Wireless temperature and humidity sensing method and system for industrial field

Abstract

This invention relates to the field of signal transmission technology, specifically disclosing a wireless temperature and humidity sensing method and system for industrial environments. It establishes a millisecond-level disturbance characteristic sensing channel to achieve dynamic generation of frequency-hopping sequences driven by single-frame snapshots. By combining decision reliability feedback with a frequency point stability weighting factor, frequency point selection considers both instantaneous state and historical performance. Furthermore, it introduces a frequency point marking preamble extension and cross-frame trend tracking to improve receiver synchronization efficiency and ensure inherent continuity in sequence evolution. Finally, it solidifies the context during node sleep / wake-up, guaranteeing the availability of the first frame under intermittent communication. This effectively suppresses the disruptive effects of broadband electromagnetic pulses on preamble acquisition and symbol synchronization, significantly reduces the first frame data packet loss rate and CRC check failure probability, and improves the robustness and temporal determinism of temperature and humidity sensing data transmission in highly disruptive industrial environments.

Description

Technical Field

[0001] This invention relates to the field of signal transmission technology, specifically to a wireless temperature and humidity sensing method and system for industrial environments. Background Technology

[0002] In industrial settings, wireless temperature and humidity sensors typically use the 2.4GHz ISM band for data transmission, relying on Zigbee, Wi-SUN, or proprietary narrowband protocols for low-power networking. Existing technologies generally employ fixed-frequency communication or slow frequency hopping mechanisms based on Channel Quality Indicator (CQI): after power-on, the node scans the entire frequency band and selects one or two channels with the highest received signal strength as the operating frequency. Some solutions introduce periodic channel reassessment at intervals of several seconds to tens of seconds, adjusting the frequency based on average received signal strength or packet error rate. These methods can meet basic requirements in electromagnetically stable environments such as offices or warehouses.

[0003] However, industrial environments are rife with strong transient electromagnetic disturbances, including nanosecond-level voltage spikes from motor contactor start-stop, broadband comb-spectrum radiation from inverter IGBT switching, and harmonic leakage from PLC high-frequency oscillators. These disturbances cover a range from 0.1MHz to 300MHz, and their energy can couple into the 2.4GHz RF link, manifesting at the receiver as sudden sampling distortion, automatic gain control lockout, and symbol timing jitter. Under such interference, existing technologies, relying on long-term statistical averaging or static threshold judgments, cannot respond to millisecond-level disturbance evolution; fixed frequency points are prone to continuously falling into the disturbance peak region, leading to preamble recognition failure and continuous CRC check errors; slow frequency hopping, due to update lag, often continues to use old frequency points even after new disturbances have formed, resulting in a significantly increased first-frame retransmission rate, increased delay in temperature and humidity data reporting, and decreased integrity. Summary of the Invention

[0004] This invention aims to provide a wireless temperature and humidity sensing method and system for industrial environments, reducing the first frame data packet loss rate and CRC check failure probability, and improving the robustness and time determinism of temperature and humidity sensing data transmission in highly disturbed industrial environments.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a wireless temperature and humidity sensing method for industrial sites, comprising:

[0006] Establish an on-site electromagnetic disturbance characteristic sensing channel and collect the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band;

[0007] Based on the initial spectral energy vector, the sub-channels are arranged in ascending order, and the first 8 sub-channels with the lowest energy are selected to generate the initial frequency hopping sequence;

[0008] The preamble parsing process is reconstructed at the receiving end, and the preamble is subjected to fast Fourier transform and weighted accumulation to improve the reliability of symbol synchronization under impulse interference.

[0009] After the first frame data packet is demodulated, the disturbance resenting is initiated to collect the spectrum energy vector of the second frame and generate the second frequency hopping sequence.

[0010] Based on the spectral energy vector of the second frame and the decision reliability index of each symbol in the first frame, a frequency point stability weighting factor is generated, and a third frequency hopping sequence is generated.

[0011] Before the third frame is transmitted, a frequency point marker preamble extension is inserted before the standard preamble, so that the receiver can synchronously confirm the frequency point and channel response through a single fast Fourier transform.

[0012] After three consecutive frames are successfully transmitted, the energy change acceleration of each sub-channel is calculated based on the energy vector of the three frames' spectrum, and a smooth transition rule for the frequency hopping sequence is generated.

[0013] Before a node goes into hibernation, the current frequency hopping sequence, frequency point stability weighting factor, and perturbation evolution matrix are compressed and stored. After waking up, the node loads and supplements a lightweight perturbation snapshot to restore communication.

[0014] Preferably, the establishment of the on-site electromagnetic disturbance characteristic sensing channel includes:

[0015] The RF transceiver chip enters monitoring mode and activates the received signal strength indication detection path.

[0016] The maximum instantaneous amplitude of the I-channel and Q-channel analog baseband signals was collected within a 125-microsecond window. The combined vector magnitude of the maximum instantaneous amplitude of the I-channel and Q-channel was calculated to obtain the energy response of 25 sub-channels.

[0017] The energy responses of the 25 sub-channels are arranged in ascending order of frequency points to form the first frame of spectrum energy snapshot and written to the buffer.

[0018] Preferably, the step of arranging the sub-channels in ascending order based on the initial spectral energy vector and extracting the first 8 sub-channels with the lowest energy to generate the initial frequency hopping sequence includes:

[0019] The first frame of the spectral energy snapshot is sorted in ascending order to obtain a sorted index sequence.

[0020] The first 8 elements of the sorted index sequence are used to form the initial frequency hopping index sub-column;

[0021] The initial frequency hopping index sub-column is loaded into the frequency hopping control register group of the RF chip, and the air propagation time is recorded as the reference time scale for subsequent frequency hopping rhythm synchronization.

[0022] Preferably, the preamble parsing reconstruction process at the receiving end includes:

[0023] The received analog intermediate frequency signal is used to generate a complex baseband sampling stream and flow into a ring buffer;

[0024] A 128-point Fast Fourier Transform is performed on the preamble sample to obtain the frequency domain response;

[0025] The frequency domain response is weighted and accumulated, with the weighting coefficients decreasing with frequency shift. The weighted spectrum is then summed in reverse and subjected to first-order difference to determine the actual main lobe center position.

[0026] Preferably, the step of initiating perturbation resensing to acquire the second frame spectral energy vector includes:

[0027] After the first frame data packet is demodulated, a command is sent to the RF chip to restart the received signal strength indicator detection path.

[0028] A new round of 25 sub-channel energy sampling is performed with the same parameter configuration as the initial spectral energy vector;

[0029] The second frame's spectral energy vector is sorted in ascending order, and the first 8 elements are extracted to form the second frequency hopping sequence. The second frequency hopping sequence is then loaded into the frequency hopping register group, replacing the initial frequency hopping sequence.

[0030] Preferably, the generated frequency point stability weighting factor includes:

[0031] The symbol decision reliability index is extracted from the demodulation process of the first frame data packet. It is obtained by scaling the inverse of the standard deviation of the sampling points. The symbol decision reliability index is then mapped to the corresponding sub-channel number.

[0032] Combined with the second frame spectral energy vector, calculate the frequency point stability weighting factor for each sub-channel; sort the frequency point stability weighting factors in descending order, and extract the first 8 elements to form the third frequency hopping sequence.

[0033] Preferably, the insertion of a frequency point marker preamble extension segment before the standard preamble includes:

[0034] A dedicated area of ​​64 symbols is reserved at the very beginning of the third frame data packet;

[0035] Eight consecutive sine cycles are transmitted at each of the eight target frequencies as frequency markers;

[0036] The receiver performs a 256-point Fast Fourier Transform on the first 64 symbol-length sampled stream;

[0037] Extract the peak position of the spectrum and map it back to the sub-channel number to confirm the actual frequency point used in the current frame.

[0038] Preferably, the smooth transition rule for generating the frequency hopping sequence includes:

[0039] After successful demodulation of the third frame, the spectral energy vector of the third frame is acquired.

[0040] The energy vectors of the three frames of spectrum are stacked vertically to form a perturbation evolution matrix;

[0041] The first-order and second-order differences are calculated for each column of the perturbation evolution matrix to obtain the sub-channel energy change acceleration;

[0042] Based on frequency stability weighting, local neighborhood constraints are applied to the candidate frequency point set to generate smooth transition rules for frequency hopping sequences.

[0043] Preferably, the method further includes performing perturbation snapshot compression and sequence memory solidification, specifically including:

[0044] Perform 8-bit uniform quantization on the frequency stability weighting factor;

[0045] Perform 8-bit quantization on the range of the energy vectors of the three frames;

[0046] The current frequency hopping sequence is stored as an 8-byte integer array. After the node is woken up, the quantized data is read and lightweight disturbance sensing is performed within a sub-channel set with low disturbance activity.

[0047] On the other hand, this invention proposes a wireless temperature and humidity sensing system for industrial environments, comprising:

[0048] The on-site electromagnetic disturbance characteristic sensing module is used to collect the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band;

[0049] The frequency hopping sequence generation module is used to sort the sub-channels in ascending order according to the initial spectral energy vector and extract the first 8 sub-channels with the lowest energy to generate the initial frequency hopping sequence.

[0050] The enhanced preamble processing module is used to reconstruct the preamble parsing process at the receiving end, perform fast Fourier transform on the preamble, and weighted accumulate it.

[0051] The disturbance re-sensing module is used to initiate disturbance re-sensing to acquire the spectral energy vector of the second frame after the first frame data packet demodulation is completed;

[0052] The frequency stability assessment module is used to generate a frequency stability weighting factor based on the second frame spectral energy vector and the decision reliability index of each symbol in the first frame.

[0053] The frequency mark preamble extension module is used to insert a frequency mark preamble extension segment before the standard preamble code before the third frame transmission;

[0054] The cross-frame disturbance trend tracking module is used to calculate the energy change acceleration of each sub-channel based on the three-frame spectral energy vector after three consecutive frames of successful transmission.

[0055] The low-power memory solidification module is used to compress and store the current frequency hopping sequence, frequency stability weighting factor and perturbation evolution matrix before the node goes into sleep mode, and load and restore communication after waking up.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] This invention establishes a millisecond-level disturbance feature sensing channel to achieve dynamic generation of frequency-hopping sequences driven by single-frame snapshots; it constructs a frequency point stability weighting factor by combining decision reliability feedback, ensuring that frequency point selection takes into account both instantaneous state and historical performance; further, it introduces a frequency point marking preamble extension segment and cross-frame trend tracking to improve receiver synchronization efficiency and ensure inherent continuity in sequence evolution; finally, it solidifies the context during node sleep-wake-up process, guaranteeing the availability of the first frame under intermittent communication. Thus, it effectively suppresses the destructive effect of broadband electromagnetic pulses on preamble acquisition and symbol synchronization, significantly reduces the first frame data packet loss rate and CRC check failure probability, and improves the robustness and temporal determinism of temperature and humidity sensing data transmission in highly disturbed industrial environments. Attached Figure Description

[0058] Figure 1 This is a flowchart of the wireless temperature and humidity sensing method for industrial sites according to the present invention.

[0059] Figure 2 This is a block diagram of the wireless temperature and humidity sensing system for industrial environments according to the present invention. Detailed Implementation

[0060] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0061] like Figure 1 As shown, this invention proposes a wireless temperature and humidity sensing method for industrial environments, designed to address broadband electromagnetic pulse interference generated by equipment start-up and shutdown and inverter operation in industrial settings. Through the synergistic effect of dynamic frequency switching and preamble enhancement, the method improves the continuous communication capability of wireless transmission under strong transient interference environments. Specifically, it includes the following steps:

[0062] A field electromagnetic disturbance characteristic sensing channel was established, and the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band were collected. Specifically, the radio frequency transceiver chip entered monitoring mode and the received signal strength indication detection path was activated. The maximum instantaneous amplitude of the I-channel and Q-channel analog baseband signals within a 125 microsecond window was collected, and the composite vector magnitude of the maximum instantaneous amplitude of the I-channel and Q-channel was calculated to obtain the energy response of the 25 sub-channels. The energy responses of the 25 sub-channels were arranged in ascending order of frequency points to form the first frame of spectral energy snapshot and written to the buffer.

[0063] This allows the system to obtain a localized snapshot of the spatial distribution of disturbances within the first millisecond window after power-on, avoiding reliance on historical statistics or external environmental assumptions, and providing an immediate, real, and unavoidable physical basis for subsequent frequency hopping decisions.

[0064] The sub-channels are sorted in ascending order based on the initial spectral energy vector, and the first 8 sub-channels with the lowest energy are extracted to generate the initial frequency hopping sequence. Specifically, this includes: sorting the first frame spectral energy snapshot in ascending order to obtain the sorted index sequence; extracting the first 8 elements of the sorted index sequence to form the initial frequency hopping index sub-column; loading the initial frequency hopping index sub-column into the frequency hopping control register group of the RF chip, and recording the air propagation time as the reference time scale for subsequent frequency hopping rhythm synchronization.

[0065] The first spatial misalignment between the transmitting radio point and the strongest disturbance area is achieved, allowing the first frame data packet to avoid the high-energy impact zone without undergoing any retransmission, thus significantly improving the first frame capture success rate.

[0066] The preamble parsing process is reconstructed at the receiving end. The preamble is subjected to Fast Fourier Transform and weighted accumulation to improve the symbol synchronization reliability under impulse interference. Specifically, this includes: generating a complex baseband sampling stream from the received analog intermediate frequency signal and feeding it into a ring buffer; performing a 128-point Fast Fourier Transform on the preamble sample to obtain the frequency domain response; performing weighted accumulation on the frequency domain response, with the weighting coefficients decaying with frequency offset; and performing inverse accumulation and summation and first-order difference on the weighted spectrum to determine the actual main lobe center position.

[0067] This enables the receiver to stably extract the main energy accumulation region from the frequency domain even under sampling distortion caused by pulse interference, significantly shortening the symbol timing setup time and reducing the preamble recognition omission rate caused by synchronization failure.

[0068] After the first frame data packet is demodulated, the disturbance resensing is initiated to acquire the second frame spectrum energy vector and generate the second frequency hopping sequence. Specifically, this includes: after the first frame data packet is demodulated, a command is sent to the RF chip to restart the received signal strength indication detection path; a new round of 25 sub-channel energy sampling is performed with the same parameter configuration as the initial spectrum energy vector; the second frame spectrum energy vector is sorted in ascending order, the first 8 elements are extracted to form the second frequency hopping sequence, the second frequency hopping sequence is loaded into the frequency hopping register group, and the initial frequency hopping sequence is replaced.

[0069] This ensures that the frequency hopping sequence update is strictly anchored to the end of the first frame transmission, guaranteeing that the disturbance state upon which the second frame transmission is based is synchronized with the environmental evolution within the actual communication gap, thus avoiding the frequency point from falling back into the new disturbance peak area due to perception delay.

[0070] Based on the second frame's spectral energy vector and the reliability indices of each symbol in the first frame, a frequency point stability weighting factor is generated, and a third frequency hopping sequence is generated. Specifically, this includes: extracting the reliability indices of each symbol from the demodulation process of the first frame's data packets, scaling them by the inverse of the standard deviation of the sampling points, and establishing a mapping relationship between the symbol decision reliability indices and the corresponding sub-channel numbers; calculating the frequency point stability weighting factor for each sub-channel by combining the second frame's spectral energy vector; arranging the frequency point stability weighting factors in descending order, and extracting the first 8 elements to form the third frequency hopping sequence.

[0071] This makes frequency hopping selection no longer solely dependent on instantaneous energy levels, but rather on the stability performance of frequency points in real-world communication, prioritizing the use of sub-channels that are both currently clean and historically reliable, thereby improving the consistency of bit error rates in continuous multi-frame transmissions.

[0072] Before the transmission of the third frame, a frequency point marker preamble extension is inserted before the standard preamble, enabling the receiver to synchronously confirm the frequency point and channel response through a single Fast Fourier Transform. Specifically, this includes: reserving a dedicated area of ​​64 symbol lengths at the very beginning of the third frame data packet; transmitting 8 consecutive sine cycles at each of the 8 target frequency points as frequency point markers; the receiver performing a 256-point Fast Fourier Transform on the first 64 symbol length sample stream; extracting the spectral peak positions and mapping them back to the sub-channel number to confirm the actual frequency point used in the current frame.

[0073] This allows the receiver to simultaneously obtain all eight transmitting radio points and their relative channel responses in a single spectrum analysis, eliminating phase discontinuities and estimation delays introduced by multiple frequency switching, and improving demodulation startup speed and amplitude compensation accuracy under multi-frequency switching.

[0074] After three consecutive frames are successfully transmitted, the energy change acceleration of each sub-channel is calculated based on the three-frame spectral energy vectors to generate a smooth transition rule for the frequency hopping sequence. Specifically, this includes: acquiring the spectral energy vector of the third frame after successful demodulation; stacking the three-frame spectral energy vectors vertically to form a perturbation evolution matrix; calculating the first-order and second-order differences for each column of the perturbation evolution matrix to obtain the energy change acceleration of the sub-channel; and applying local neighborhood constraints to the candidate frequency point set based on frequency point stability weighting to generate a smooth transition rule for the frequency hopping sequence.

[0075] This allows the frequency hopping sequence to maintain its disturbance avoidance capability while suppressing disordered frequency jumps between adjacent frames, reducing RF front-end switching stress and receiver acquisition jitter, and enhancing timing stability during multi-frame communication.

[0076] Before the node goes into sleep mode, the current frequency hopping sequence, frequency point stability weighting factor and perturbation evolution matrix are compressed and stored. After waking up, the light perturbation snapshot is loaded and supplemented to restore communication.

[0077] The method also includes performing perturbation snapshot compression and sequence memory solidification, specifically including: performing 8-bit uniform quantization on the frequency point stability weighting factor; performing 8-bit quantization on the range of the energy vectors of the three frames of spectrum; storing the current frequency hopping sequence in the form of an 8-byte integer array, reading the quantized data after the node is woken up, and performing lightweight perturbation sensing within a sub-channel set with low perturbation activity.

[0078] This allows nodes to quickly reconstruct frequency hopping sequences adapted to the current disturbance state without repeating full-band scanning after waking up from low-power sleep. This significantly shortens the initial communication establishment time and ensures the timeliness and continuity of temperature and humidity data reporting in intermittent power supply scenarios.

[0079] On the other hand, this invention proposes a wireless temperature and humidity sensing system for industrial environments, such as... Figure 2 As shown, it includes:

[0080] The on-site electromagnetic disturbance characteristic sensing module is used to collect the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band;

[0081] The frequency hopping sequence generation module is used to sort the sub-channels in ascending order according to the initial spectral energy vector and extract the first 8 sub-channels with the lowest energy to generate the initial frequency hopping sequence.

[0082] The enhanced preamble processing module is used to reconstruct the preamble parsing process at the receiving end, perform fast Fourier transform on the preamble, and weighted accumulate it.

[0083] The disturbance re-sensing module is used to initiate disturbance re-sensing to acquire the spectral energy vector of the second frame after the first frame data packet demodulation is completed;

[0084] The frequency stability assessment module is used to generate a frequency stability weighting factor based on the second frame spectral energy vector and the decision reliability index of each symbol in the first frame.

[0085] The frequency mark preamble extension module is used to insert a frequency mark preamble extension segment before the standard preamble code before the third frame transmission;

[0086] The cross-frame disturbance trend tracking module is used to calculate the energy change acceleration of each sub-channel based on the three-frame spectral energy vector after three consecutive frames of successful transmission.

[0087] The low-power memory solidification module is used to compress and store the current frequency hopping sequence, frequency stability weighting factor and perturbation evolution matrix before the node goes into sleep mode, and load and restore communication after waking up.

[0088] Furthermore, each module in the above system is also used to implement other steps of the aforementioned wireless temperature and humidity sensing method for industrial sites, as detailed below:

[0089] Step 1: Establish a field electromagnetic disturbance characteristic sensing channel and complete an initial frequency band energy distribution snapshot.

[0090] This step begins within the first millisecond-level observation window after the wireless sensor node is powered on and initialized. Its core lies in generating the original basis for subsequent frequency hopping decisions solely based on the real-time response of the radio frequency spectrum, without relying on historical data or pre-set interference models. In industrial settings, the steep rising current change generated when a motor contactor engages excites multi-peak resonant pulse radiation in the 0.1MHz to 300MHz frequency domain; the switching action of the inverter's IGBT forms discrete comb-like spectral lines near integer multiples of the carrier frequency, superimposed with a broadband noise floor. These disturbances are not static but fluctuate in real time with the equipment's operating conditions, therefore, a fixed shielded frequency band approach cannot be used to address them.

[0091] This step abandons the reliance on "interference type identification" or "source location" and instead uses the energy density of the analog baseband signal at the front end of the receiving link as the only observable measure. Without demodulation or unpacking, it completes coarse-grained energy sampling of 25 equally spaced sub-channels in the 2.4GHz ISM band to form the first frame spectral energy vector, which serves as the spatiotemporal coordinate origin of the entire robust transmission process.

[0092] After the node master control unit completes clock synchronization and ADC reference voltage stabilization, the RF transceiver chip enters monitoring mode, activates the on-chip RSSI detection path, and configures the sampling period to 125 microseconds, covering the minimum integer multiple of a complete carrier cycle to ensure phase independence. At this time, the antenna port receives the signal conditioned by the impedance matching network, which is amplified by the low-noise amplifier and then sent to the quadrature downconverter circuit to output two analog baseband signals, I and Q. These two signals do not enter the digital demodulation module, but are directly connected to the on-chip integrated peak-hold analog voltage acquisition unit to record the maximum instantaneous amplitude of the I and Q channels within each 125 microsecond window. Since pulse interference in industrial environments often manifests as amplitude steps rather than frequency shifts, taking the extreme value of the amplitude is more effective in reflecting the transient impact intensity than taking the root mean square.

[0093] Using the maximum amplitude values ​​within 125 microseconds acquired by the I and Q channels respectively as inputs, calculate the composite vector magnitude at that moment, denoted as . ,in Indicates the first Each sampling window is numbered from 1 to 25, corresponding to the center frequency points of 25 sub-channels divided within the range of 2.400 GHz to 2.4835 GHz, with an interval of 3.34 MHz between adjacent center frequencies. This modulus is not normalized and retains its original voltage dimensions; its value directly reflects the degree of disturbance to the sub-channel under the current transient event. Due to slight differences in the response bandwidth of different sub-channel filters, and a slight roll-off in LNA gain at the frequency band edges, therefore... The inherent hardware response non-uniformity is not required and no additional compensation is needed.

[0094] The 25 sub-channels are corresponding Arranged in ascending order of frequency, they form a one-dimensional real number sequence of length 25. This is the first frame of the spectrum energy snapshot; this sequence is not smoothed or truncated, and the peak structure under the original pulse excitation is completely preserved; for example, when an 11kW three-phase asynchronous motor starts, the tens of nanosecond-level voltage change caused by the contactor contact bounce will simultaneously excite voltages more than 6dB higher than the background value on two sub-channels, 2.412GHz and 2.437GHz. Peak values ​​are observed in the sequence, while the remaining channels remain close to the baseline level; this phenomenon is observed in the sequence. The pattern manifests as a local bimodal shape, which serves as the direct basis for subsequent frequency hopping avoidance.

[0095] Will The entire data is written to the designated buffer in the on-chip SRAM and a lightweight interrupt is triggered to notify the main control unit to start the next stage of processing. At this time, the buffer does not store the original sampling points, but rather a set of instantaneous energy responses at 25 defined frequency points, with timestamps indicating these responses. This is the absolute moment when the node completes its first full-band scan after power-on; this timestamp is not involved in subsequent calculations and is only used to identify the time boundary of this frame snapshot; due to the sudden and short-lived nature of disturbances in industrial settings, any data not updated within 10 milliseconds... If the signal fails, this step must be repeated. At this point, the on-site electromagnetic disturbance characteristic sensing channel is officially established, providing irreplaceable localized initial values ​​for subsequent dynamic frequency selection.

[0096] Step 2: Generate an initial frequency hopping sequence based on the first frame snapshot to achieve the initial spatial misalignment between the transmission frequency point and the interference energy peak.

[0097] The previous step obtained Although the sequence consists of single-frame snapshots, it is sufficient to reveal the area of ​​most intense interference at the current moment; if we continue at high... Transmitting on a sub-channel significantly increases the probability of data packets encountering pulse attacks, thus raising the risk of CRC check failures. Therefore, this step no longer waits for multiple sampling statistics but immediately performs the sampling based on... An initial frequency hopping sequence of length N is constructed so that N consecutive transmission actions fall sequentially on the N sub-channels with the lowest energy response, thereby increasing the spatial distance between each transmission and the strongest interference source in the time dimension. This design avoids the dependence on "interference duration prediction" or "pulse repetition period estimation" and completes the first avoidance using only a single observation result, which reduces latency and avoids missing the best communication window due to waiting for multiple frames.

[0098] Perform an ascending sort operation to obtain the sorted index sequence. ,in Indicates the first position in the original sequence Small elements The original position number in the data ranges from 1 to 25; for example, if If the minimum value of the entire sequence is found, then... The index sequence directly corresponds to the arrangement order of the 25 sub-channels from low to high perturbation, without the need to introduce any weighting coefficients or attenuation factors; the sorting process is completed in the on-chip hardware accelerator, taking no more than 3 microseconds, and does not affect real-time performance.

[0099] from Extract the first N elements from the middle to form the initial frequency hopping index sub-column. Where N is a preset constant with a value of 8; this value is determined by the wireless frame structure: a complete temperature and humidity data packet includes a physical layer preamble, synchronization word, payload, and tail bits, requiring a total of 8 symbol cycles to reliably acquire; therefore, the frequency hopping sequence length is set to match it; each It corresponds to a specific sub-channel center frequency point, that is ;thus, This refers to the frequency sequence followed by the eight symbol transmissions of the first data packet.

[0100] Will The frequency hopping control register set of the RF chip is loaded, so that the chip, in the next transmit cycle, follows... The local oscillator frequency is switched sequentially in a specified order. The switching action occurs before the start edge of each symbol, with a frequency settling time of no less than 200 nanoseconds reserved to ensure that the local oscillator signal is stable before the power amplifier is turned on. During this process, the center frequency of the RF front-end filter synchronously follows the local oscillator offset to ensure that the passband always covers the current transmitting RF point. Since all sub-channels are located within the coverage area of ​​the same filter bank, switching does not require the replacement of external filter components, and path reconstruction is completed solely by the on-chip switching matrix.

[0101] Node complete After all 8 frequency switching operations and the transmission of the first data packet, the current frequency hopping sequence status is automatically marked as "activated" and then frozen. Content is frozen, prohibiting any write operations; this freeze action prevents subsequent disturbance perception results from mistakenly overwriting the initial decision; simultaneously, the main control unit reads the actual air propagation time during this transmission process and records it as... This duration is equal to the sum of 8 symbol periods plus the sum of fixed delays introduced by each frequency switching, and its value is constant and does not change with disturbances. It is temporarily stored in the register as the reference time scale for subsequent frequency hopping rhythm synchronization; at this point, the initial frequency hopping sequence is deployed, and the transmission behavior is freed from the high-energy disturbance region for the first time, achieving spatial misalignment.

[0102] Step 3: Inject an enhanced preamble structure at the receiver of the first data packet to improve the reliability of symbol synchronization under impulse interference.

[0103] The previous step avoided the transmission frequency point. While the signal-to-noise ratio (SNR) is within the mid-energy peak region, pulse interference in industrial settings has strong penetrating power and can still couple into the receiving link, causing sampling clock jitter, abnormal automatic gain control, or decision threshold drift, ultimately leading to preamble recognition failure. The traditional 802.15.4 standard preamble only contains a 32-bit fixed pattern, making it difficult to maintain synchronization accuracy when the SNR drops sharply below 5dB. This step does not change the frequency hopping logic at the transmitter end, but reconstructs the preamble parsing process at the receiver end, expanding the original preamble into a composite structure with autocorrelation peak sharpening characteristics and embedding a frequency point marker field synchronized with the frequency hopping sequence. This allows the receiver to simultaneously complete frequency point confirmation and symbol timing estimation in a single FFT analysis, significantly reducing the synchronization time and the probability of exposure to interference windows.

[0104] After the receiving node completes the RF front-end gain configuration and ADC sampling rate lock, the received analog intermediate frequency signal is sent to the digital down-conversion module to generate a complex baseband sampling stream. This sampling stream continuously flows into the ring buffer at a rate of 8 sampling points per symbol. The buffer depth is set to 128 points, which is sufficient to accommodate two complete preamble lengths. When a sudden increase in energy is detected and continues to exceed 3 sampling points, it is determined that the preamble has arrived, and the parsing process is started. This energy detection does not rely on a preset threshold, but adopts the sliding window local variance comparison method, that is, it is triggered when the variance of the current window is greater than 1.8 times the variance of the previous window, adapting to different background noise levels.

[0105] A 128-point complex sample is extracted from the start position of the preamble, and a 128-point Fast Fourier Transform is performed to obtain the frequency domain response. In the transformation result, each Each corresponds to a frequency resolution unit, with a resolution equal to the sampling rate divided by 128. Since the preamble is discretely distributed in the frequency domain after frequency hopping modulation, its energy is mainly concentrated in several adjacent frequency points corresponding to the current transmit sub-channel. To enhance recognition robustness, [the following is omitted as the text is incomplete and requires further context]. Perform weighted summation, weight coefficients Defined as:

[0106] ;

[0107] in The theoretical main lobe center index is derived by working backward from the current expected receiving frequency. The normalization broadening factor is set to 4. Its physical meaning is to make the weights extend by 4 frequency units on both sides of the main lobe and then decay to below 0.5. This weighting does not change the original spectral structure, but only strengthens the contribution of the main energy accumulation region and suppresses misjudgments caused by sidelobe fluctuations.

[0108] Weighted spectrum along Perform reverse cumulative summation on the axis to generate the cumulative energy curve. The curve shows a clear inflection point at the main lobe, because the left side is a high-energy region and the right side is a low-energy region, and the accumulation process amplifies this transition feature; subsequently, Find the first difference, and we get Its peak position is the actual main lobe center. This peak value is less susceptible to narrowband interference than traditional amplitude detection because if the interference only occupies a few frequencies, its impact on the slope of the cumulative curve is minimal; once determined... This allows us to deduce the actual center frequency of the current receiving frequency. ,in For FFT reference frequency, This refers to the frequency resolution.

[0109] in accordance with The corresponding sub-channel number can be obtained by looking up the table. The frequency point is then compared with the expected number of the local frequency hopping sequence. If they match, the frequency point is confirmed to be correct, and the process proceeds to the symbol timing fine estimation stage. At this point, the middle 64 points are extracted from the preamble sampling stream to construct the autocorrelation function. ,in It is a complex sampling sequence. The time delay step is 0 to 32; the autocorrelation function is in The global maximum value is obtained at the point where the rising edge slope directly reflects the clarity of the symbol boundary. When the slope is lower than the preset threshold, it indicates that there is serious inter-symbol interference. At this time, the timing lock is not performed, but the current preamble is discarded and the system waits for the next frame to retry. This judgment logic ensures that data demodulation is only advanced when the synchronization quality meets the standard, avoiding batch errors caused by incorrect synchronization. Thus, the enhanced preamble structure completes closed-loop verification, and the synchronization reliability is substantially improved.

[0110] Step 4: After demodulating the first frame data packet at the receiving end, initiate disturbance re-sensing and online frequency hopping sequence update.

[0111] In the previous step, the receiving end successfully extracted the temperature and humidity values ​​and timestamp carried in the first frame data packet, proving that the current frequency hopping strategy is initially effective; however, industrial site disturbances are time-varying, and the first frame snapshot... Only reflects The time status, and the transmission time of the first frame. During this period, the equipment operating conditions may have changed; for example, if a frequency converter increases its frequency from 50Hz to 60Hz during the first frame transmission, its radiation spectrum will shift accordingly, and some frequency points in the original frequency hopping sequence may fall into the new disturbance peak region; therefore, this step is not satisfied with a single avoidance, but immediately starts the second round of disturbance sensing after the first frame demodulation is completed, and collects a second frame spectrum energy snapshot. And based on this, a second frequency hopping sequence is generated. This enables the closed-loop evolution of the frequency hopping strategy; the update process is seamless and does not interrupt the data flow, and the update timing is strictly anchored to the moment the first frame demodulation is completed, ensuring that the strategy iteration and the evolution rhythm of physical perturbation are consistent.

[0112] After the main control unit confirms that the CRC check of the first frame data packet has passed and the payload has been parsed, it immediately sends a command to the RF chip to restart the RSSI detection path, reset the sampling counter, and clear the previous buffer; at this time, the distance... It's over This refers to the time taken for the first frame to propagate through the air; to maintain time alignment, the start time of this sampling is set to... This synchronization is strictly synchronized with the end of the first frame transmission; this synchronization does not depend on an external clock source, but is determined by the internal timer of the RF chip. Preset values ​​are automatically triggered, eliminating jitter introduced by software scheduling.

[0113] Perform a new round of 25 sub-channel energy sampling with the exact same parameter configuration as in step one to obtain the second frame spectral energy vector. The vector and The structures are the same, but the numerical distributions may differ significantly; for example, if a DC charging station is put into operation during the transmission of the first frame, its high-frequency switching noise will have an additional energy spike around 2.452 GHz. It will be significantly higher than This change requires no manual annotation or threshold determination; it is directly reflected in… In each component value; because the interval between two samplings is only Typically less than 2 milliseconds, therefore Can be regarded as Differential approximation in the direction of perturbation evolution.

[0114] right Perform the same ascending sort operation as in step two to obtain a new index sequence. Then, the first 8 elements are extracted to form the second frequency hopping sequence. This sequence is not related to Any comparisons or fusions are performed, and the data is generated completely independently. Its physical meaning is: for the new state after the perturbation evolution, the eight sub-channels with the lowest energy are reselected as the transmission paths for the next frame; if the perturbation change is weak, then... High degree of overlap; if the disturbance undergoes a sudden change, then Automatically migrates to a new low-disturbance region; this design avoids introducing hysteresis compensation or predictive extrapolation, and all decisions are driven by current measured data.

[0115] Will Load the frequency hopping register group, and at the same time... The status is changed from "activated" to "historical sequence," allowing it to be overwritten by subsequent sequences; all subsequent launch actions are based on this. Execution; to ensure a smooth sequence switch, the following regulations are stipulated. The effective time is the start edge of the first symbol of the next frame's data packet, that is, it is enabled immediately after the current frame is sent; this switch has no handshake process and does not increase protocol overhead; in addition, the master control unit will With corresponding timestamp The data is stored together in a circular historical buffer to reserve an interface for possible long-term trend analysis later; however, this step does not perform any trend analysis, but only completes the data writing to disk; at this point, the frequency hopping sequence has completed its first online update, and the system enters a dynamic adaptation state, laying the foundation for subsequent continuous communication.

[0116] Step 5: Based on the second frame snapshot and the transmission quality feedback of the first frame, generate a frequency stability weighting factor to guide the frequency hopping sequence to shift towards the long-term low-disturbance region.

[0117] In the previous step, Data collection is completed at all times. Deployed and ready for the next frame of transmission; however, relying solely on a single energy snapshot has limitations—some subchannels, although... While the energy level is relatively low, it may be affected by out-of-band leakage due to nearby strong interference peaks, exhibiting time-varying jitter in continuous communication. For example, the 2.425GHz subchannel itself has no obvious radiation source, but there is an old-fashioned PLC high-frequency oscillator located immediately adjacent to 2.422GHz, whose harmonic sidebands intermittently rise in this channel. Such phenomena are difficult to distinguish in a single-frame snapshot, but will manifest as bit error rate fluctuations in demodulation over multiple consecutive frames. Therefore, this step is no longer considered in isolation. Instead, it analyzes the frequency along with the actual received quality of the first frame to construct a weighting factor that reflects the inherent stability of the frequency point. This makes subsequent frequency hopping selection not only focus on the instantaneous energy level, but also favor those sub-channels that maintain low bit error rate performance amidst disturbances.

[0118] The reliability index of each symbol decision recorded by the receiver during the demodulation process of the first frame data packet is extracted and denoted as follows: ,in correspond The Middle The number of symbols carried by each frequency point; this indicator is not a soft-decision likelihood value, but is derived from the dispersion of sampling points on the hard-decision path inside the digital baseband module: the standard deviation is calculated for 8 ADC sampling points within each symbol period. Then take its reciprocal and obtain it by linear scaling. This is a dimensionless normalization constant, with a value of 0.12, to ensure... The value falls in the range of 0.3 to 1.0; The higher the value, the less interference suppression the symbol experiences at the corresponding frequency point, and the clearer the decision boundary.

[0119] Will Its corresponding sub-channel number Establish mapping relationships to form 8 pairs. Then iterate through the 25 sub-channels, for each number... Statistics on its Number of times it appears in and all corresponding ;like Not appeared In the middle, then This statistic does not involve interpolation or extrapolation; it faithfully reflects the actual usage frequency and performance sum of each sub-channel in the first frame. For example, if... ;all Together, they constitute the first frame of frequency point behavior profile.

[0120] Combined with the second frame snapshot For each subchannel Calculate the frequency stability weighting factor Defined as:

[0121] ;

[0122] in The Middle The component is the measured energy response of the subchannel in the second frame; The median value of all elements is used as a dynamic reference benchmark to avoid the adaptability defects caused by a fixed threshold; when hour, Simultaneously reflecting historical performance (average reliability) and current state (degree of energy suppression); when At that time, weights are assigned only based on the current energy response, ensuring that unused sub-channels still have the opportunity to be selected; this segmentation definition guarantees that all All are positive real numbers with consistent dimensions, and can be directly used for sorting.

[0123] 25 according to Sort in ascending order to generate a weighted stability vector. The vector is then sorted in descending order to obtain the index sequence. ,in express The Middle The original position number of the large element; the first 8 elements are extracted from it to form the third frequency hopping sequence. This sequence is the new path that the next frame will follow; at this point, the system has moved from the initial single-point avoidance to a collaborative selection stage that takes into account both historical performance and the current state. After generation, the main control unit synchronously clears the behavior file of the first frame to reserve space for updates after receiving the next frame; at this point, the frequency selection logic has completed a qualitative leap, from static snapshot-driven to spatiotemporal joint-driven.

[0124] Step Six: Inject a frequency point marking preamble before the third frame is transmitted, so that the receiver can simultaneously complete frequency point confirmation and channel estimation in one FFT.

[0125] The previous step generated While it already possesses spatiotemporal joint optimization characteristics, if the receiver still uses the traditional preamble parsing method, it needs to first perform coarse frequency point determination, then switch to the corresponding channel, and finally perform channel estimation. The entire process is time-consuming and easily interrupted by intermediate disturbances; especially when When the interval between adjacent frequency points is small, the phase discontinuity caused by frequency switching will further degrade the channel estimation accuracy. This step, while maintaining the original frequency hopping structure, inserts a fixed-length frequency marker preamble at the very beginning of each data packet frame. This preamble does not carry temperature and humidity information; it consists only of a set of narrowband tones modulated by preset amplitude, and its spectral distribution is similar to... The eight frequency points correspond one-to-one, allowing the receiver to read all eight frequency point numbers used in the current frame simultaneously after a single wide-window FFT, and to invert the channel attenuation characteristics based on the amplitude of each tone, eliminating the need for multiple switching and repeated estimation steps.

[0126] When the node prepares to encapsulate the third frame data packet, the main control unit reads the currently available frequency hopping sequence. And convert it into the corresponding center frequency set. Subsequently, in the physical layer frame structure, a dedicated area of ​​64 symbols is reserved before the standard preamble as a frequency point marking preamble extension segment. This area does not use spread spectrum or coding, but instead transmits 8 consecutive sine cycles at 8 target frequencies with constant power, with each cycle corresponding to one symbol, for a total of 64 symbols. The start times of each tone are strictly aligned to ensure that they remain spectrally separable after being superimposed at the receiver.

[0127] Upon detecting an increase in energy, the receiver does not immediately initiate standard preamble parsing. Instead, it first extracts a sample stream of the first 64 symbols and feeds it into a 256-point Fast Fourier Transform module. Since the transmitter's eight tone frequencies all fall on the center frequencies of the 25 sub-channels, and the bandwidth of each tone is much smaller than the sub-channel spacing, the transformed spectrum... Lieutenant General in 8 specific A sharp peak appears near the value; to enhance noise immunity, [the following is applied]: Perform a sliding median filter with a window width of 5 frequency units to filter out isolated spikes without blurring the main peak; after filtering, retain all peak positions that are 1.6 times higher than the global mean, denoted as the set. ,in The number of valid peak values ​​detected.

[0128] Will Each Mapping back to physical frequency Then look up the table to convert it into a sub-channel number. This set of numbers represents the list of frequency points actually used in the current frame, inferred by the receiver; if And all All belong to If the frequency point marking is complete, the system proceeds to the next step. If some frequency points are missing (e.g., only 6 are detected), a fault tolerance mechanism is activated: for undetected frequency points, the system uses their corresponding frequencies. The average channel gain of adjacent detected frequency points is the initial estimate; this fault tolerance does not change the frequency hopping sequence itself, but only affects the channel compensation accuracy. It is a passive response and does not affect the subsequent process.

[0129] For each confirmed frequency point Extract the peak amplitude of the corresponding tone in the frequency domain. This amplitude is related to the transmit power, path loss, antenna gain, and noise floor, but within the same frame, the transmit power and antenna parameters are constant. The differences mainly reflect the actual attenuation characteristics of each sub-channel at the current moment; Normalization process yields the relative channel response. This group The equalizer coefficient register, which is directly loaded into the digital baseband module, participates in amplitude compensation in real time during the demodulation of the subsequent standard preamble and data symbols. Since the compensation coefficients are ready in the frame header, there is no need to wait for data segment feedback, which greatly shortens the channel adaptation delay. At this point, the frequency point marking preamble extension has completed its mission, and the receiver realizes the integrated acquisition of frequency point confirmation and channel estimation.

[0130] Step 7: After three consecutive frames are successfully transmitted, start cross-frame perturbation trend tracking and generate frequency hopping sequence smooth transition rules.

[0131] The first six steps have achieved single-frame avoidance, double-frame update, three-frame optimization, and four-frame coordination. However, industrial disturbances are not random walks but often exhibit directional evolution. For example, the frequency ramp-up process of a frequency converter is accompanied by a uniform rightward shift of the radiation spectrum, and the temperature rise of the motor causes a slow drift in the center frequency of the switching noise. If the frequency hopping sequence is completely reset each time, the frequency selection may drastically change between adjacent frames, exacerbating the switching stress of the RF front-end and increasing the difficulty of acquisition at the receiver. Therefore, this step does not pursue the absolute optimality of the frequency hopping sequence, but introduces a sequence evolution continuity constraint under the premise of ensuring communication availability, so that... Maintaining a minimum Hamming distance between them to avoid unnecessary oscillations; this constraint is based on the previous three frame perturbation snapshots. The temporal variation characteristics are automatically generated without the need to preset the step size or direction.

[0132] After the third frame is successfully demodulated, the system automatically triggers the third disturbance sensing and captures a snapshot of the third frame. The time of its collection is ,in For the second frame's airborne propagation time, and The same applies, as the frame length and symbol rate remain unchanged; at this point, there are already three frame snapshots: Stack the three items vertically in chronological order to form... 3D perturbation evolution matrix , of which The row corresponds to the first Frame snapshot.

[0133] right Each column (i.e., each sub-channel) Calculate the first difference of its time series, and obtain Then calculate the second difference. Its physical meaning is the acceleration of the energy change in the sub-channel; if This indicates that the channel's interference level is deteriorating rapidly; if It is advised to avoid this in advance.

[0134] Define the smoothness cost function of frequency hopping sequences To measure the strength of change between the old and new sequences:

[0135] ;

[0136] The first term is a linear approximation of the Hamming distance, reflecting the overall offset of the frequency point numbering; the second term is a sequence difference consistency penalty term. The balancing coefficient, set to 0.35, aims to ensure that the frequency jump step size of the new sequence maintains the rhythm of the old sequence as much as possible. This function does not seek a global minimum, but rather... Based on the weighted ranking, a local neighborhood constraint is imposed on the candidate frequency point set: only allowing... The selection is within ±2 of the location; if it exceeds this range, the weighting is automatically reduced. This constraint scale is determined by... The maximum absolute value is used to deduce the dynamic adaptation to the perturbation evolution rate.

[0137] In the weighted stability vector Based on this, for each position Constructing a restricted candidate set Then Selected from The largest one as The selection process proceeds digit by digit without backtracking; if a certain If empty, return to the global state. Take the next unselected number from the sequence; the final generated Maintain and It achieves structural similarity without sacrificing the essential perturbation avoidance mechanism; after generation, the system will... Marked as "reference sequence" Mark as "current sequence" and clear. The matrix makes room for the accumulation of the next three frames; at this point, the frequency hopping behavior is transformed from discrete decision-making into continuous evolution, and the system gains an inherent sense of rhythm.

[0138] Step 8: During the node sleep / wake cycle, perform low-power perturbation snapshot compression and sequence memory persistence to ensure context continuity under intermittent communication.

[0139] In industrial environments, wireless sensor nodes often employ periodic sleep cycles to extend battery life, with a typical wake-up interval of 30 seconds. During this period, the disturbance environment may change significantly. If the frequency hopping sequence is rebuilt from scratch with each wake-up, it will lead to an increased first-frame retransmission rate and an accumulated risk of synchronization failure. Therefore, this step does not rely on continuous online sensing, but rather, before the node enters deep sleep, it retrieves the current optimal frequency hopping sequence. and its supporting evidence -- including the most recent Vector and three-frame perturbation matrix The compressed representation is written to the non-volatile storage area with extremely low overhead; after waking up, instead of re-executing the entire perception and generation process, the solidified context is loaded first, and communication can be restored by supplementing only a lightweight snapshot, which greatly reduces startup latency and power consumption.

[0140] The node completes the transmission and confirmation of the third frame. Once effective, the main control unit initiates the hibernation preparation process; at this time, it reads the current... The vector is subjected to 8-bit uniform quantization, that is, each Mapping to the integer range of 0–255, the mapping relationship is as follows: ,in and for The minimum and maximum values ​​are determined; the quantization preserves the relative ordering relationship, with the error controlled within ±0.4%, meeting the frequency hopping selection accuracy requirements; the quantized sequence... It occupies a total of 25 bytes and is written to a specified sector of the on-chip EEPROM.

[0141] For the three-frame perturbation matrix Instead of storing the original floating-point values, it extracts the range of each column. And then Perform the same 8-bit quantization to obtain This range reflects the perturbation activity of each sub-channel within the observation window, and is more robust than a single-point value; 25 It occupies a total of 25 bytes, and is related to quantization. Adjacent storage; in addition, the acquisition times of the three frames are recorded relative to each other. offset It is stored as a 4-byte integer and used to calibrate the timeline after wake-up.

[0142] Will It is stored directly as an 8-byte integer array, with each element occupying 1 byte; the sequence length is fixed at 8 bytes, requiring no additional length field; the storage location is immediately adjacent to... Then, the four items, totaling 77 bytes, are all stored in a single page of EEPROM. The write operation is completed within 1.2 milliseconds before the hibernation command is issued, ensuring that the data is solidified before power failure. This storage structure has no redundant fields, no checksums, and no file headers. It is purely designed to be the minimum set of information required for recovery, minimizing write power consumption.

[0143] The node enters deep sleep mode, with only the real-time clock and ultra-low-power interrupt controller maintaining power. Upon waking, the main control unit first reads 77 bytes of content from the EEPROM to restore... Quantification Quantitative range And the time offset; then immediately perform a single-frame perturbation sensing to obtain a new snapshot. ;use Select the sub-channel set with the smaller range This snapshot sampling is limited to only The process is performed internally, sampling up to 12 sub-channels; this focused sampling reduces sensing time to half of the original and power consumption by approximately 40%; finally, By fusing with historical quantitative data, a fourth frequency hopping sequence is generated. Thus, the system maintains complete context memory and fast recovery capability under low power consumption constraints, truly achieving sustainable and robust communication for industrial sites.

[0144] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A wireless temperature and humidity sensing method for industrial sites, characterized by, include: Establish an on-site electromagnetic disturbance characteristic sensing channel and collect the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band; Based on the initial spectral energy vector, the sub-channels are arranged in ascending order, and the first 8 sub-channels with the lowest energy are selected to generate the initial frequency hopping sequence; The preamble parsing process is reconstructed at the receiving end, and the preamble is subjected to fast Fourier transform and weighted accumulation to improve the reliability of symbol synchronization under impulse interference. After the first frame data packet is demodulated, the disturbance resensing is initiated to acquire the second frame spectrum energy vector and generate the second frequency hopping sequence. Specifically, this includes: after the first frame data packet is demodulated, a command is sent to the RF chip to restart the received signal strength indication detection path; a new round of 25 sub-channel energy sampling is performed with the same parameter configuration as the initial spectrum energy vector; the second frame spectrum energy vector is sorted in ascending order, the first 8 elements are extracted to form the second frequency hopping sequence, the second frequency hopping sequence is loaded into the frequency hopping register group, and the initial frequency hopping sequence is replaced. Based on the spectral energy vector of the second frame and the decision reliability index of each symbol in the first frame, a frequency point stability weighting factor is generated, and a third frequency hopping sequence is generated. Before the third frame is transmitted, a frequency point marker preamble extension is inserted before the standard preamble, so that the receiver can synchronously confirm the frequency point and channel response through a single fast Fourier transform. After three consecutive successful transmissions, the energy change acceleration of each sub-channel is calculated based on the three-frame spectral energy vectors to generate a smooth transition rule for the frequency hopping sequence. Specifically, this includes: acquiring the spectral energy vector of the third frame after successful demodulation; stacking the three-frame spectral energy vectors vertically to form a perturbation evolution matrix; calculating the first-order and second-order differences for each column of the perturbation evolution matrix to obtain the energy change acceleration of the sub-channel; and applying local neighborhood constraints to the candidate frequency point set based on frequency point stability weighting to generate a smooth transition rule for the frequency hopping sequence. Before a node goes into hibernation, the current frequency hopping sequence, frequency point stability weighting factor, and perturbation evolution matrix are compressed and stored. After waking up, the node loads and supplements a lightweight perturbation snapshot to restore communication.

2. The wireless temperature and humidity sensing method for industrial field according to claim 1, wherein, The establishment of the on-site electromagnetic disturbance characteristic sensing channel includes: The RF transceiver chip enters monitoring mode and activates the received signal strength indication detection path. The maximum instantaneous amplitude of the I-channel and Q-channel analog baseband signals was collected within a 125-microsecond window. The combined vector magnitude of the maximum instantaneous amplitude of the I-channel and Q-channel was calculated to obtain the energy response of 25 sub-channels. The energy responses of the 25 sub-channels are arranged in ascending order of frequency points to form the first frame of spectrum energy snapshot and written to the buffer.

3. The wireless temperature and humidity sensing method for industrial field according to claim 1, wherein, The step of arranging the sub-channels in ascending order based on the initial spectral energy vector and extracting the first 8 sub-channels with the lowest energy to generate the initial frequency hopping sequence includes: The first frame of the spectral energy snapshot is sorted in ascending order to obtain a sorted index sequence. The first 8 elements of the sorted index sequence are used to form the initial frequency hopping index sub-column; The initial frequency hopping index sub-column is loaded into the frequency hopping control register group of the RF chip, and the air propagation time is recorded as the reference time scale for subsequent frequency hopping rhythm synchronization.

4. The wireless temperature and humidity sensing method for industrial field according to claim 1, characterized in that, The process of reconstructing the preamble parsing at the receiving end includes: The received analog intermediate frequency signal is used to generate a complex baseband sampling stream and flow into a ring buffer; A 128-point Fast Fourier Transform is performed on the preamble sample to obtain the frequency domain response; The frequency domain response is weighted and accumulated, with the weighting coefficients decreasing with frequency shift. The weighted spectrum is then summed in reverse and subjected to first-order difference to determine the actual main lobe center position.

5. The wireless temperature and humidity sensing method for industrial field according to claim 1, characterized in that, The second frame of spectrum energy vector is acquired by initiating disturbance resensing.

6. The wireless temperature and humidity sensing method for industrial field according to claim 1, wherein, The generated frequency point stability weighting factor includes: The symbol decision reliability index is extracted from the demodulation process of the first frame data packet. It is obtained by scaling the inverse of the standard deviation of the sampling points. The symbol decision reliability index is then mapped to the corresponding sub-channel number. Combined with the second frame spectral energy vector, calculate the frequency point stability weighting factor for each sub-channel; sort the frequency point stability weighting factors in descending order, and extract the first 8 elements to form the third frequency hopping sequence.

7. The wireless temperature and humidity sensing method for industrial field according to claim 1, characterized in that, The insertion of a frequency point marker preamble extension segment before the standard preamble includes: A dedicated area of ​​64 symbols is reserved at the very beginning of the third frame data packet; Eight consecutive sine cycles are transmitted at each of the eight target frequencies as frequency markers; The receiver performs a 256-point Fast Fourier Transform on the first 64 symbol-length sampled stream; Extract the peak position of the spectrum and map it back to the sub-channel number to confirm the actual frequency point used in the current frame.

8. The wireless temperature and humidity sensing method for industrial field according to claim 1, characterized in that, The smooth transition rule for generating frequency hopping sequences.

9. The wireless temperature and humidity sensing method for industrial field according to claim 1, wherein, The method further includes performing perturbation snapshot compression and sequence memory solidification, specifically including: Perform 8-bit uniform quantization on the frequency stability weighting factor; Perform 8-bit quantization on the range of the energy vectors of the three frames; The current frequency hopping sequence is stored as an 8-byte integer array. After the node is woken up, the quantized data is read and lightweight disturbance sensing is performed within a sub-channel set with low disturbance activity.

10. An industrial field oriented wireless temperature and humidity sensor system for implementing the method according to any one of claims 1 to 9, characterized by include: The on-site electromagnetic disturbance characteristic sensing module is used to collect the initial spectral energy vectors of 25 sub-channels in the 2.4 GHz band; The frequency hopping sequence generation module is used to sort the sub-channels in ascending order according to the initial spectral energy vector and extract the first 8 sub-channels with the lowest energy to generate the initial frequency hopping sequence. The enhanced preamble processing module is used to reconstruct the preamble parsing process at the receiving end, perform fast Fourier transform on the preamble, and weighted accumulate it. The disturbance re-sensing module is used to initiate disturbance re-sensing to acquire the second frame's spectral energy vector after the first frame's data packet demodulation is completed. Specifically, it includes: sending a command to the RF chip to restart the received signal strength indication detection path after the first frame's data packet demodulation is completed; performing a new round of 25 sub-channel energy sampling with the same parameter configuration as the initial spectral energy vector; sorting the second frame's spectral energy vector in ascending order, extracting the first 8 elements to form the second frequency hopping sequence, loading the second frequency hopping sequence into the frequency hopping register group, and replacing the initial frequency hopping sequence. The frequency stability assessment module is used to generate a frequency stability weighting factor based on the second frame spectral energy vector and the decision reliability index of each symbol in the first frame. The frequency mark preamble extension module is used to insert a frequency mark preamble extension segment before the standard preamble code before the third frame transmission; The cross-frame perturbation trend tracking module is used to calculate the energy change acceleration of each sub-channel based on the three-frame spectral energy vectors after three consecutive successful transmissions. Specifically, it includes: acquiring the spectral energy vector of the third frame after successful demodulation; stacking the three-frame spectral energy vectors vertically to form a perturbation evolution matrix; calculating the first-order and second-order differences for each column of the perturbation evolution matrix to obtain the energy change acceleration of the sub-channel; and applying local neighborhood constraints to the candidate frequency point set based on frequency point stability weighting to generate a smooth transition rule for frequency hopping sequences. The low-power memory solidification module is used to compress and store the current frequency hopping sequence, frequency stability weighting factor and perturbation evolution matrix before the node goes into sleep mode, and load and restore communication after waking up.

Images