An intelligent baby crib monitoring system based on wireless transmission

Abstract

The present invention discloses an intelligent crib monitoring system based on wireless transmission, which relates to the technical field of infant physiological monitoring. The system includes the following steps: dynamically allocating node communication time slots through wireless networking, adjusting the laser array tilt angle and activating a dual-band metamaterial antenna to synchronously collect light intensity fluctuation signals and electrocardiogram signals to generate a mixed signal stream; extracting physiological features from the mixed signal using adaptive filtering and noise joint analysis; generating clean respiratory waveforms and electrocardiogram waveforms through wavelet decomposition and dynamic threshold suppression; identifying respiratory and heart rate anomalies based on zero-crossing point detection and R-wave energy verification, and determining transmission priority by combining chaotic key generation and a Markov decision model. The present invention achieves hardware-level synchronous acquisition of light intensity fluctuation signals and electrocardiogram signals through multi-band collaborative networking and dynamic time slot allocation, and controls signal synchronization errors by combining dual-channel noise capture and differential common-mode suppression of the metamaterial antenna.

Description

Technical Field

[0001] The present invention relates to the technical field of infant physiological monitoring, in particular to an intelligent baby crib monitoring system based on wireless transmission. Background Art

[0002] In recent years, infant health monitoring technology has evolved from single-sensor monitoring (such as piezoelectric film respiration detection and photoplethysmography) to multimodal fusion. Existing physiological signal acquisition systems based on wireless body area networks (WBANs) have achieved preliminary synchronized monitoring of ECG and respiration, with data transmission via Bluetooth or ZigBee protocols. However, these systems generally utilize fixed-frequency communication and static encryption strategies, resulting in bottlenecks in multi-node coordination, dynamic anti-interference, and real-time anomaly response.

[0003] The main shortcomings of existing technologies are concentrated in two aspects: first, the synchronous acquisition of multimodal signals is limited by hardware-level clock deviation and radio frequency interference, and the time domain alignment error between the light intensity fluctuation signal and the electrocardiogram signal often leads to misjudgment of the respiratory cycle; second, traditional encryption mechanisms (such as AES static encryption) and fixed-priority transmission strategies have difficulty balancing security and real-time performance in high-concurrency data flow scenarios. Key update delays and transmission conflicts may exacerbate system response delays. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an intelligent baby crib monitoring system based on wireless transmission to solve the problems of insufficient synchronous acquisition accuracy of multi-modal physiological signals and low safety transmission efficiency.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides an intelligent baby crib monitoring system based on wireless transmission, which comprises:

[0008] The network acquisition module dynamically allocates node communication time slots through wireless networking, adjusts the laser array tilt angle and activates the dual-band metamaterial antenna to synchronously collect light intensity fluctuation signals and electrocardiogram signals to generate a mixed signal stream;

[0009] The signal optimization module uses adaptive filtering and noise joint analysis to extract physiological features from mixed signals, and generates clean respiratory and ECG waveforms through wavelet decomposition and dynamic threshold suppression;

[0010] The abnormality decision module identifies respiratory and heart rate abnormalities based on zero-crossing point detection and R-wave energy verification, and determines transmission priority by combining chaotic key generation and Markov decision model;

[0011] The transmission control module triggers multi-band coordinated transmission strategies based on priority, implements differentiated encryption and dual-link protocol encapsulation, and establishes a time-slotted response and dynamic retransmission mechanism;

[0012] The data fusion module uses the extended Kalman filter to fuse multi-node observation data and generates bed posture adjustment instructions through the model predictive controller;

[0013] The intervention execution module, based on the fuzzy logic controller, triggers the acoustic-optical-mechanical composite intervention in a hierarchical manner to execute the bed posture adjustment instructions.

[0014] As a preferred solution of the wireless transmission-based intelligent crib monitoring system of the present invention, the generating of the mixed signal stream includes the following steps:

[0015] Dynamically allocate node communication time slots through wireless networking, adjust the tilt angle of the quantum cascade laser array, synchronously activate the dual-band metamaterial antenna, use the laser pulse reflection signal to generate a light intensity sequence, and simultaneously collect electrocardiogram signals and environmental noise;

[0016] Common-mode interference is eliminated through differential amplification, and transmission delay is compensated by a sliding window to generate a mixed signal stream.

[0017] As a preferred solution of the intelligent baby crib monitoring system based on wireless transmission of the present invention, wherein: the physiological feature extraction of the mixed signal includes the following steps:

[0018] Inputting the mixed signal stream into a low-noise amplifier to eliminate signal reflection interference and generate an amplified analog signal stream;

[0019] The amplified analog signal stream is inputted into the respiratory signal channel processing branch and the electrocardiogram signal channel processing branch in parallel;

[0020] Filtering the respiratory signal of the respiratory signal channel processing branch to generate a stable respiratory simulation signal;

[0021] Perform anti-interference processing on the ECG signal of the ECG signal channel processing branch and output an ECG analog signal with normalized amplitude.

[0022] As a preferred solution of the intelligent crib monitoring system based on wireless transmission described in the present invention, the wavelet decomposition and dynamic threshold suppression refer to using Daubechies wavelet to decompose the respiratory waveform and the electrocardiogram waveform, jointly analyzing the noise characteristics based on the high-frequency detail layer energy distribution, dynamically suppressing the interference components through an adaptive threshold algorithm, and reconstructing a clean respiratory waveform that retains physiological characteristics and an electrocardiogram waveform with a complete QRS complex.

[0023] As a preferred solution of the wireless transmission-based intelligent crib monitoring system of the present invention, the identification of abnormal breathing and heart rate includes the following steps:

[0024] The respiratory cycle is detected based on the zero-crossing point, and a respiratory abnormality event is triggered when the instantaneous frequency in multiple consecutive sliding windows exceeds the safe range;

[0025] The R-wave position is identified through first-order difference and energy verification, and a dynamic heart rate curve is constructed. When multiple consecutive data points of the smoothed heart rate exceed the safety limit, an abnormal heart rate event is triggered.

[0026] As a preferred solution of the wireless transmission-based smart crib monitoring system of the present invention, the triggering multi-band coordinated transmission strategy includes the following steps:

[0027] The microsecond time component is used to generate the Lorenz chaotic initial key, and the dynamic key stream is generated through the fourth-order Runge-Kutta method iteration.

[0028] Combined with the Markov decision model, the abnormal state is mapped into a four-dimensional state vector, and the multi-band collaborative transmission strategy is triggered according to the transmission priority weight ratio.

[0029] As a preferred solution of the wireless transmission-based intelligent crib monitoring system of the present invention, the transmission strategy refers to setting a high priority threshold as , the medium priority threshold is , when the breathing weight is greater than or equal to When the high frequency band orthogonal frequency division multiplexing is activated, when the ECG weight is When the link quality is within the range of , the direct sequence spread spectrum of the relay node is enabled, otherwise the narrowband sleep mode is used to transmit environmental data, and the modulation order and spreading code rules are dynamically adjusted based on the link quality.

[0030] As a preferred solution of the wireless transmission-based intelligent crib monitoring system of the present invention, the implementation of differential encryption and dual-link protocol encapsulation, and the establishment of a time-slotted response and dynamic retransmission mechanism include the following steps:

[0031] Bind data slices with chaotic keys at the symbol level, bit level, and phase level to perform differentiated encryption operations;

[0032] Encrypted data is encapsulated through a dual-link protocol, a cyclic redundancy check code is added to the main link, control instructions are transmitted on the backup link, and dynamic retransmission is triggered based on a slotted response mechanism.

[0033] As a preferred solution of the intelligent baby bed monitoring system based on wireless transmission described in the present invention, the generation of bed posture adjustment instructions refers to using an extended Kalman filter to fuse multi-node observation data, constructing a latent state transfer model of respiratory frequency and heart rate, generating bed tilt angle adjustment instructions through a model predictive controller, and injecting the adjustment amount into the Kalman filter to update the state estimation.

[0034] As a preferred solution of the intelligent baby crib monitoring system based on wireless transmission described in the present invention, the execution of the bed posture adjustment instruction refers to triggering sound and light soothing, mechanical vibration and robotic arm patting actions based on the abnormal breathing level classification, closing the laser and antenna receiving channels during the execution of the bed posture adjustment instruction, and restarting data collection after the posture adjustment is completed to compensate for the state estimation loss during the interruption period.

[0035] The beneficial effects of the present invention are: through multi-band collaborative networking and dynamic time slot allocation, hardware-level synchronous acquisition of light intensity fluctuation signals and electrocardiogram signals is realized, and the signal synchronization error is controlled by combining the dual-channel noise capture and differential common mode suppression of the metamaterial antenna; the adaptive threshold algorithm based on the joint analysis of wavelet decomposition and statistical noise characteristics can dynamically distinguish motion artifacts from respiratory characteristic signals, thereby improving the suppression rate of respiratory waveform baseline drift, while retaining the energy integrity of the QRS wave group and avoiding the effective signal loss caused by the fixed threshold algorithm; through chaotic key generation and Markov decision model, dynamic mapping of abnormal events and transmission priority is realized. In the abnormal respiratory state, the orthogonal frequency division multiplexing strategy is adopted to compress the data transmission delay, and the anti-interception capability is enhanced through phase rotation encryption to ensure the end-to-end security of key physiological data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0037] Figure 1 This is a flow chart of mixed signal generation in this embodiment.

[0038] Figure 2 This is a flowchart of physiological signal optimization in this embodiment.

[0039] Figure 3 This is a flowchart of abnormal decision-making and transmission control in this embodiment.

[0040] Figure 4 This is a flowchart of data fusion and intervention execution in this embodiment. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0042] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0043] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0044] In this embodiment, refer to Figures 1 to 4 , this embodiment provides a smart baby crib monitoring system based on wireless transmission, comprising the following steps:

[0045] The quantum cascade laser array is activated, and a mesh network topology is established through the wireless networking coordination module, dynamically allocating node communication time slots. The laser array tilt is precisely adjusted under a temperature feedback control system to ensure that the beam covers the infant's chest area. The dual-band metamaterial RF antenna within the wireless network (for example, 2.4 GHz and 5.8 GHz) is synchronously activated to calibrate the antenna pattern and synchronize phases between network nodes. The master FPGA broadcasts hardware-level synchronization pulses through the wireless network, and the laser array and the network node antennas synchronously collect data. The laser captures light intensity fluctuations, emitting multiple light pulses during each sampling cycle. The reflected light signal is received by the avalanche photodiode (APD), generating a light intensity sequence that contains breathing characteristics. Simultaneously, the metamaterial antenna's dual channels synchronously collect ECG signals and environmental interference noise (full frequency band). After common-mode interference is eliminated by the differential amplifier, the light intensity sequence, ECG signals, and environmental interference noise are output as two analog signals. The wireless network time synchronization engine uses a sliding window algorithm to match the timestamp differences between the two analog signals, dynamically compensate for transmission delays, and output a time-synchronized mixed signal stream.

[0046] To further illustrate, the dual-channel metamaterial antenna includes a main channel and an auxiliary channel. The main channel (e.g., 2.4 GHz) activates the H-shaped resonator array and loads the pre-stored ECG signal reception mode. In this embodiment, the bandwidth is set to 0.5-40 Hz and the gain is adjusted to 20 dB. The auxiliary channel (e.g., 5.8 GHz) starts the full-band scanning mode to dynamically capture environmental electromagnetic noise (interference sources such as Wi-Fi and Bluetooth), and the sensitivity is set to -120 dBm.

[0047] The mixed signal stream is input into a low-noise amplifier (LNA) to set the gain and noise figure; input impedance matching is achieved through an adjustable resistor network to eliminate signal reflection interference and generate an amplified analog signal stream; the amplified analog signal stream enters the respiratory signal channel processing branch for filtering processing, and is simultaneously input into the ECG signal channel processing branch for anti-interference processing.

[0048] It is further explained that the filtering process of the respiratory signal channel processing branch includes pre-filtering, main bandpass filtering and post-stage conditioning.

[0049] Pre-filtering refers to the use of a second-order Butterworth high-pass filter to filter out low-frequency motion artifacts caused by the baby turning over; main bandpass filtering refers to the use of a cascaded fourth-order Chebyshev bandpass filter to suppress high-frequency noise caused by ambient light fluctuations; post-conditioning refers to the use of an automatic gain control (AGC) circuit to dynamically adjust the amplitude of the amplified analog signal to ensure the stability of the output peak and generate a stable breathing analog signal.

[0050] The ECG signal channel processing branch includes power frequency notching, dynamic bandpass filtering and common mode suppression.

[0051] Power frequency notching refers to the use of double-T active filtering to eliminate power line interference; dynamic bandpass filtering refers to the use of programmable switched capacitor filtering to adaptively suppress myoelectric noise; common-mode suppression refers to the elimination of electrode contact noise through instrument amplification and the output of an amplitude-normalized ECG analog signal.

[0052] The stable respiratory analog signal is quantized, and modulation technology is used to improve the effective resolution to generate a digital respiratory waveform; the noise floor of the ECG analog signal is reduced by combining oversampling technology, and a digital ECG waveform with effective resolution is output; the digital respiratory waveform is decomposed into five levels using the fourth-order Daubechies wavelet basis to separate the high-frequency detail coefficients containing noise and the low-frequency approximate coefficients reflecting physiological characteristics; the digital ECG waveform is simultaneously decomposed into six levels, and the intermediate frequency component reflecting the QRS wave group is retained; based on the energy distribution of the high-frequency detail layer after decomposition, combined with the joint analysis of the noise characteristics between wireless networking nodes, the noise area is identified, and the intensity of the motion artifact in the digital respiratory waveform and the myoelectricity in the digital ECG waveform are quantified by statistical methods. The activity level of the interference is measured to generate a noise characteristic distribution map; an adaptive threshold algorithm based on statistical characteristics is used to dynamically suppress the noise characteristic distribution map, and a soft threshold is used to cut off the high-frequency interference components in the digital respiratory waveform. At the same time, a protection mechanism is introduced to the QRS wave group characteristic area in the digital ECG waveform to avoid effective signal loss; after completing the noise suppression, the approximate coefficients that retain the effective physiological components and the optimized detail coefficients are selected to reconstruct the signal to generate a clean respiratory waveform with baseline drift removed and an ECG waveform with a complete QRS wave group; the clean respiratory waveform with baseline drift removed and the ECG waveform with a complete QRS wave group are superimposed with physiological status markers (such as apnea event markers and arrhythmia warning signs) to generate a pure physiological signal.

[0053] It should be noted that communication time slots are dynamically allocated through the Mesh network topology to reduce the probability of multi-node transmission conflicts; the inclination angle is adjusted based on laser temperature feedback to ensure that the optical pulse covers the infant's chest area and eliminate signal loss caused by body position deviation; the main channel focuses on ECG signal acquisition, and the auxiliary channel captures full-band environmental noise to achieve synchronous identification of interference sources; common-mode interference (such as power line noise) is eliminated through differential amplifiers to retain effective physiological signal characteristics; a second-order Butterworth high-pass filter (cutoff frequency 0.1Hz) effectively suppresses low-frequency interference caused by the infant's turning over; a dual-T active notch filter (center frequency 50 / 60Hz) suppresses power line interference and retains the QRS complex; hard threshold protection is used for the QRS complex area (intermediate frequency component) to avoid effective signal loss; and the respiratory waveform is reconstructed to eliminate baseline offset caused by infant movement.

[0054] The respiratory cycle of the clean respiratory waveform in the pure physiological signal is identified by the zero-crossing point detection method. Specifically, when the waveform crosses the zero axis from the negative direction, it is marked as the starting point of inspiration, and when the waveform crosses the zero axis from the positive direction, it is marked as the starting point of exhalation. The time difference between adjacent inhalation starting points is the length of a single respiratory cycle, and a continuous respiratory cycle sequence is obtained; a sliding time window is established to calculate the instantaneous frequency.

[0055] To further illustrate, in this embodiment, a window is set to 5 cycles. When the instantaneous frequency is lower than 20 beats / minute or higher than 60 beats / minute for 3 consecutive windows (such as 15 cycles), a respiratory abnormality event is triggered, and the abnormality start timestamp and frequency deviation value are recorded; a first-order difference is performed on the ECG waveform of the complete QRS complex, and when the difference value exceeds 3 times the standard deviation of the baseline noise, the candidate R wave (the maximum positive peak generated during rapid ventricular depolarization) is marked; the signal energy in the 50ms interval before and after the candidate R wave is calculated, and the artifact points with sudden energy changes are eliminated to obtain the R wave position sequence; a dynamic heart rate curve is constructed based on the R wave sequence, the instantaneous heart rate is calculated, and a moving average is performed on the heart rate data in the past 10 seconds to obtain a smoothed heart rate curve; when the smoothed heart rate is lower than 100 beats / minute or higher than 160 beats / minute for 5 consecutive data points (≥10 seconds), an abnormal heart rate event is triggered, and the abnormality start time and the maximum heart rate deviation value are recorded.

[0056] It should be noted that the error in respiratory cycle detection is reduced through double verification of the waveform's negative zero crossing (starting point of inhalation) and positive back-crossing (end point of exhalation); the baseline fluctuation interference caused by the baby's movement is eliminated based on the pre-processed clean respiratory waveform; and the transient interference misjudgment is avoided through continuous judgment of over-limit triggering.

[0057] The microsecond component of the current time is obtained and the SHA-256 hash function is used to generate a 256-bit entropy source. The last 24 bits are truncated as the initial conditions of the Lorenz chaotic key sequence to complete the hard-coded loading of the chaotic key parameters. The initial conditions are iteratively calculated using the fourth-order Runge-Kutta method to generate a continuous chaotic variable sequence. Each chaotic variable is normalized and the valid bits are extracted to obtain a chaotic variable fragment. The chaotic variable fragments are then bit-wise XORed to generate a dynamic key block.

[0058] To further illustrate, the bitwise XOR obfuscation includes preliminary concatenation, nonlinear obfuscation and cyclic shift.

[0059] Initial splicing is to combine chaotic variable fragments into a key block; nonlinear obfuscation is to replace each byte in the key block through S-box permutation (using AES's S-Box); circular shift is to shift the key after nonlinear obfuscation left by 3 bits.

[0060] A double-buffered queue is established to alternately store newly generated key blocks. When the main control module issues an encryption request, the key is extracted from the active buffer in FIFO order. At the same time, the last 4 bits are randomly perturbed (the last 4 bits of each byte are replaced) through the hardware true random number generator, ultimately generating a dynamic Lorenz encryption key stream.

[0061] The triggered respiratory and heart rate abnormality event states (normal, respiratory abnormality, heart rate abnormality, and combined abnormality) are mapped to the four-dimensional state variables of the Markov decision model; based on the activation of the four-dimensional state vector, a three-dimensional action space is established, which includes three actions: prioritizing the transmission of respiratory data, prioritizing the transmission of ECG data, and balanced transmission. In addition, an immediate reward rule is defined based on the degree of matching between the action and the state. For example, when the priority transmission of respiratory data is selected in the respiratory abnormality state, a reward value of +5 is given to form a reward matrix; a numerical integration method is used to approximate the continuous time value function, and the future time axis is divided into discrete intervals; based on the historical state transition data (such as the past The algorithm uses the following steps: (1) the average time it takes for abnormal breathing to return to normal within 1 hour) and counts the frequency of transitions within discrete intervals to generate a state transition rate matrix; (2) the immediate benefits provided by the reward matrix and the cumulative expected benefits of each state-action pair under the discount factor provided by the state transition rate matrix; (3) an initial weight matrix is ​​set, and the marginal benefits of respiratory and ECG data are calculated based on the cumulative expected benefits; (4) the weight ratio of the transmission priority of the infant's respiratory and ECG data is quantified based on the marginal benefits; (5) the infant's crying status (crying mode and quiet mode) is obtained through an acoustic sensor, and the quantified weight ratio is dynamically corrected to generate a priority weight vector.

[0062] It should be noted that by using the SHA-256 hash function to generate the initial key and adopting the Lorenz chaotic system and the fourth-order Runge-Kutta method to iteratively generate the dynamic key block, the complexity and unpredictability of the encryption algorithm are increased, and the security of data transmission is improved; the Markov decision process (MDP) is used to determine the optimal transmission strategy for respiratory data and electrocardiogram data, which can intelligently adjust the priority of data transmission according to the current status of the infant (normal, abnormal breathing, abnormal heart rate or combined abnormalities), thereby ensuring the timely transmission of critical information in emergency situations; the design of a double-buffered queue and the disturbance injected by the hardware true random number generator ensure the dynamic update and real-time nature of the key stream, which helps to quickly respond to changes in the infant's status.

[0063] Based on the priority weight vector, the weight ratio is analyzed and the priority is divided to trigger the wireless network multi-band cooperative transmission strategy. In this embodiment, a high priority threshold is set. Greater than or equal to 0.6, medium priority threshold The interval range is [0.3, 0.6). The specific operations are as follows:

[0064] When the respiration weight is ≥0.6, high priority is triggered, the wireless network high-band OFDM is activated, non-essential RF channels are shut down, OFDM parameters are loaded, multiple subcarriers are configured and specific intervals are set, high-order quadrature amplitude modulation is used, and the respiration data is converted into a time-domain OFDM baseband signal through inverse fast Fourier transform.

[0065] When 0.3≤ECG weight<0.6, the medium priority is triggered, and the direct sequence spread spectrum of the wireless networking relay node is enabled. Gold code is used to enhance anti-interference capabilities. The lowest latency path within 3 hops is selected through network topology optimization. A sequence code of a specific length is used for direct sequence spread spectrum. The phase modulation symbols are logically operated with the spread spectrum code sequence to generate a baseband chip stream that is resistant to narrowband interference.

[0066] When the trigger conditions of high priority and medium priority are not met, the wireless networking narrowband sleep mode is enabled to perform default allocation of low-priority environmental data, and a semi-static scheduling strategy is implemented in the narrowband channel. The environmental data is encapsulated into fixed-length data packets and differential phase modulation is performed at periodic intervals.

[0067] The modulated baseband signals of each mode are input into the multi-channel RF front-end of the metamaterial antenna. The high-frequency channel uses a power amplifier to boost the signal to a high power level, the mid-frequency channel maintains a medium power level, and the narrowband channel limits the power spectral density.

[0068] It is further explained that during the RF signal transmission process, the link quality indicators fed back by the receiving end are collected in real time. When it is detected that the quality difference of the orthogonal frequency division multiplexing subcarriers exceeds the difference judgment critical value, the adaptive modulation algorithm is triggered to reduce the modulation order of the low-quality subcarriers; when the direct sequence spread spectrum link encounters interference, the spread spectrum code generation rules are dynamically switched to enhance the anti-interference capability; when the narrowband transmission fails continuously, the scheduling cycle is shortened and the transmission power is increased.

[0069] It should be noted that different types of data (respiration, ECG and environmental data) are prioritized based on the priority weight vector to ensure that key information (such as data when breathing or ECG is abnormal) can be transmitted with higher efficiency and quality, thereby optimizing the use of wireless network resources; orthogonal frequency division multiplexing (OFDM), direct sequence spread spectrum (DSSS) and Gold code are used to enhance the system's anti-interference capability and stability, especially in complex electromagnetic environments, which can effectively avoid signal loss or errors; through measures such as adaptive modulation algorithms, dynamic switching of spread spectrum code generation rules and adjustment of transmission parameters according to link quality, the reliability and efficiency of data transmission are further improved, ensuring stable data communication even under adverse conditions; different RF channel power settings are used for data of different priorities, which not only helps to extend the battery life of the device, but also reduces unnecessary energy consumption, while ensuring the efficient transmission of high-priority data.

[0070] The high-frequency orthogonal frequency division multiplexing baseband signal, the mid-frequency direct sequence spread spectrum code chip stream and the narrowband environmental data generated by the RF front-end processing are divided into data slice units, and the data slice units are integrated to form standardized data blocks to be encrypted; the standardized data blocks to be encrypted are bound to the dynamic Lorenz encryption key stream, in which the high-frequency data slices are aligned with the chaotic key at the byte level, the mid-frequency code chip stream is matched with the key at the bit level, and the narrowband data symbols are associated with the key phase to obtain time-synchronized key-data pairs; the dynamic Lorenz key stream is obtained through the wireless networking key distribution center to implement differentiated encryption.

[0071] To further explain, the differentiated encryption operation refers to performing symbol-level nonlinear replacement on high-frequency band signals, implementing logical operation confusion on mid-frequency band code streams, and implementing phase rotation encryption on narrowband data to generate encrypted RF baseband signals.

[0072] The encrypted RF baseband signal is processed by the dual-link protocol encapsulation module. The main link is transmitted through multiple paths through the wireless networking backbone nodes, encapsulating the high-frequency and mid-frequency encrypted data and appending a cyclic redundancy check code. The backup link uses the Bluetooth Mesh of the networking edge node to encapsulate the narrowband control instructions and link status identification to obtain a dual-link protocol data frame with an integrity verification tag.

[0073] It is further explained that a time slotted response mechanism is established during the main link transmission process, and the receiving end returns an encrypted confirmation signal within a predetermined time window. The critical condition judgment value for triggering the backup link retransmission mechanism is set based on the comprehensive channel quality, historical statistics and system requirements. When it is detected that the number of consecutive unconfirmed time slots exceeds the critical judgment value, the backup link retransmission control process is triggered, and the lost data frame number is sent through the backup channel first and the RF channel resources are switched synchronously.

[0074] The receiving end performs layered decryption verification on the dual-link protocol data frames with integrity verification tags. The main link data is reversely restored to the original physiological signal through the chaotic key stream. The backup link instructions are decrypted and drive the power control module, which outputs the decrypted multimodal physiological signal to execute the control instructions.

[0075] It is further explained that when a hash value conflict of the primary and backup link data is detected, the arbitration mechanism is started to give priority to the backup link data and mark the security event to trigger the chaotic key regeneration module to update the dynamic key sequence. A forward security protection mechanism is performed during the encryption process, and the last key perturbation factor of the current encryption operation is injected into the subsequent chaotic iteration process. At the same time, the used key storage area is physically erased.

[0076] It should be noted that the security of data transmission is enhanced by binding different types of data fragments generated by the RF front-end processing with dynamic Lorenz encryption key streams and implementing differentiated encryption. A dual-link protocol encapsulation module is used to process the encrypted RF baseband signal, and a cyclic redundancy check code is added to the main link to ensure the integrity of data transmission. At the same time, the backup link mechanism provides additional data protection. When problems occur in the main link, it can quickly switch to the backup link to ensure reliable data transmission. By establishing a time-slotted response mechanism and setting critical condition judgment values ​​for triggering the backup link retransmission mechanism, effective resource management and improved transmission efficiency are achieved, which not only improves the system response speed but also reduces unnecessary energy consumption. When a hash value conflict is detected, the arbitration mechanism is activated, the backup link data is preferentially used and the security event is marked. At the same time, the dynamic key sequence is updated to enhance the system's fault recovery capability and security. The last key perturbation factor of the current encryption operation is injected into the subsequent chaotic iteration process, and the used key storage area is physically erased, further strengthening the security of the encryption algorithm and preventing the risks brought by the leakage of historical keys.

[0077] The instantaneous frequency sequence of the respiratory waveform and the smoothed heart rate curve of the electrocardiogram are used as observation vectors. The observation data of multiple network nodes are fused based on the extended Kalman filter. The state vector is defined as the latent state of respiratory frequency and the latent state of heart rate. The state transfer matrix and control input matrix are constructed. The process noise covariance and observation noise covariance are set to complete the extended Kalman filter parameter initialization.

[0078] Further explanation: The respiratory rate latent state in the state transition matrix is ​​used to describe the dynamic evolution of respiratory rate, reflecting the physiological mechanism by which respiration is weakly regulated by heart rate. For example, the current respiratory rate = 95% of the respiratory rate in the previous period + 3% of the heart rate in the previous period. The heart rate latent state corresponds to the dynamic relationship of heart rate changes, reflecting the high self-stability of heart rate. For example, the current heart rate = 2% of the respiratory influence in the previous period + 98% of the heart rate in the previous period. The control input matrix is ​​used to map the impact of external intervention (such as bed adjustment) on the state, define the intervention intensity of mechanical bed adjustment on physiological signals, optimize ventilation efficiency by changing the thoracic cavity volume, and quantify the adjustment amplitude of the heart rate caused by the same control action (changes in the tilt angle cause venous return fluctuations, resulting in a slight increase in heart rate). For example, every 1° increase in the bed tilt angle increases the respiratory rate by 0.1 breaths / minute.

[0079] Taking the respiratory abnormality event marker as the trigger condition, when an abnormality is detected, the initialization of the extended Kalman filter parameter iteration is performed as follows:

[0080] Based on the state estimate and control command execution of the previous time slot, the prior state and prior covariance are calculated. Combined with the current observation, the posterior state and covariance are updated to obtain the latent state estimate of the respiratory / heart rate. The latent state estimate of the respiratory / heart rate output by the extended Kalman filter is input into the model predictive controller. Within the prediction time domain, the state space equation is constructed, and the discretization step size is aligned with the physiological signal sampling period. A cost function is defined with physiological information tracking accuracy and control action amplitude as optimization objectives. The respiratory rate safety range and heart rate safety range are used as hard constraints, and the optimal control command sequence is solved through quadratic programming. The output respiratory rate sequence is monitored in real time. When multiple consecutive predicted points fall below the respiratory rate safety range and the actual observed value deviates synchronously, it is determined to be a persistent respiratory abnormality. The bed tilt angle adjustment is dynamically calculated, and the bed tilt angle is increased by the stepper motor, while the maximum adjustment range is limited to avoid excessive intervention. The calculated bed tilt angle adjustment is injected as a feedforward into the extended Kalman filter control input matrix. The actual angle is fed back through the bed pressure sensor to generate the bed posture control command containing the target adjustment parameters and safety limits.

[0081] To further illustrate, in this embodiment, the respiratory rate safety limit is 20≤respiratory rate≤60 times / minute, and the heart rate safety limit is 100≤heart rate≤160 times / minute.

[0082] It should be noted that by using the instantaneous frequency sequence of the respiratory waveform and the smoothed ECG heart rate curve as observation vectors and performing data fusion based on the extended Kalman filter, the latent state of the respiratory rate and heart rate can be estimated more accurately, thereby improving the accuracy of monitoring the infant's physiological state. By using the extended Kalman filter combined with the model predictive controller method, when abnormal breathing or heart rate is detected, it can respond quickly and calculate the optimal control instructions, which helps to take timely measures to prevent the occurrence of potentially dangerous situations. By quantifying the impact of mechanical adjustment of the bed (such as changes in the tilt angle) on the respiratory rate and heart rate, and dynamically adjusting the bed posture according to real-time monitoring results, it can effectively improve ventilation efficiency, stabilize the heart rate, and avoid the risks of excessive intervention.

[0083] The tilt angle target value, maximum allowable adjustment rate and anti-overshoot safety boundary in the bed posture control command are extracted, and the respiratory rate deviation, heart rate variation coefficient and control residual are input into the fuzzy logic controller. Based on the respiratory abnormality level classification rules, the real-time respiratory rate is matched with the respiratory rate safety limit. When the respiratory rate is detected to be mildly abnormal, the wireless network drives the sound and light composite stimulation module, encodes the sound and light soothing command into a PWM dimming signal and a Class D audio amplifier driving signal, and sends it to the LED array and speaker through the metamaterial antenna link, starting the progressive enhancement strategy, gradually increasing the light intensity, and expanding the white noise frequency band at the same time; when the respiratory rate is moderately abnormal, the wireless network sends vibration parameters and emergency braking conditions to the motor drive sequence. The link sends a control frame containing the start and stop vibration cycle and working duty cycle, and emergency braking conditions (such as the bed pressure sensor detects a reaction force of >10N), drives the eccentric wheel mechanism to generate mechanical vibration, and synchronously enables the current loop Feedback mechanism: When the motor current fluctuation exceeds the rated value, it automatically switches to slow start mode; when an apnea limit event is identified, the robot arm motion trajectory is decomposed into a Cartesian space coordinate sequence, and the joint angle instructions are redundantly transmitted through the wireless network dual link. The safe operating space is verified by the laser rangefinder before execution. When an obstacle is detected, it automatically switches to impedance control mode, and the contact force feedback is monitored in real time during the tapping action. During the execution of all control instructions, the APD photodetector power supply is turned off (disabling the laser drive current) and the metamaterial antenna receiving channel is cut off through a hardware interrupt signal, maintaining only a minimized communication link for execution status feedback; after the posture adjustment is completed or the abnormality is resolved, the quantum cascade laser array is restarted (such as a 500ms preheating delay), and full-scale data acquisition is reactivated after the respiratory signal-to-noise ratio is restored. The Kalman filter based on the wireless network multi-node compensates for the state estimation loss during the interruption and reconstructs the physiological signals during the interruption.

[0084] It should be noted that the fuzzy logic controller is used to process complex input parameters (such as respiratory rate deviation, heart rate variation coefficient and control residual) to achieve dynamic adjustment based on real-time data, thereby improving the effectiveness and pertinence of intervention measures; wireless networking technology is used to achieve rapid transmission of instructions, and dual-link redundant transmission of key instructions (such as the motion trajectory of the robotic arm) is used to ensure the reliable execution of commands in emergency situations. At the same time, unnecessary functions are turned off during the execution of control instructions to reduce potential interference sources; multiple operating modes are designed (such as slow start mode, impedance control mode, etc.), and can be automatically switched according to actual conditions, which not only protects the equipment but also avoids discomfort or harm to the baby.

[0085] In summary, the present invention realizes hardware-level synchronous acquisition of light intensity fluctuation signals and electrocardiogram signals through multi-band collaborative networking and dynamic time slot allocation, and combines the dual-channel noise capture and differential common mode suppression of metamaterial antennas to control signal synchronization error; the adaptive threshold algorithm based on wavelet decomposition and statistical noise feature joint analysis can dynamically distinguish motion artifacts from respiratory characteristic signals, thereby improving the suppression rate of respiratory waveform baseline drift, while retaining the energy integrity of the QRS wave group and avoiding the effective signal loss caused by the fixed threshold algorithm; through chaotic key generation and Markov decision model, dynamic mapping of abnormal events and transmission priority is realized. In the abnormal respiratory state, orthogonal frequency division multiplexing strategy is adopted to compress data transmission delay, and phase rotation encryption is used to enhance anti-interception capability to ensure end-to-end security of key physiological data.

[0086] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A smart baby crib monitoring system based on wireless transmission, characterized by: include, The network acquisition module dynamically allocates node communication time slots through wireless networking, adjusts the laser array tilt angle and activates the dual-band metamaterial antenna to synchronously collect light intensity fluctuation signals and electrocardiogram signals to generate a mixed signal stream; The signal optimization module uses adaptive filtering and noise joint analysis to extract physiological features from mixed signals, and generates clean respiratory and ECG waveforms through wavelet decomposition and dynamic threshold suppression; Abnormal decision module, based on zero crossing point detection and R wave energy verification to identify respiratory and heart rate abnormalities, combined with chaotic key generation Markov decision model to determine the transmission priority; the identification of respiratory and heart rate abnormalities The following steps are included: The respiratory cycle is detected based on the zero-crossing point, and a respiratory abnormality event is triggered when the instantaneous frequency in multiple consecutive sliding windows exceeds the safe range; The R-wave position is identified through first-order difference and energy verification, and a dynamic heart rate curve is constructed. When multiple consecutive data points of the smoothed heart rate exceed the safety limit, an abnormal heart rate event is triggered; The transmission control module triggers the multi-band coordinated transmission strategy according to priority, implements differentiated encryption and dual-link protocol encapsulation, and establishes a time-slotted response and dynamic retransmission mechanism; the triggering of the multi-band coordinated transmission strategy includes the following steps: The microsecond time component is used to generate the Lorenz chaotic initial key, and the dynamic key stream is generated through the fourth-order Runge-Kutta method iteration. Combined with the Markov decision model, the abnormal state is mapped into a four-dimensional state vector, and the multi-band coordinated transmission strategy is triggered according to the transmission priority weight ratio; The transmission strategy refers to setting the high priority threshold as , the medium priority threshold is , when the breathing weight is greater than or equal to When the high frequency band orthogonal frequency division multiplexing is activated, when the ECG weight is When the link quality is within the range of , the direct sequence spread spectrum of the relay node is enabled; otherwise, the narrowband sleep mode is used to transmit environmental data, and the modulation order and spreading code rules are dynamically adjusted based on the link quality; The data fusion module uses the extended Kalman filter to fuse multi-node observation data and generates bed posture adjustment instructions through the model predictive controller; The intervention execution module, based on the fuzzy logic controller, triggers the acoustic-optical-mechanical composite intervention in a hierarchical manner to execute the bed posture adjustment instructions.

2. The wireless transmission-based intelligent crib monitoring system according to claim 1, characterized in that: The generating of the mixed signal stream comprises the following steps: Dynamically allocate node communication time slots through wireless networking, adjust the tilt angle of the quantum cascade laser array, synchronously activate the dual-band metamaterial antenna, use the laser pulse reflection signal to generate a light intensity sequence, and simultaneously collect electrocardiogram signals and environmental noise; Common-mode interference is eliminated through differential amplification, and transmission delay is compensated by sliding window to generate a mixed signal stream.

3. The wireless transmission-based intelligent crib monitoring system according to claim 1, characterized in that: The physiological feature extraction of the mixed signal The following steps are included: Inputting the mixed signal stream into a low-noise amplifier to eliminate signal reflection interference and generate an amplified analog signal stream; The amplified analog signal stream is inputted into the respiratory signal channel processing branch and the electrocardiogram signal channel processing branch in parallel; Filtering the respiratory signal of the respiratory signal channel processing branch to generate a stable respiratory simulation signal; Perform anti-interference processing on the ECG signal of the ECG signal channel processing branch and output an ECG analog signal with normalized amplitude.

4. The wireless transmission-based intelligent crib monitoring system according to claim 1, wherein: The wavelet decomposition and dynamic threshold suppression refer to using Daubechies wavelet to decompose the respiratory waveform and the ECG waveform, jointly analyzing the noise characteristics based on the energy distribution of the high-frequency detail layer, dynamically suppressing the interference components through an adaptive threshold algorithm, and reconstructing a clean respiratory waveform that retains physiological characteristics and an ECG waveform with a complete QRS complex.

5. The wireless transmission-based intelligent crib monitoring system according to claim 1, characterized in that: The implementation of differentiated encryption and dual-link protocol encapsulation, and the establishment of a time-slotted response and dynamic retransmission mechanism include the following steps: Bind data slices with chaotic keys at the symbol level, bit level, and phase level to perform differentiated encryption operations; Encrypted data is encapsulated through a dual-link protocol, a cyclic redundancy check code is added to the main link, control instructions are transmitted on the backup link, and dynamic retransmission is triggered based on a slotted response mechanism.

6. The wireless transmission-based intelligent crib monitoring system according to claim 1, characterized in that: Generating bed posture adjustment instructions refers to using an extended Kalman filter to fuse multi-node observation data, constructing a latent state transfer model of respiratory frequency and heart rate, generating bed tilt angle adjustment instructions through a model predictive controller, and injecting the adjustment amount into the Kalman filter to update the state estimation.

7. The wireless transmission-based intelligent crib monitoring system according to claim 1, characterized in that: The execution of the bed posture adjustment instruction refers to triggering sound and light soothing, mechanical vibration and robotic arm patting actions based on the abnormal breathing level classification, shutting down the laser and antenna receiving channels during the execution of the bed posture adjustment instruction, restarting data collection after the posture adjustment is completed and compensating for the state estimation loss during the interruption.

Images