technology, the levels are dynamically merged and the time slots are adjusted to prioritize the wake-up of high-efficiency equipment, reduce the waste of idle time slots, and reduce the energy consumption of equipment during sleep.

[0045] 3. The sub-nodes utilize bidirectional delay calibration and error sequence management to accurately calculate wake-up time, align network time slots, significantly reduce idle listening power consumption, and improve synchronization accuracy.

[0046] 4. Based on the hierarchy and subnet depth, the relay duration is dynamically adjusted. The relay timing is optimized by using a double logarithmic slow-increment term to reduce redundant relays and improve relay efficiency.

[0047] 5. It has an automatic network reconfiguration mechanism. When it detects sudden changes in network topology, signal or device status, it automatically reconfigures the network to maintain the optimal state and enhance adaptability. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a schematic diagram of the structural process of a multi-device low-power fast networking method based on star flash technology; Detailed Implementation

[0050] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0051] Numerous specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways than those described herein, and therefore the invention is not limited to the specific embodiments disclosed in the following specification.

[0052] In this example, with the rapid growth of IoT devices, the number of devices simultaneously connected to the network has increased significantly in scenarios such as smart homes, industrial monitoring, and environmental monitoring. Existing networking solutions often suffer from high idle power consumption, large network access latency, severe time slot conflicts, and weak dynamic adjustment capabilities when accessing large numbers of devices, making it difficult to simultaneously meet the requirements for rapid networking and low power consumption, and thus failing to satisfy real-time and reliability requirements. To address these issues, this example proposes a multi-device low-power rapid networking method based on a hierarchical auction and asynchronous time slot allocation using StarFlash technology. Through refined hierarchical division, dynamic time slot management, and bidirectional latency calibration, it achieves dual optimization of network access efficiency and power consumption. The specific implementation process is as follows: Figure 1 As shown.

[0053] First, the gateway sends out network frames. The devices to be networked calculate their optimal joining level based on real-time data collected about their own load, signal strength with the gateway, and remaining battery power, using an innovative optimal joining level strategy. Specifically, the calculation of the optimal joining level using this innovative strategy is implemented as follows: At the start of network formation, the gateway broadcasts an auction invitation frame, specifying the maximum number of layers, the bidding parameter set, auction round limits, network formation period, and penalty increment, notifying all devices to be networked to enter the bidding preparation state. Each device in the network, upon receiving the invitation frame, immediately collects and standardizes its data, including load. Signal strength Battery For each level Calculate the initial price for each piece of equipment. ,in As a set of bidding parameters, the device broadcasts its bids for all tiers within the allocated time slots and collects the bid lists of its neighbors for each tier in parallel. After each round of bidding, the gateway performs a check on each tier. Parallel execution of the second-bid auction: From the received bid list, select the lowest bid and the second lowest bid, designate the device with the lowest bid as the winner, and announce the bidding results online; if a device wins in a multi-level auction, then for each winning bid level... The net utility is calculated as follows: ,in, Represents the net effect. To preset the benchmark return, Net effect of equipment selection for winning bid price The highest tier is the final tier to be added, and the rest are considered automatically forfeited. For devices that fail to confirm their bids within the specified time slot after winning the bid, the gateway will add a penalty increment to their original bid in the next round of bidding. At the same time, for any vacant tiers created by this, the gateway will immediately initiate a supplementary auction among the second lowest bidders. The entire process forms a closed loop through tiered auctions, timely feedback, and adaptive penalties. This allows for the rapid construction of the initial network topology while maintaining fair competition and optimal energy consumption during large-scale device access. It also boasts advantages such as strong scalability and flexible deployment.

[0054] Next, to reduce idle power consumption and slot conflict rate, an asynchronous slot allocation mechanism based on StarFlash technology was adopted, along with dynamic slot merging and release effects. The gateway allocates an independent time window for each level based on the hierarchical distribution reported by each device, combined with the asynchronous slot allocation mechanism of StarFlash technology. The specific operation of the gateway allocating an independent time window for each level based on the hierarchical distribution reported by each device and the asynchronous slot allocation mechanism of StarFlash technology is as follows: The gateway sorts all levels according to the number of nodes to be networked reported by each device. If the number of nodes in any level is lower than a preset merging threshold, that level is merged with adjacent levels to form a joint level, thereby reducing idle slots. A shortest guaranteed slot is first allocated to each level, and the remaining slots are distributed proportionally according to the number of nodes in each level. Before each time slot begins, the gateway sends a wake-up preparation message to the top k devices with the highest net utility value in that layer. Only these k devices are given priority to enter the answering state, while the remaining devices remain dormant until the time slot officially begins. If a layer completes all network access confirmations within its time slot, the gateway immediately issues a time slot release, dynamically allocating the remaining time slot to the next layer for emergency re-encoding or accelerated network access. The gateway calculates the network access completion rate for each layer and compares it with the target completion rate. If a layer fails to meet expectations, it automatically increases its guaranteed time slot or merges its threshold in the next networking cycle until the network access rates of all layers converge to the target. Specifically, after receiving network access intention reports from all devices, the gateway first stores the number of nodes reported by each layer into an array and then performs quick sorting (such as using merge or quicksort algorithms) in ascending or descending order to obtain an ordered layer list. The system then iterates through the list. If the number of nodes at any level in the sorted sequence is lower than a preset merging threshold (this threshold can be adaptively adjusted based on network size and historical statistics), the level is merged with the level that is closest in the sorted sequence in terms of the number of nodes. During merging, the gateway generates a new joint level identifier and merges the registration information of all devices at both levels into a list to be allocated. Simultaneously, the level mapping table is updated to map the original level number to the joint level number, ensuring consistency in subsequent time slot allocation logic. After merging, the gateway broadcasts a notice to the affected devices announcing the new level number and the range of time slots to be allocated, and synchronously updates the internal scheduling queue. The system pre-allocates a fixed-length minimum guaranteed time slot for each level. This length is calculated based on the minimum expected number of network access requests and the maximum single frame duration, ensuring that at least basic network access can be completed even under extremely low latency and high load conditions. The total number of remaining time slots is obtained by subtracting the sum of all guaranteed time slots from the total available time slots. Next, the gateway calculates the weight ratio for each layer based on the number of devices currently waiting to join the network. Finally, it allocates the remaining time slots to each layer according to their respective weight ratios. The allocation results are then overlaid with the shortest guaranteed time slot to form the final total available time slot segment for each layer. The gateway distributes the allocation table via multicast or broadcast and synchronizes the start and end times of the time slots across the entire network to ensure that devices at each layer accurately know their network access window.Upon completing the above steps, the gateway has already determined the net utility value of devices at each level. Within a controllable reserved timeframe before the official start of a time slot, the gateway selects the top K devices (K can be a constant or adaptively set according to the level size) with the highest net utility values ​​for each level and sends them individual wake-up preparation instructions—these instructions include the upcoming time slot number, duration, and necessary synchronization timestamps. Devices receiving the instructions immediately exit deep sleep, warm up their RF modules, and open their receive links; unselected devices continue to enter deep sleep until their level's time slot actually begins, at which point they receive a unified wake-up broadcast. Simultaneously, the gateway starts a countdown timer to ensure reliable delivery or retransmission of the instructions within the controllable reserved timeframe. Next, the gateway continuously listens for and counts network access confirmation frames within each level's time slot; once it detects that all devices at that level have successfully joined the network (or have reached the preset minimum confirmation ratio), the gateway immediately generates and broadcasts a time slot release notification, which includes the release start time and the available allocation duration. The released time slot will be dynamically added to the emergency time slot pool of the next level for re-encoding or early network access attempts at that level. The gateway updates the global time slot allocation table and issues change events, enabling next-level devices to receive new wake-up commands and quickly attempt network access before the original time slot. Furthermore, if the network access demand of the next level exceeds the dynamic allocation capacity, the gateway will perform a short-term reallocation based on the remaining total time slots to ensure that the system does not experience time slot waste or congestion. After completing all level time slots, the gateway calculates the completion rate as the number of actual network-connected devices and the total number of devices waiting to join at that level, dividing the completed time slots by the total number of devices waiting to join. This is compared to the preset target completion rate: if the completion rate is greater than or equal to the preset target completion rate, the minimum guaranteed time slot and merging threshold for that level remain unchanged; otherwise, the guaranteed time slot length for that level in the next networking cycle is increased proportionally or its merging threshold is raised, and the adjustment plan is issued via configuration frames. After all level adjustments are completed, the gateway starts the next networking cycle. Through continuous iteration, the network access completion rate is improved with each round until the completion rate of all levels stably reaches or exceeds the target value, thereby ensuring rapid and reliable networking in large-scale multi-device environments.

[0055] The device to be networked wakes up within its assigned time slot and sends a join request. After the gateway sends back an acknowledgment, the device completes its network entry. Specifically, to achieve timely network entry, the device wakes up from deep sleep 10ms before the start of its assigned time slot, opens the RF link, and starts the receiver module to reduce power consumption. At the start of the time slot, the device immediately sends a join request frame, which includes the device ID, layer identifier, capability parameters, and calibration information, enhancing fast connection. After receiving the request frame, the gateway performs address pool and security authentication verification, allocates a short address and time slot information according to the current network topology, and constructs an acknowledgment frame, enhancing reliable network entry. After the acknowledgment frame is sent to the device via the response link, the device parses the acknowledgment information, updates its network parameters, switches to normal network entry status, establishes the data transmission channel, and ensures stable network entry. If the device does not receive an acknowledgment within the time slot, the gateway will trigger a retry mechanism and delay retransmission of the request, guiding the device into a backup time slot, enhancing fault tolerance.

[0056] After joining the network, devices within one hop form a parent-child relationship. The child node utilizes the next-cycle scheduling information dynamically broadcast by the parent node and the bidirectional delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and minimizing idle power consumption. Specifically, the implementation of the child node's use of the next-cycle scheduling information dynamically broadcast by the parent node and the bidirectional delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and minimizing idle power consumption, is as follows: After completing the initial network entry, the child node initializes an error sequence and stores the time difference between the actual arrival time of the parent node's network frames and its own preset wake-up time in a circular queue. When the parent node sends a network entry confirmation frame, it appends its frame transmission plan for the next cycle to the frame payload in the form of a time offset table, which lists the sequence of predetermined frame transmission times relative to the end of the current cycle.

[0057] After receiving and parsing the frame transmission plan, the child node immediately initiates a bidirectional delay detection process, sending a detection request frame with a local transmission timestamp to the parent node. After receiving the data, the parent node sends back a message containing... and its own received timestamp Response frames with local reply timestamps One-way delay of child nodes , The child nodes are based on the parsed frame transmission schedule. Measured one-way delay and the average deviation of the error sequence Generate a baseline wake-up point , Then, based on the current network load indicators and its remaining power, a compensation buffer time is set to form a candidate time set. The time corresponding to the smallest deviation in the error sequence is selected from the candidate time set as the initial wake-up time. Simultaneously, the compensation buffer time is divided into a main monitoring segment and an auxiliary monitoring segment to form a complete wake-up execution window. The child node calculates the confidence interval based on the initial wake-up time and the standard deviation of the error sequence. If the standard deviation is less than or equal to the threshold, the initial wake-up time is directly used; if the standard deviation is greater than the threshold, it is corrected based on the baseline wake-up point as the final wake-up time. If the parent node frame is not captured in either the main monitoring segment or the auxiliary monitoring segment, the difference between the actual end time of monitoring and the parent node's frame transmission time is recorded again in the error sequence. In the next cycle, when re-executing the calibration steps such as bidirectional delay measurement, baseline wake-up point calculation, and confidence interval determination, the error is included in the average deviation of the error sequence. The calculation is performed, and a deviation correction is applied, wherein the deviation correction is...

[0058] The error sequence management mechanism is implemented as follows: The child node sets the dynamic length N of the error sequence and adjusts it in real time based on signal stability. When the error standard deviation is less than or equal to a set threshold for three consecutive periods, the length N changes to 1.25N to accumulate more historical data and improve trend prediction accuracy. When the error standard deviation exceeds the set threshold for three consecutive periods, the length N changes to 0.75N to prevent stale errors from interfering with the current calculation. Then, a sliding window filter is applied to the error sequence to remove outliers. If the error in a certain period... satisfy ,in If the error standard deviation is not specified, it is marked as an outlier, removed, and replaced with the error value from the previous cycle. After each network cycle ends, the child node calculates the first-order difference of the error sequence. This first-order difference is obtained by subtracting the current cycle error and the previous cycle error from the error sequence after each network cycle ends. If the first-order difference is positive twice consecutively, then the reference wake-up point is... Make corrections. The reference wake-up point. To be corrected ,in, The corrected wake-up point, It represents the first-order difference of the current period and the first-order difference of the previous period.

[0059] During each networking cycle, all network-connected devices adjust their relay duration in real time based on their current layer and subnet depth, ensuring that relay timing converges with load changes and reducing redundant relays. Specifically, this real-time adjustment of relay duration based on the current layer and subnet depth involves the device first obtaining its own layer L and the maximum subnet depth S, then calculating a double logarithmic slowly increasing term, which is respectively... , The adjusted relay duration is calculated and updated based on the preset base relay duration. ,in To preset the basic relay duration, This is the adjusted relay duration.

[0060] Finally, when a sudden change in network topology, signal strength, or device status is detected exceeding a threshold, the entire process described above is automatically re-executed to maintain optimal network performance. Specifically, during network operation, the gateway continuously monitors the link quality, topology, and device status of each node. If any link is disconnected, signal strength drops sharply, or key indicators such as device power consumption or load fluctuate beyond a preset threshold, a self-healing process is triggered. At this point, the gateway first suspends normal communication and broadcasts a network reconfiguration notification. Upon receiving this notification, all devices enter a ready state and report their latest status information according to a predetermined time window. Subsequently, the gateway re-executes steps such as hierarchical auctions, asynchronous time slot allocation, latency calibration, and relay duration adjustment to reconstruct the optimal network topology with the latest network parameters. During the reconfiguration process, the gateway can compare the current network access completion rate and latency accuracy with the previous one, dynamically adjusting auction parameters, time slot quotas, and compensation strategies until network performance recovers to the target level.

[0061] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments for application in other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for low-power, fast networking of multiple devices based on star-flash technology, characterized in that, Includes the following steps: S1. The gateway sends network frames. The device to be networked calculates the optimal joining level based on its own load, signal strength with the gateway, and remaining battery power collected in real time, through an innovative optimal joining level strategy. S2. The gateway allocates an independent time window for each level based on the hierarchical distribution reported by each device and the asynchronous time slot allocation mechanism of StarFlash technology. S3. The device to be networked wakes up in its own time slot and sends a join request. After the gateway sends back confirmation, the network access of the device is completed. S4. After joining the network, devices within one hop form a parent-child relationship. The child node uses the next cycle scheduling information dynamically broadcast by the parent node and the two-way delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and reducing idle power consumption. S5. All network-connected devices adjust the relay duration in real time based on their level and subnet level during each network cycle, so that the relay timing converges with load changes and reduces redundant relays. S6. When a network topology, signal change, or significant change in device status is detected that exceeds the threshold, steps S1 to S5 are automatically re-executed to maintain the optimal network state. The specific implementation of calculating the optimal joining level through the innovative optimal joining level strategy in step S1 is as follows: S11. When the network starts, the gateway broadcasts an auction invitation frame, which specifies the maximum number of layers, the set of bidding parameters, the auction round limit, the network period and the penalty increment, and notifies all devices to be networked to enter the bidding preparation state. S12. Upon receiving the invitation frame, each device in each network to be formed immediately collects and standardizes the data, including load. Signal strength Battery For each level Calculate the initial price for each piece of equipment. ,in As a set of bidding parameters, the device broadcasts its bids at all levels within the allocated time slots and collects the bid lists of its neighbors at each level in parallel. S13. After each round of bidding, the gateway will process each level. Parallel execution of the second-bid auction: From the list of received bids, select the lowest bid and the second lowest bid, designate the equipment with the lowest bid as the winning bidder, and announce the bidding results online; S14. If a device wins a bid in a multi-level auction, then for each winning bid level... The net utility is calculated as follows: ,in, Represents the net effect. To preset the benchmark return, Net effect of equipment selection for winning bid price The highest level is the final level to join; all others are considered to have been automatically forfeited. S15. For equipment that is not confirmed within the specified time slot after winning the bid, the gateway will add a penalty increment to its original bid in the next round of bidding; at the same time, for the vacant level caused by it, the gateway will immediately initiate a supplementary bidding among the second lowest bidders.

2. The method for low-power, fast networking of multiple devices based on star-flash technology according to claim 1, characterized in that, The specific operation of step S2, in which the gateway allocates an independent time window for each level based on the hierarchical distribution reported by each device and in conjunction with the asynchronous time slot allocation mechanism of StarFlash technology, is as follows: S21. The gateway sorts all levels according to the number of nodes to be networked reported by each device. If the number of nodes in any level is lower than the preset merging threshold, the level is merged with the adjacent level into a joint level to reduce idle time slots. S22. First, allocate a shortest guaranteed time slot to each level, and then allocate the remaining time slots according to the proportion of the number of nodes in each level. S23. Before each time slot begins, the gateway sends a wake-up preparation message to the top k devices with the highest net utility value in that level. Only these k devices are given priority to enter the answering state, while the remaining devices continue to sleep until the time slot officially begins. S24. If a certain level completes all network access confirmations within its time slot, the gateway immediately issues a time slot release, dynamically allocating the remaining time slot to the next level for emergency re-shooting or accelerated network access. S25. The gateway calculates the network access completion rate of each level and compares it with the target completion rate. If a level does not meet expectations, it automatically increases its guarantee time slot or merge threshold in the next networking cycle until the network access rate of all levels converges to the target.

3. The method for low-power, fast networking of multiple devices based on star flash technology according to claim 1, characterized in that, Before the parent-child relationship is formed between devices within one hop after network entry as described in step S4, it is necessary to determine the parent node. Specifically, this is implemented as follows: The child node sends the parent node request signal within the brief query slot of its own level. The sending slot is randomly selected by the child node ID using a hash algorithm to avoid simultaneous transmission by all nodes. All potential parent nodes, within a fixed waiting time slot after receiving the request, send out serviceable responses in ascending order of their respective device hash values, with the response time slots having the same length. After waiting for a response, the child node temporarily adds the first, second, and third responding nodes to the candidate parent node pool to achieve fast local aggregation. The child node is based on the round-robin priority logic and shakes hands with the candidate nodes in the order of its local historical rotation. The first node to successfully establish a handshake is confirmed as the parent node.

4. The method for low-power, fast networking of multiple devices based on star flash technology according to claim 1, characterized in that, In step S4, the child node utilizes the next cycle scheduling information dynamically broadcast by the parent node and the bidirectional delay calibration mechanism to accurately calculate the wake-up time of the next cycle, thereby aligning the network time slots and reducing idle power consumption. The specific implementation of this is as follows: S41. After completing the first network entry, the child node initializes the error sequence and stores the time difference between the actual arrival time of the parent node's network frame and its own preset wake-up time in the most recent N periods in a circular queue. S42. When the parent node sends a network access confirmation frame, it appends its frame transmission plan for the next cycle to the frame payload in the form of a time offset table. The table lists the sequence of scheduled frame transmission times relative to the end of the current cycle. S43. After receiving and parsing the frame transmission plan, the child node immediately initiates the bidirectional delay detection process and sends a detection request frame with a local transmission timestamp to the parent node. After receiving the data, the parent node sends back a message containing... and its own received timestamp Response frames with local reply timestamps One-way delay of child nodes , ; S44, the child nodes are based on the parsed frame transmission schedule. Measured one-way delay and the average deviation of the error sequence Generate a baseline wake-up point , ; Then, based on the current network load indicators and its remaining power, a compensation buffer time is set to form a candidate time set; S45. Select the moment corresponding to the smallest deviation in the error sequence from the candidate moment set as the initial wake-up time, and divide the compensation buffer time into a main monitoring segment and an auxiliary monitoring segment to form a complete wake-up execution window. S46. The sub-node calculates the confidence interval based on the initial wake-up time and the standard deviation of the error series. If the standard deviation is less than or equal to the threshold, the initial wake-up time is used directly. If the standard deviation is greater than the threshold, it is corrected based on the benchmark wake-up point and used as the final wake-up time. S47. If the parent node frame is not captured in either the main listening segment or the auxiliary listening segment, the difference between the actual listening end time and the parent node frame transmission time is recorded again in the error sequence. When S42 to S46 are re-executed in the next cycle, the error is included in the average deviation of the error sequence. The calculation is performed, and deviation correction is applied.

5. A method for low-power, fast networking of multiple devices based on star-flash technology according to claim 4, characterized in that, The error sequence management mechanism in step S41 is specifically implemented as follows: S411. The sub-node sets the dynamic length N of the error sequence and adjusts it in real time according to the signal stability. When the error standard deviation is less than or equal to the set threshold for three consecutive periods, the length N changes to 1.25N to accumulate more historical data and improve the accuracy of trend prediction. If the standard deviation of the error exceeds the set threshold for three consecutive periods, the length N changes to 0.75N to avoid stale errors interfering with the current calculation. S412. Perform sliding window filtering on the error sequence to remove outliers. If a certain period of error... satisfy ,in If the error is the standard deviation, it is marked as an outlier, the outlier is removed, and the error value of the previous period is used to replace it; S413. After each network cycle ends, the child node calculates the first-order difference of the error sequence. The first-order difference is obtained by extracting the current cycle error and the previous cycle error from the error sequence after each network cycle ends, and subtracting them. If the first-order difference is positive twice consecutively, then in step S44, the reference wake-up point is... Make corrections.

6. The method for low-power, fast networking of multiple devices based on star-flash technology according to claim 1, characterized in that, In step S5, all network-connected devices adjust their relay duration in real time within each networking cycle based on their current layer and subnet layer number, ensuring that the relay timing converges with load changes and reducing redundant relays. Specifically, this is achieved by: first, the device obtains its own layer L and the maximum subnet layer depth S, and calculates a double logarithmic slowly increasing term, which is respectively... , The adjusted relay duration is calculated and updated based on the preset base relay duration. ,in To preset the basic relay duration, This is the adjusted relay duration.

Images