Intelligent multi-mode insomnia intervention system based on sleep monitoring

Abstract

The invention discloses an intelligent multi-mode insomnia intervention system based on sleep monitoring, belongs to the technical field of sleep, and solves the problems that insomnia and reasons thereof cannot be diagnosed in a non-contact manner without professional equipment at present, and targeted intervention is difficult to realize by a household sleep aid instrument. Comprising a main controller, and the main controller is connected with a perturbation sensing head array, a background field pickup antenna array, a distributed noise phase-locked amplification array and a data output end of an optical phase-locked thermodynamic imaging ring through high-speed data interfaces. According to the invention, human body multi-dimensional physical field information is sensed synchronously in a non-contact manner in a household natural sleep environment, physiological and psychological root causes of insomnia are diagnosed dynamically by using a causal entropy network model, and a physical field targeting intervention strategy with high specificity is generated and executed accordingly. Therefore, a complete closed loop from accurate detection and root cause analysis to personalized treatment is realized under the condition that hospital examination is not needed.

Description

Technical Field

[0001] This invention relates to the field of sleep technology, and in particular to an intelligent multimodal insomnia intervention system based on sleep monitoring. Background Technology

[0002] Insomnia is a common sleep disorder, mainly characterized by difficulty falling asleep, light sleep, frequent awakenings, difficulty falling back asleep after waking up early, or daytime fatigue and poor concentration. Long-term insomnia can affect physical and mental health and quality of life. Insomnia testing requires a combination of subjective feelings and objective assessments. Professional doctors typically conduct detailed consultations to understand sleep habits, symptom duration, and triggers. They may also use sleep monitoring equipment to record data such as EEG, ECG, and respiration during sleep to accurately determine the type and severity of insomnia. Some routine testing can be done by keeping a sleep diary, including sleep onset time, number of awakenings, and pre-sleep activities. Improving sleep requires a multi-pronged approach. A regular sleep schedule is fundamental; consistent bedtime and wake-up time stabilize the biological clock. Avoiding caffeine and strong tea before bed, reducing electronic device use, and creating a quiet, dark, and temperature-appropriate sleep environment are also crucial. In addition, moderate exercise, relaxation techniques such as deep breathing and meditation, or relieving stress through gentle music and foot baths can all help improve sleep. If insomnia persists, professional medical help should be sought promptly to avoid self-medication.

[0003] Current insomnia diagnosis often requires detailed testing in hospitals followed by a thorough assessment by experienced doctors to determine the possible cause of the insomnia. In such cases, medication and other sleep aids are used to guide the patient into hypnosis, gradually improving the insomnia. However, for most patients, frequent visits to hospitals for sleep aids are not readily accepted. Self-medication at home may even lead to incorrect use and health problems. Currently, it is impossible to diagnose insomnia without the aid of specialized hospital equipment, including non-contact methods to determine its presence and cause. Therefore, even if insomnia patients have appropriate sleep aids at home, they still cannot receive targeted sleep support.

[0004] Therefore, an intelligent multimodal insomnia intervention system based on sleep monitoring is proposed to solve or alleviate the above problems. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an intelligent multimodal insomnia intervention system based on sleep monitoring.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An intelligent multimodal insomnia intervention system based on sleep monitoring includes a main controller. The main controller is connected to the data output terminals of a perturbation sensor head array, a background field pickup antenna array, a distributed noise phase-locked amplification array, and an optical phase-locked thermodynamic imaging ring via high-speed data interfaces. The main controller is also connected to the control terminals of a metasurface wavefront sensor and a phase conjugate plate via a configuration bus. Furthermore, the main controller is connected to a nonlinear decoupling engine via a high-speed expansion interface and a closed-loop topological insulator intervention module via a control signal interface. The main controller coordinates the synchronous acquisition of signals by each functional module, processes and fuses the original signals into a multi-dimensional feature vector, constructs and dynamically updates the causal entropy network model based on this, analyzes the topology and dynamic characteristics of the network to diagnose the immediate dominant subtype and root cause of insomnia, generates targeted intervention strategies based on the diagnostic results and converts them into control commands, re-acquires data after intervention to evaluate the effect and adaptively optimizes the model parameters. The perturbation sensor array collects extremely low-frequency physiological phonon signals naturally emitted by the human body during sleep and converts them into electrical signals characterizing cellular metabolic oscillations. The background field pickup antenna array collects chaotic background electromagnetic field signals modulated by human bioelectromagnetic activity to reflect the activity synchronicity and chaotic characteristics of the cerebral cortex. The metasurface wavefront sensor and phase conjugate plate collect wavefront phase distortion information of environmental radio frequency signals caused by human breathing and micro-movement, and can adjust its unit state under the control of the main controller to participate in sensing. The distributed noise lock-in amplification array collects ultra-weak magnetic field signals generated by neural currents and outputs quantum noise data reflecting neural cluster activity through multi-channel correlation measurement. The nonlinear decoupling engine receives the feature data stream after front-end processing and calculates the causal and information transmission relationship between feature pairs at high speed through its dedicated hardware circuit to assist the main controller in building a causal entropy network. The optical phase-locked thermodynamic imaging ring acquires infrared thermal radiation photons from specific brain regions of the human body and obtains thermodynamic entropy change information characterizing the metabolic efficiency of local tissues through phase-locked photon counting and statistics. The closed-loop topological insulator intervention module receives and executes specific control commands issued by the main controller to drive the topological insulator thin film to generate non-contact physical field intervention with specific waveforms, intensities and frequencies to act on the human body for targeted regulation.

[0007] Preferably, the perturbation sensing head array includes multiple heterojunction sensing units, a first precision voltage reference source, a first digital potentiometer, a first buffer amplifier, a first transimpedance amplifier, a first analog multiplexer, a first programmable filter, and a first analog-to-digital converter. The positive current terminal of each heterojunction sensing unit is connected to the inverting input of the first transimpedance amplifier through a first current-limiting resistor. The negative current terminals of each heterojunction sensing unit are connected to a first negative bias reference ground. The positive voltage output pin of the first precision voltage reference source is connected to the power input pin of the first digital potentiometer. The sliding output pin of the first digital potentiometer is connected to the non-inverting input of the first buffer amplifier. The output pin of the first buffer amplifier provides a negative bias voltage to the first negative bias reference ground network. The serial interface pin of the first digital potentiometer is connected to the main control... The first serial peripheral interface on the processing system side of the main controller has a first feedback resistor and a first feedback capacitor connected in parallel between the inverting input terminal and the output terminal of the first transimpedance amplifier. Its output pin is connected to the common input pin of the first analog multiplexer. The multiple channel selection output pins of the first analog multiplexer are connected in parallel and then connected to the signal input pin of the first programmable filter. Its channel address selection pin is connected to the general-purpose input / output pin on the processing system side of the main controller. The signal output pin of the first programmable filter is connected to the analog input positive pin of the first analog-to-digital converter. The serial clock pin, serial data input pin, and serial data output pin of the first analog-to-digital converter are respectively connected to the corresponding pins of the second serial peripheral interface on the programmable logic side of the main controller. Its conversion start control pin is driven by the general-purpose input / output pin of the main controller.

[0008] Preferably, the background field pickup antenna array includes multiple ultra-wideband chip antennas, a high-frequency circuit board, multiple first low-noise amplifiers, multiple first mixers, a first frequency synthesizer, multiple first variable gain amplifiers, a first ultra-high-speed analog-to-digital converter, and a first field-programmable gate array. The signal output pin of each ultra-wideband chip antenna is soldered to the feed point of the high-frequency circuit board and connected to the RF input pin of the corresponding first low-noise amplifier via a microstrip transmission line. The RF output pin of each first low-noise amplifier is connected to the RF input pin of the corresponding first mixer via a microstrip transmission line. The local oscillator input pin of the first mixer is connected in parallel to the RF output pin of the first frequency synthesizer via a power distribution network. The clock and data pins of the serial configuration interface of the first frequency synthesizer are connected to the second serial peripheral interface on the processing system side of the main controller. The intermediate frequency output pin of each first mixer is connected to... The differential input positive and negative pins of the first variable gain amplifier are connected via coaxial cables or differential traces. The differential output positive and negative pins of each of the first variable gain amplifiers are connected via parallel differential pairs to the analog input differential channel positive and negative pins of the first ultra-high-speed analog-to-digital converter. All conversion channels of the first ultra-high-speed analog-to-digital converter share an external reference clock input pin to receive the sampling clock signal. Its multiple pairs of transmit differential data positive and negative pins are connected via differential lines on the circuit board to the corresponding multiple pairs of gigabit transceivers on the first field-programmable gate array to receive differential data positive and negative pins. The synchronization control pins and power management pins of the first ultra-high-speed analog-to-digital converter are connected to the general-purpose input and output pins of the first field-programmable gate array. Its digital output data stream is preprocessed by the first field-programmable gate array and then transmitted to the main controller via a high-speed expansion interface.

[0009] Preferably, the metasurface wavefront sensing and phase conjugate plate includes multiple liquid crystal metasurface units, a first high-voltage driving chip, a first integrated transceiver, an audio digital signal processing chip, a first high-voltage digital output driver, and at least one reference antenna. The upper electrode pin of each liquid crystal metasurface unit is connected to a corresponding current output pin of the first high-voltage driving chip, and the lower electrode pin of each liquid crystal metasurface unit is connected to the system analog ground. The serial clock pin and serial data pin of the first high-voltage driving chip are connected to the first internal integrated circuit interface on the processing system side of the main controller, and its multiple high-voltage output pins are respectively connected to the corresponding liquid crystal metasurface units. The reference antenna is connected to the first receive channel input pin of the first integrated transceiver via a coaxial cable. The local oscillator and transmit sections of the first integrated transceiver are disabled. The data input port of the first integrated transceiver is connected to the audio digital signal processing chip. Multiple general-purpose input / output extension pins of the audio digital signal processing chip are connected to the parallel digital input pins of the first high-voltage digital output driver. The first high-voltage digital output driver receives parallel digital control words from the audio digital signal processing chip and connects its corresponding multiple high-level digital output pins to the upper electrode pins of the corresponding liquid crystal metasurface unit via second current-limiting resistors.

[0010] Preferably, the distributed noise lock-in amplifier array includes multiple Josephson junction sensing units, a cryostat, multiple first high-speed current sources, multiple first instrumentation amplifiers, a first synchronous sampling analog-to-digital converter array, and a second field-programmable gate array. The two electrodes of each Josephson junction sensing unit are connected via superconducting wires inside the cryostat to the current output pin of the corresponding first high-speed current source and the non-inverting input of the first instrumentation amplifier through a first cryogenic coaxial connector. The shielding layer of each first cryogenic coaxial connector is connected to an independent low-noise reference ground. The digital control interface of each first high-speed current source is connected to the third serial peripheral interface on the processing system side of the main controller. The output pin of an instrumentation amplifier is connected to the analog input positive pin of the analog-to-digital converter in the first synchronous sampling analog-to-digital converter array. The reference ground pin of the first instrumentation amplifier is connected to the same low-noise analog ground plane. The synchronous sampling control pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are interconnected and receive the sampling clock signal together. The serial data output pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are connected in parallel to the serial peripheral interface master input slave output pin of the second field-programmable gate array. The independent chip select pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are controlled by different general-purpose input / output pins of the second field-programmable gate array.

[0011] Preferably, the nonlinear decoupling engine includes a dedicated causal computing integrated circuit and a high-speed data buffer memory. The multi-channel high-speed parallel data input bus of the dedicated causal computing integrated circuit is connected to the general-purpose input / output pin group of the second field-programmable gate array in the distributed noise lock-in amplifier array via a parallel bus on the circuit board. The dedicated causal computing integrated circuit is connected to the fourth serial peripheral interface on the processing system side of the main controller via a configuration bus. The calculation result output of the dedicated causal computing integrated circuit is connected to the corresponding high-speed peripheral component interconnect interface pin on the programmable logic side of the main controller via multiple pairs of transmit differential pins and receive differential pins of its high-speed peripheral component interconnect data link. The data pins, address pins, and control pins of the high-speed data buffer memory are respectively connected to the corresponding external memory interface pins of the dedicated causal computing integrated circuit.

[0012] Preferably, the optical phase-locked thermodynamic imaging ring includes a single-photon avalanche diode array, a matching readout chip, a direct digital frequency synthesizer, and an electro-optic phase modulator array. The single-photon avalanche diode array is integrated from multiple single-photon avalanche diode pixels in a two-dimensional matrix. The serial peripheral interface clock pin, serial data input pin, and serial data output pin of the matching readout chip are connected to the corresponding pins of the third serial peripheral interface on the programmable logic side of the first field-programmable gate array in the background field pickup antenna array. The multiple digital pulse output pins of the single-photon avalanche diode array are respectively connected to the multiple dedicated general-purpose input / output pins of the first field-programmable gate array in the background field pickup antenna array. The serial configuration interface of the direct digital frequency synthesizer is connected to the first internal integrated circuit interface on the processing system side of the main controller. The positive and negative terminals of its differential analog output pins are connected to the driving electrodes of the electro-optic phase modulator array. The electro-optic phase modulator array is placed at the front end of the optical receiving path of the single-photon avalanche diode array.

[0013] Preferably, the closed-loop topological insulator intervention module includes a topological insulator thin film, a first power electrode, a second power electrode, an isolated gate driver, a full-bridge power amplifier circuit, a current monitoring amplifier, and a sampling resistor. The first edge of the topological insulator thin film is in ohmic contact with the first power electrode, and its second edge is in ohmic contact with the second power electrode. The first and second power electrodes are respectively welded to the first and second power output nodes of the full-bridge power amplifier circuit. The full-bridge power amplifier circuit is composed of four gallium nitride power transistors in a full-bridge topology, wherein the connection point between the first high-side transistor and the first low-side transistor forms the first power output node. The connection point between the high-side transistor and the low-side transistor forms the second power output node. The gate drive pins of the four transistors in the full-bridge power amplifier circuit are respectively connected to the four independent output pins of the isolated gate driver. The input power supply and logic input pins of the isolated gate driver are connected to the general-purpose input / output pin group of the programmable logic side of the main controller. The sampling resistor is connected in series between the first power output node and the first power electrode, and the voltage difference across its two ends is connected to the non-inverting input and inverting input of the current monitoring amplifier. The voltage output pin of the current monitoring amplifier is connected to the second analog input channel pin of the first analog-to-digital converter in the perturbation sensor array.

[0014] Preferably, it also includes a system-level clock circuit, the output of which is connected to the clock input of the main controller and each module to provide a unified synchronous clock reference.

[0015] Preferably, the system-level clock circuit includes a system master clock generator and a clock buffer. The first clock output of the system master clock generator is connected to the reference clock input pin of the processing system side of the main controller. The second clock output of the system master clock generator is connected to the clock input of the clock buffer. The multiple clock outputs of the clock buffer are respectively connected to the reference clock input pins of the first ultra-high-speed analog-to-digital converter in the background field pickup antenna array, the first synchronous sampling analog-to-digital converter array in the distributed noise phase-locked amplification array, and the direct digital frequency synthesizer in the optical phase-locked thermodynamic imaging loop. The third clock output of the system master clock generator is connected to the reference clock input pin of the first integrated transceiver in the metasurface wavefront sensing and phase conjugate plate.

[0016] The present invention has the following beneficial effects: This invention achieves a complete closed loop from accurate detection and root cause analysis to personalized treatment without the need for hospital examinations by non-contact synchronous sensing of multidimensional physical field information of the human body in a natural home sleep environment, using a causal entropy network model to dynamically diagnose the physiological and psychological root causes of insomnia, and generating and executing highly specific physical field-targeted intervention strategies accordingly. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a structural block diagram of the present invention.

[0019] In the figure: 1. Main controller; 2. Perturbation sensor head array; 3. Background field pickup antenna array; 4. Metasurface wavefront sensing and phase conjugate plate; 5. Distributed noise phase-locked amplifier array; 6. Nonlinear decoupling engine; 7. Optical phase-locked thermodynamic imaging loop; 8. Closed-loop topology insulator intervention module; 9. System-level clock circuit. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0021] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] In the description of this invention, it should be understood that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used to facilitate the description of this invention and to simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0024] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0025] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0026] An intelligent multimodal insomnia intervention system based on sleep monitoring, such as Figure 1 As shown, the main controller 1, model ZU19EG-2FFVC1760I, is connected to the data output terminals of the perturbation sensor array 2, background field pickup antenna array 3, distributed noise phase-locked amplifier array 5, and optical phase-locked thermodynamic imaging ring 7 via high-speed data interfaces. The main controller 1 is connected to the control terminal of the metasurface wavefront sensor and phase conjugate plate 4 via a configuration bus. The main controller 1 is connected to the nonlinear decoupling engine 6 via a high-speed expansion interface. The main controller 1 is connected to the closed-loop topology insulator intervention module 8 via a control signal interface. The main controller 1 coordinates the synchronous acquisition of signals by various functional modules, processes and fuses the raw signals into multidimensional feature vectors, constructs and dynamically updates the causal entropy network model based on this, analyzes the topology and dynamic characteristics of the network to diagnose the immediate dominant subtype and root cause of insomnia, generates targeted intervention strategies based on the diagnostic results and converts them into control commands, re-acquires data after intervention to evaluate the effect and adaptively optimizes the model parameters. The perturbation sensor array 2 collects extremely low-frequency physiological phonon signals naturally emitted by the human body during sleep and converts them into electrical signals characterizing cellular metabolic oscillations. The perturbation sensor array 2 includes multiple heterojunction sensing units, a first precision voltage reference source LT6657BHMS8-2.5, a first digital potentiometer AD5122BRUZ10, a first buffer amplifier OPA2188AIDR, a first transimpedance amplifier LTC6268IS8-10, a first analog multiplexer ADG704BRMZ, and a first programmable filter LT. The C1560-1CS8 and the first analog-to-digital converter AD4003BCPZ-RL7 each have a heterojunction sensing unit. Each heterojunction sensing unit is composed of a superlattice structure formed by alternating growth of gallium arsenide and aluminum gallium arsenide materials through molecular beam epitaxy and is encapsulated in an electromagnetic shielding shell. The positive current terminal of the heterojunction sensing unit is connected to the inverting input terminal of the first transimpedance amplifier through a first current-limiting resistor. The negative current terminals of the heterojunction sensing units are connected to the first negative bias reference ground. The positive voltage output pin of the first precision voltage reference source is connected to the power supply of the first digital potentiometer. The input pin of the first digital potentiometer is connected to the non-inverting input of the first buffer amplifier. The output pin of the first buffer amplifier provides a negative bias voltage to the first negative bias reference ground network. The serial interface pin of the first digital potentiometer is connected to the first serial peripheral interface on the processing system side of the main controller 1. A first feedback resistor and a first feedback capacitor are connected in parallel between the inverting input and output of the first transimpedance amplifier. Its output pin is connected to the common input pin of the first analog multiplexer. The multiple channel selection output pins of the first analog multiplexer are connected in parallel and then connected to the signal input pin of the first programmable filter. Its channel address selection pin is connected to the general-purpose input / output pin on the processing system side of the main controller 1. The signal output pin of the first programmable filter is connected to the analog input positive pin of the first analog-to-digital converter. The serial clock pin, serial data input pin, and serial data output pin of the first analog-to-digital converter are respectively connected to the corresponding pins of the second serial peripheral interface on the programmable logic side of the main controller 1. Its conversion start control pin is driven by the general-purpose input / output pin of the main controller 1. Background field pickup antenna array 3 collects chaotic background electromagnetic field signals modulated by human bioelectromagnetic activity to reflect the synchronicity and chaotic characteristics of cerebral cortex activity. Background field pickup antenna array 3 includes multiple ultra-wideband chip antennas ANAU-0808-A2-T, a high-frequency circuit board, multiple first low-noise amplifiers SKY65404-31, multiple first mixers ADL5801ACPZ-R7, a first frequency synthesizer ADF4351BCPZ, multiple first variable gain amplifiers AD8370ACPZ-R7, a first ultra-high-speed analog-to-digital converter ADC12DJ3200RFMR, and a first field-programmable gate array XCKU060-2FFVA1156I. The signal output pin of each ultra-wideband chip antenna is soldered to the feed point of the high-frequency circuit board and connected to the RF input pin of the corresponding first low-noise amplifier via a microstrip transmission line. The RF output pin of each first low-noise amplifier is connected to the RF input pin of the corresponding first mixer via a microstrip transmission line. The local oscillator input pin of the first mixer is connected in parallel to the first frequency synthesizer via a power distribution network. The RF output pins of the first frequency synthesizer, the clock pins and data pins of the serial configuration interface of the first frequency synthesizer are connected to the second serial peripheral interface on the processing system side of the main controller 1. The intermediate frequency output pins of each first mixer are connected to the differential input positive pins and differential input negative pins of the corresponding first variable gain amplifier through coaxial cables or differential traces. The differential output positive pins and differential output negative pins of each first variable gain amplifier are connected to the analog input differential channel positive pins and negative pins of the corresponding first ultra-high speed analog-to-digital converter through parallel differential pairs. All conversion channels of the first ultra-high speed analog-to-digital converter share an external reference clock input pin to receive the sampling clock signal. Its multiple pairs of transmit differential data positive pins and negative pins are connected to the corresponding multiple pairs of gigabit transceivers on the first field-programmable gate array through differential line pairs on the circuit board to receive differential data positive pins and negative pins. The synchronization control pins and power management pins of the first ultra-high speed analog-to-digital converter are connected to the general-purpose input and output pins of the first field-programmable gate array. Its digital output data stream is preprocessed by the first field-programmable gate array and then transmitted to the main controller 1 through the high-speed expansion interface. The metasurface wavefront sensing and phase conjugate plate 4 collects wavefront phase distortion information of environmental radio frequency signals caused by human respiration and micro-movements, and can adjust its unit state under the control of the main controller 1 to participate in sensing. The metasurface wavefront sensing and phase conjugate plate 4 includes multiple liquid crystal metasurface units, a first high-voltage driver chip MAX25210ATE / V+, a first integrated transceiver AD9371BBCZ, an audio digital signal processing chip ADAU1452WBCPZ, a first high-voltage digital output driver MAX14900AEWC+, and at least one reference antenna. Each liquid crystal metasurface unit is composed of two flexible circuit boards sandwiching a liquid crystal material layer. The upper flexible circuit board is etched with a metal resonant structure pattern to form an electromagnetic metasurface. The upper electrode pin of each liquid crystal metasurface unit is connected to a corresponding current output pin of the first high-voltage driver chip. The lower electrode pins are connected to the system analog ground. The serial clock pin and serial data pin of the first high voltage driver chip are connected to the first internal integrated circuit interface of the processing system side of the main controller 1. Its multiple high voltage output pins are respectively connected to the upper electrode pins of the corresponding liquid crystal metasurface unit. The reference antenna is connected to the first receiving channel input pin of the first integrated transceiver through a coaxial cable. The local oscillator and transmitting sections of the first integrated transceiver are disabled. The data input port of the first integrated transceiver is connected to the audio digital signal processing chip. Multiple general-purpose input / output extension pins of the audio digital signal processing chip are connected to the parallel digital input pins of the first high voltage digital output driver. The first high voltage digital output driver receives the parallel digital control word from the audio digital signal processing chip and connects its corresponding multiple high-level digital output pins to the upper electrode pins of the corresponding liquid crystal metasurface unit through the second current-limiting resistor. The distributed noise lock-in amplifier array 5 acquires ultra-weak magnetic field signals generated by neural currents and outputs quantum noise data reflecting neural cluster activity through multi-channel correlation measurement. The distributed noise lock-in amplifier array 5 includes multiple Josephson junction sensing units, a cryostat, multiple first high-speed current sources ADN2890ACPZ-R2, multiple first instrumentation amplifiers AD8429ARMZ-R7, a first synchronous sampling analog-to-digital converter array ADS9224RSAR, and a second field-programmable gate array 10AX066H3F34E2SG. Each Josephson junction sensing unit is made of niobium-alumina-niobium superconducting material and encapsulated within the vacuum cavity of the cryostat, operating in the liquid nitrogen temperature range. The two electrodes of each Josephson junction sensing unit are connected via superconducting wires inside the cryostat to the current output pin of the corresponding first high-speed current source and the non-inverting input terminal of the first instrumentation amplifier through a first cryogenic coaxial connector. The shielding layer of each first cryogenic coaxial connector is connected to an independent low-noise reference ground. The digital control interface of each first high-speed current source... The interface is connected to the third serial peripheral interface on the processing system side of the main controller 1. The output pin of each first instrumentation amplifier is connected to the analog input positive pin of the analog-to-digital converter in the first synchronous sampling analog-to-digital converter array. The reference ground pin of the first instrumentation amplifier is connected to the same low-noise analog ground plane. The synchronous sampling control pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are interconnected and receive the sampling clock signal together. The serial data output pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are connected in parallel to the master input and slave output pins of the serial peripheral interface of the second field-programmable gate array. The independent chip select pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are controlled by different general-purpose input and output pins of the second field-programmable gate array. The second field-programmable gate array is configured with parallel cross-correlation operation logic to perform real-time correlation matrix calculation on all channel digital noise data received from the first synchronous sampling analog-to-digital converter array. The calculated correlation data is transmitted to the main controller 1 through the high-speed expansion interface for subsequent causal entropy network model construction. The nonlinear decoupling engine 6 receives the feature data stream from the front-end processing and uses its dedicated hardware circuitry to calculate the causal and information transmission relationships between feature pairs at high speed to assist the main controller 1 in constructing a causal entropy network. The nonlinear decoupling engine 6 includes a dedicated causal calculation integrated circuit and a high-speed data buffer memory MT40A512M16LY-062EIT:F. The dedicated causal calculation integrated circuit is manufactured using a dedicated semiconductor process and internally contains a parallel computing unit array for calculating Granger causality strength and transmission entropy. The dedicated causal calculation integrated circuit's multi-channel high-speed parallel data input bus is connected to the general-purpose input / output pin group of the second field-programmable gate array in the distributed noise lock-in amplifier array 5 via a parallel bus on the circuit board. The dedicated causal calculation integrated circuit is connected to the main controller 1 via a configuration bus. The fourth serial peripheral interface on the processing system side connects the computation result output of the dedicated causal computing integrated circuit to the corresponding high-speed peripheral component interconnection interface pins on the programmable logic side of the main controller 1 through multiple pairs of transmit differential pins and receive differential pins of its high-speed peripheral component interconnection data link. The data pins, address pins and control pins of the high-speed data buffer memory are respectively connected to the corresponding external memory interface pins of the dedicated causal computing integrated circuit. The programmable logic resources in the second field-programmable gate array are configured as a dynamic Bayesian network update circuit. The real-time data input of this circuit receives the fused feature data stream output from the direct memory access controller on the programmable logic side of the main controller 1 through the advanced extensible interface stream protocol interface. The parameter update output of this circuit affects the subsequent diagnostic decision logic through the on-chip interconnection bus inside the main controller 1. The optical phase-locked thermodynamic imaging ring 7 acquires infrared thermal radiation photons from specific brain regions of the human body and obtains thermodynamic entropy change information characterizing the metabolic efficiency of local tissues through phase-locked photon counting and statistics. The optical phase-locked thermodynamic imaging ring 7 includes a single-photon avalanche diode array ArrayC-30035-4P, a matching readout chip, a direct digital frequency synthesizer AD9914BCPZ, and an electro-optic phase modulator array. The single-photon avalanche diode array is composed of multiple single-photon avalanche diode pixels integrated in a two-dimensional matrix. The serial peripheral interface clock pin, serial data input pin, and serial data output pin of the matching readout chip are connected to the corresponding pins of the third serial peripheral interface on the programmable logic side of the first field-programmable gate array in the background field pickup antenna array 3. The multiple digital pulse output pins of the single-photon avalanche diode array are respectively connected to the first field-programmable gate array in the background field pickup antenna array 3. The array has multiple dedicated general-purpose input / output pins. The serial configuration interface of the direct digital frequency synthesizer is connected to the first internal integrated circuit interface of the processing system side of the main controller 1. The positive and negative terminals of its differential analog output pins are connected to the driving electrodes of the electro-optic phase modulator array. The electro-optic phase modulator array is placed at the front end of the optical receiving path of the single-photon avalanche diode array. The multiple start signal channels of the time-to-digital converter logic implemented in the first field-programmable gate array are synchronized with the reference clock output by the direct digital frequency synthesizer. Its multiple stop signal channels receive pulse signals from each pixel of the single-photon avalanche diode array, thereby accurately measuring the flight time of each photon arrival event relative to the modulation period. The processed photon arrival time histogram data is uploaded to the main controller 1 through the high-speed gigabit transceiver of the first field-programmable gate array for calculating the thermodynamic entropy change of the local tissue. The closed-loop topological insulator intervention module 8 receives and executes specific control commands from the main controller 1 to drive the topological insulator thin film to generate a non-contact physical field intervention with specific waveforms, intensities, and frequencies to act on the human body for targeted regulation. The closed-loop topological insulator intervention module 8 includes a topological insulator thin film, a first power electrode, a second power electrode, an isolated gate driver ADUM4121ARIZ, a full-bridge power amplifier circuit, a current monitoring amplifier INA210AIDCKR, and a sampling resistor. The topological insulator thin film is made of bismuth selenide crystal material. The first edge of the topological insulator thin film is in ohmic contact with the first power electrode, and its second edge is in ohmic contact with the second power electrode. The first and second power electrodes are respectively welded to the first and second power output nodes of the full-bridge power amplifier circuit. The large circuit consists of four gallium nitride power transistors in a full-bridge topology. The connection point between the first high-side transistor and the first low-side transistor forms the first power output node, and the connection point between the second high-side transistor and the second low-side transistor forms the second power output node. The gate drive pins of the four transistors in the full-bridge power amplifier circuit are respectively connected to the four independent output pins of the isolated gate driver. The input power supply and logic input pins of the isolated gate driver are connected to the general-purpose input / output pin group of the programmable logic side of the main controller 1. The sampling resistor is connected in series between the first power output node and the first power electrode, and the voltage difference across its two ends is connected to the non-inverting input and the inverting input of the current monitoring amplifier. The voltage output pin of the current monitoring amplifier is connected to the second analog input channel pin of the first analog-to-digital converter in the perturbation sensor array 2.

[0027] It also includes a system-level clock circuit 9. The output of the system-level clock circuit 9 is connected to the clock input of the main controller 1 and each module to provide a unified synchronous clock reference. The system-level clock circuit 9 includes a system master clock generator Si5345B-D-GM and a clock buffer ADCLK946BCPZ. The first clock output of the system master clock generator is connected to the reference clock input pin of the processing system side of the main controller 1. The second clock output of the system master clock generator is connected to the clock input of the clock buffer. The multiple clock outputs of the clock buffer are respectively connected to the reference clock input pins of the first ultra-high-speed analog-to-digital converter in the background field pickup antenna array 3, the first synchronous sampling analog-to-digital converter array in the distributed noise phase-locked amplifier array 5, and the direct digital frequency synthesizer in the optical phase-locked thermodynamic imaging ring 7. The third clock output of the system master clock generator is connected to the reference clock input pin of the first integrated transceiver in the metasurface wavefront sensing and phase conjugate plate 4.

[0028] When the aforementioned intelligent multimodal insomnia intervention system based on sleep monitoring is in operation, an intelligent multimodal insomnia intervention method based on sleep monitoring is provided to cooperate with the intelligent multimodal insomnia intervention system based on sleep monitoring, including the following steps: Step S1: The main controller 1 controls and synchronizes the perturbation sensor head array 2, the background field pickup antenna array 3, the metasurface wavefront sensing and phase conjugate plate 4, the distributed noise lock-in amplifier array 5, and the optical lock-in thermodynamic imaging ring 7 to acquire non-contact raw physical field signals, and processes these signals to convert them into standardized digital features that characterize specific physiological and psychological functions in order to form a multidimensional feature vector. The main controller 1 receives the signal from the quantum perturbation sensor array 2 through its programmable logic side high-speed expansion interface, extracts the main peak frequency of its power spectrum in the extremely low frequency band as the main frequency feature of metabolic oscillation, and calculates the ratio of the main peak energy to the total energy in the extremely low frequency band as the spectral purity feature. The main controller 1 receives the multi-channel signal of the chaotic background field pickup antenna array 3 from the background field pickup antenna array 3 through its programmable logic side high-speed expansion interface, calculates the attenuation rate of the coherence coefficient of the signal originating from the forehead region and the pillow region in a specific frequency band, and uses it as a feature of the disintegration rate of cortical functional connectivity. The main controller 1 receives signals from the metasurface wavefront sensing plate 4 through its internal integrated circuit interface on the processing system side and the general input / output interface on the programmable logic side, and analyzes the breathing fundamental frequency and its harmonic components. The ratio of the total harmonic amplitude to the fundamental frequency amplitude is used as the breathing harmonic distortion characteristic. At the same time, it detects body movement micro-events and calculates the Shannon entropy of their time intervals as the micro-motion event entropy characteristic. The main controller 1 receives the multi-channel noise cross-correlation matrix data of the distributed quantum noise lock-in amplifier array 5 from the distributed noise lock-in amplifier array 5 through its programmable logic side high-speed expansion interface, calculates the correlation dimension of the eigenvalue distribution after eigenvalue decomposition of the matrix, and uses it as the neural magnetic fluctuation complexity feature. The main controller 1 receives signals from the optical phase-locked thermodynamic imaging ring 7 through its programmable logic side general input / output interface, calculates the thermodynamic entropy of the orbitofrontal cortex region and the primary motor cortex region based on the photon arrival time interval distribution, and calculates the difference between the two entropy values ​​as the characteristic of the brain region thermodynamic entropy difference. The main controller 1 combines the extracted metabolic oscillation main frequency, spectral purity, cortical functional connectivity disintegration rate, respiratory harmonic distortion, micromotion event entropy, neural magnetic fluctuation complexity, and brain region thermodynamic entropy difference features into a multidimensional feature vector. Step S2: Based on the time series data of the multidimensional feature vector obtained in step S1, the main controller 1 dynamically constructs and updates a causal entropy network model; wherein, the nodes in the causal entropy network model correspond to standardized digital features, and the weight of the directed edges between nodes is synthesized by the linear causal influence strength and the nonlinear information transmission strength between the corresponding feature pairs. For any two different time series features in the multidimensional feature vector, the main controller 1 uses a regularized vector autoregression model for analysis; by comparing the difference in the variance of the prediction error generated by including and not including the historical information of the second feature when predicting the future value of the first feature, the linear causal influence strength from the second feature to the first feature is calculated. For the same two feature time series, the main controller 1 quantifies the nonlinear information transmission strength from the second feature to the first feature by calculating the portion reduced by the historical information of the second feature in the uncertainty of the future value of the first feature; The main controller 1 normalizes the calculated linear causal influence intensity and nonlinear information transmission intensity respectively, and then combines them linearly through preset weighting coefficients. The combination result is used as the real-time weight of the directed edge from the second feature node to the first feature node in the causal entropy network model, thereby completing the dynamic construction and update of the network. Step S3: The main controller 1 analyzes the real-time topology and dynamic characteristics of the causal entropy network model. By calculating the causal dominance index of network nodes, detecting the abnormal feedback loop gain in the network, and comparing it with the dynamic threshold established based on historical data, the immediate dominant subtype and root cause of insomnia are diagnosed. For each node in the causal entropy network model, the main controller 1 calculates the difference between the sum of the weights of all outgoing edges and the sum of the weights of all incoming edges to obtain the causal dominance index of that node; nodes with a significantly positive causal dominance index are identified as the driving source of the network. The main controller 1 searches for strongly connected node loops in the causal entropy network model, and obtains the maximum magnitude of the loop by calculating the eigenvalue of the Jacobian matrix of the weight matrix corresponding to the loop. Loops with a gain greater than 1 are determined to be loops with non-decaying abnormal oscillations. The main controller 1 compares the causal dominance index of the cortical functional connection disintegration rate feature and the edge weight from the feature node to the respiratory harmonic distortion feature node with the first dynamic threshold and the second dynamic threshold, respectively. If both exceed their corresponding thresholds, the current insomnia subtype is determined to be the psychogenic hyperarousal dominant type. The main controller 1 compares the gain of the loop formed by the metabolic oscillation main frequency characteristic node and the respiratory harmonic distortion characteristic node, as well as the causal dominance index of the brain region thermodynamic entropy difference characteristic, with the third dynamic threshold. If the loop gain is greater than 1 and the causal dominance index of the entropy difference characteristic is lower than the third dynamic threshold, the current insomnia subtype is determined to be the physiological rhythm disorder dominant type. The main controller 1 compares the nonlinear information transmission strength from the neural magnetic fluctuation complexity feature node to the metabolic oscillation main frequency feature node, and the causal influence strength from the metabolic oscillation main frequency feature node to the neural magnetic fluctuation complexity feature node with the fourth dynamic threshold and the fifth dynamic threshold, respectively. If both are lower than their corresponding thresholds, the current insomnia subtype is determined to be the neural-metabolic decoupling type. In step S4, the main controller 1 generates a targeted physical field intervention strategy based on the diagnostic results of step S3, and converts the strategy into specific control instructions for the closed-loop topology insulator intervention module 8; the closed-loop topology insulator intervention module 8 performs non-contact physical field intervention based on the control instructions. If diagnosed as psychogenic hyperarousal dominant insomnia, the main controller 1 generates an intervention strategy aimed at inhibiting hyperarousal EEG patterns and disrupting abnormal causal connections. The main controller 1 sends control commands to the closed-loop topological insulator intervention module 8 through its programmable logic side general input / output interface, causing it to generate a modulation field that is out of phase with the hyperarousal EEG pattern estimated by inversion of the cortical functional connection disintegration rate characteristics. The amplitude of the intervention field is proportional to the causal dominance index of the cortical functional connection disintegration rate characteristics, and its frequency is locked to the peak frequency of the inverted EEG pattern. If the diagnosis is physiological rhythm disorder-dominated insomnia, the main controller 1 generates an intervention strategy aimed at reducing the gain of the abnormal metabolic-respiratory loop. The main controller 1 sends control commands to the closed-loop topology insulator intervention module 8 through its programmable logic side general input / output interface, so that it generates a rhythmic traction field that is synchronized with the real-time respiratory rate and is one-quarter cycle ahead in phase. The frequency of the intervention field is equal to the current respiratory rate, and its intensity is proportional to the difference between the gain of the abnormal loop and the value of 1. If diagnosed as neuro-metabolic decoupling insomnia, the main controller 1 generates an intervention strategy aimed at enhancing the information coupling between neural and metabolic features. The main controller 1 sends control commands to the closed-loop topology insulator intervention module 8 through its programmable logic side general input / output interface, causing it to generate a low-frequency broadband random resonant noise field. The intensity of the noise field is proportional to the difference between the fourth dynamic threshold and the current nonlinear information transmission intensity value from the neural magnetic fluctuation complexity feature node to the metabolic oscillation main frequency feature node. Step S5: After a preset time following the intervention, the main controller 1 controls the perturbation sensor array 2, background field pickup antenna array 3, metasurface wavefront sensing and phase conjugate plate 4, distributed noise phase-locked amplification array 5, nonlinear decoupling engine 6, and optical phase-locked thermodynamic imaging loop 7 to reacquire data, evaluate the changes in key indicators of the causal entropy network model, calculate the network reconstruction degree to quantify the intervention effect, and use closed-loop data to adaptively optimize the diagnostic model and intervention parameters. When the preset time window is reached after the intervention begins in the closed-loop topology insulator intervention module 8, the main controller 1 re-executes steps S1 to S2 to obtain the causal entropy network model after intervention. The main controller 1 calculates the change in weights of the critical anomalous edges identified in step S3 after intervention, and calculates the Frobenius norm distance between the network model after intervention and the user's healthy sleep benchmark network model stored in the main controller 1. The main controller 1 determines that if the weight change of the key abnormal edge is in the desired direction and the Frobenius norm distance decreases, then the intervention is deemed effective. The main controller 1 stores the network status data before and after the intervention, the executed intervention instructions, and the effect evaluation results in its associated storage unit for periodic retraining and optimization of various dynamic thresholds in step S3 and various proportional coefficients in step S4.

[0029] When the system is working, the main controller 1 writes parameters to peripherals that need to be configured, such as the direct digital frequency synthesizer and the LCD driver chip, through its internal integrated circuit interface on the processing system side, and loads firmware programs to the first and second field programmable gate arrays through the high-speed expansion interface on the programmable logic side, thus completing the startup preparation of the entire system.

[0030] When the system enters the working state, its non-contact detection process immediately unfolds in multiple dimensions. The quantum perturbation sensor array 2 in the perturbation sensor array 2 starts to work. Its gallium arsenide and aluminum gallium arsenide superlattice heterojunction enters a metastable state under a precise bias voltage, and generates a coupling response to the extremely low frequency physiological phonons naturally emitted by the human body during sleep. The extremely low frequency physiological phonons are quasi-particle vibrations originating from life activities such as cell metabolism. This coupling will perturb the energy level distribution of electrons in the heterojunction, causing a measurable change in tunneling current. This pA-level current is captured by the first transimpedance amplifier and converted into a voltage signal. This process captures the oscillation information of the human body's deep metabolic activities non-contactly.

[0031] Almost simultaneously, the ultra-wideband chip antenna array in the background field pickup antenna array 3 begins to capture the chaotic electromagnetic background field of the environment modulated by human bio-electromagnetic activity. After being amplified by the first low-noise amplifier, the signal is down-converted to intermediate frequency by the first mixer, and then conditioned by the first variable gain amplifier before being digitized by the first ultra-high-speed analog-to-digital converter. This process acquires electromagnetic radiation information reflecting the synchronicity and chaotic characteristics of neural clusters in different regions of the brain without contact.

[0032] Simultaneously, the metasurface wavefront sensing plate in the metasurface wavefront sensing and phase conjugate plate 4 begins to sense the wavefront phase distortion generated by the environmental radio frequency signal in the sleep environment after encountering the human body. These subtle distortions are captured and demodulated by the reference antenna and the first integrated transceiver. Meanwhile, the metasurface unit integrated around the bed or mattress, under the control of the audio digital signal processing chip and the first high-voltage digital output driver, adjusts the orientation of the liquid crystal molecules in real time to change their electromagnetic properties, and detects the spatial electromagnetic field disturbances caused by breathing fluctuations and limb micro-movements in a non-contact manner.

[0033] Inside the cryostat, the Josephson junction array in the distributed noise lock-in amplifier array 5 operates at liquid nitrogen temperature. It is extremely sensitive to weak magnetic fields at the fettesta level generated by neural currents. Its quantum state changes due to these magnetic field changes and is converted into a voltage signal by the readout circuit composed of the first high-speed current source and the first instrumentation amplifier. This signal is then synchronously acquired by the first synchronous sampling analog-to-digital converter array, realizing non-contact correlation measurement of the magnetic field of deep neural activity.

[0034] The single-photon avalanche diode array in the optical phase-locked thermodynamic imaging ring 7 is driven by the modulation signal generated by the direct digital frequency synthesizer to count the extremely weak infrared thermal radiation photons of a specific brain region of the human body within its field of view. By statistically analyzing the distribution of the time intervals between photon arrivals, the thermodynamic entropy change of the local tissue can be inferred non-contactly.

[0035] All these raw data streams from different physical dimensions, which are redundant yet complementary to each other, are strictly synchronized at the nanosecond level under the timestamp generated by the system master clock generator uniformly distributed by the main controller 1. They are continuously fed into the first field-programmable gate array, the second field-programmable gate array, and the main controller 1 through various interfaces and parallel buses. Thus, the system has completed the all-round, non-intrusive, and high-fidelity acquisition of the human body's multidimensional physical field information in the natural sleep state.

[0036] The massive amount of raw data collected is then fed into the causal entropy network diagnosis and target field intervention algorithm process led by the main controller 1.

[0037] In the first step of the algorithm, namely multi-physics synchronous sensing and field feature transformation, the main controller 1 and its coprocessor field programmable gate array perform real-time digital signal processing on the original data stream.

[0038] They perform fast Fourier transform and Welch power spectrum estimation on quantum perturbation signals to extract the metabolic oscillation main frequency and its spectral purity that characterize the cellular metabolic rhythm.

[0039] Wavelet coherence analysis and exponential fitting were performed on chaotic electromagnetic field signals to calculate the cortical functional connectivity disintegration rate, which reflects the stability of functional connectivity between regions of the cerebral cortex.

[0040] By performing fundamental frequency separation and event detection on wavefront distortion signals, we can obtain respiratory harmonic distortion, which indicates respiratory quality, and micromotion event entropy, which reflects the degree of bodily relaxation.

[0041] By performing eigenvalue decomposition and entropy calculation on the quantum noise cross-correlation matrix, the neural magnetic fluctuation correlation dimension, which reflects the complexity of neural cluster activity, is derived.

[0042] Statistical mechanical analysis was performed on the photon arrival time series to calculate the thermodynamic entropy difference between the orbitofrontal and motor cortex, which characterizes the differences in metabolic efficiency between different brain regions.

[0043] These feature extraction processes transform raw, seemingly unrelated physical quantities into a set of standardized digital features with clear physiological and psychological meanings, which together form a dynamically updated multidimensional feature vector. The effect of this step is to unify multimodal physical observations into biological digital twin signals that can be used for advanced analysis, laying a data foundation for subsequent diagnosis.

[0044] Next, the second step of the algorithm, the construction and update of the dynamic causal entropy network, is initiated. The main controller 1 uses its powerful computing power to perform vector autoregression modeling and transfer entropy calculation on the time series of the above multidimensional feature vectors. It analyzes whether there is a predictive relationship and nonlinear information flow between any two features, and quantifies these two relationships into weights, thereby dynamically generating a causal entropy network graph model with features as nodes and causality and information flow as directed edges.

[0045] It has achieved the modeling of the complex system problem of insomnia, linking scattered features into an organic whole, thereby enabling insights into the driving chains and coupling relationships between various physiological factors behind insomnia.

[0046] The algorithm then moves to its third step: network dynamics analysis and insomnia root cause diagnosis. The main controller 1 performs deep graph analysis on this real-time updated causal entropy network. It calculates the causal dominance index of each network node to find the pivot points that drive changes in the network state. For example, it finds that the rate of disintegration of cortical functional connectivity is the node that strongly drives other nodes, rather than the other way around.

[0047] It searches for strong connection loops in the network and calculates their loop gain to determine whether there are self-reinforcing anomalous oscillations. For example, it finds a vicious cycle with a gain greater than one between the dominant frequency of metabolic oscillations and the distortion of respiratory harmonics.

[0048] Then, it compares the analysis results with personalized dynamic thresholds learned from long-term user data, and matches them using a pre-set diagnostic rule base with embedded medical knowledge to accurately determine the subtype of the current insomnia state. For example, it may determine whether it is a psychogenic hyperarousal type caused by excessive activity of the cerebral cortex due to anxious thoughts, which in turn drives respiratory disorders; a physiological rhythm disorder type caused by decoupling of the internal biological clock and respiratory rhythm, forming a self-disrupting circuit; or a neurometabolic decoupling type caused by the loss of normal communication between brain neural electrical activity and deep cell metabolic signals. As a result, the system no longer simply outputs how long you slept or how many times you woke up, but diagnoses the root cause of insomnia and accurately distinguishes sleep disorders of different natures and sources.

[0049] Once the diagnosis is complete, the fourth step of the algorithm, the generation of personalized intervention strategies based on network modulation, is triggered. The main controller 1 calls the corresponding strategy mapping function according to the diagnosed insomnia subtype. This function translates the abstract network modulation goals, such as inhibiting high-drive nodes, breaking abnormal loops, and strengthening weak connections, into specific, executable physical instruction parameters for the closed-loop topological insulator intervention module 8. For example, for the psychogenic hyperarousal dominant type, the instruction is to generate a topological insulator surface exciton field that is precisely out of phase with the real-time inverted hyperarousal EEG pattern, whose amplitude is proportional to the causal dominance index, and whose frequency is locked.

[0050] For the physiological rhythm disorder-dominant type, the instruction is to generate a rhythmic traction field that is synchronized with the current respiratory rate but leads by 90 degrees in phase and whose intensity is proportional to the overshoot of the abnormal circuit.

[0051] For the decoupled neuro-metabolic type, the instruction is to generate a low-frequency broadband random resonant noise field whose intensity is proportional to the degree of information loss. The effect of this step is to non-destructively transform the diagnostic conclusions of the digital world into an energy intervention prescription that can be applied in the physical world.

[0052] The prescription is the instruction. The closed-loop topological insulator intervention module 8 immediately takes action. Its isolated gate driver receives pulse instructions from the main controller 1 and precisely controls the full-bridge circuit composed of four gallium nitride power transistors to inject high-speed, high-current specific waveform pulses into the topological insulator thin film electrode. This excites phonon polaritons on the surface of the thin film with propagation direction and mode protected by the topology. This special energy wave radiates to the human body in a non-contact manner and interacts specifically with the body's own disordered physiological field. For example, the anti-phase modulation field aims to counteract oversynchronized brain electrical activity, the rhythm traction field aims to bring the detuned breathing rhythm back on track, and the random resonance field aims to enhance the perception of weak metabolic signals to promote system recoupling. This is the specific sleep aid method recommended by the system. It is not just general music or white noise, but a highly personalized physical targeted therapy based on in-depth etiological diagnosis and energy form and parameters.

[0053] Intervention is not the end point. The fifth step of the algorithm, network reconstruction degree evaluation and model parameter adaptation, follows immediately. After a period of intervention, the main controller 1 restarts the first optical phase-locked thermodynamic imaging loop 7 to carry out a new round of data acquisition and feature extraction, reconstructs the causal entropy network after intervention, and objectively and quantitatively evaluates the immediate effect of this intervention by calculating the changes in the weights of key edges and comparing the overall distance between the network after intervention and the user's personal health baseline network, i.e., the Frobenius norm.

[0054] Regardless of the effectiveness, the complete closed-loop data of perception, diagnosis, intervention, and re-perception will be recorded by the system. This data will be used to regularly update the user's personal baseline model and optimize the proportions of diagnostic thresholds and intervention parameters. This gives the system the ability to learn and adapt to the user's unique physiological response patterns as the usage time increases, thereby continuously optimizing its ability to solve the user's insomnia problem in the long term.

[0055] As a result, the entire process, from insomnia detection and root cause determination to personalized sleep intervention recommendations, can be completed contactlessly without the need for hospital examinations.

[0056] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent multimodal insomnia intervention system based on sleep monitoring, characterized in that, The system includes a main controller (1), which is connected to the data output terminals of a perturbation sensor head array (2), a background field pickup antenna array (3), a distributed noise phase-locked amplifier array (5), and an optical phase-locked thermodynamic imaging ring (7) via a high-speed data interface. The main controller (1) is connected to the control terminal of a metasurface wavefront sensor and a phase conjugate plate (4) via a configuration bus. The main controller (1) is connected to a nonlinear decoupling engine (6) via a high-speed expansion interface. The main controller (1) is connected to a closed-loop topological insulator intervention module (8) via a control signal interface. The main controller (1) coordinates the synchronous acquisition of signals by each functional module, processes and fuses the original signals into a multidimensional feature vector, constructs and dynamically updates the causal entropy network model based on this, analyzes the topology and dynamic characteristics of the network to diagnose the immediate dominant subtype and root cause of insomnia, generates targeted intervention strategies based on the diagnostic results and converts them into control instructions, re-acquires data after intervention to evaluate the effect and adaptively optimizes the model parameters. The perturbation sensor array (2) collects the extremely low frequency physiological phonon signals naturally emitted by the human body during sleep and converts them into electrical signals that characterize cellular metabolic oscillations. The background field pickup antenna array (3) collects chaotic background electromagnetic field signals formed by human bioelectromagnetic activity modulation to reflect the activity synchronicity and chaotic characteristics of the cerebral cortex. The metasurface wavefront sensor and phase conjugate plate (4) collects the wavefront phase distortion information of environmental radio frequency signals caused by human breathing and micro-movement, and can adjust its unit state under the control of the main controller (1) to participate in sensing. The distributed noise lock-in amplifier array (5) collects ultra-weak magnetic field signals generated by nerve currents and outputs quantum noise data reflecting the activity of nerve clusters through multi-channel correlation measurement; The nonlinear decoupling engine (6) receives the feature data stream after front-end processing and calculates the causal and information transmission relationship between feature pairs at high speed through its dedicated hardware circuit to assist the main controller (1) in constructing a causal entropy network; The optical phase-locked thermodynamic imaging ring (7) collects infrared thermal radiation photons from specific brain regions of the human body and obtains thermodynamic entropy change information characterizing the metabolic efficiency of local tissues through phase-locked photon counting statistics. The closed-loop topological insulator intervention module (8) receives and executes specific control commands issued by the main controller (1) to drive the topological insulator thin film to generate non-contact physical field intervention with specific waveforms, intensities and frequencies to act on the human body to achieve targeted regulation.

2. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The perturbation sensor array (2) includes multiple heterojunction sensing units, a first precision voltage reference source, a first digital potentiometer, a first buffer amplifier, a first transimpedance amplifier, a first analog multiplexer, a first programmable filter, and a first analog-to-digital converter. The positive current terminal of the heterojunction sensing unit is connected to the inverting input terminal of the first transimpedance amplifier through a first current-limiting resistor. The negative current terminals of the heterojunction sensing units are connected to a first negative bias reference ground. The positive voltage output pin of the first precision voltage reference source is connected to the power input pin of the first digital potentiometer. The sliding output pin of the first digital potentiometer is connected to the non-inverting input terminal of the first buffer amplifier. The output pin of the first buffer amplifier provides a negative bias voltage to the first negative bias reference ground network. The serial interface pin of the first digital potentiometer is connected to the main controller (1). The first serial peripheral interface on the processing system side has a first feedback resistor and a first feedback capacitor connected in parallel between the inverting input terminal and the output terminal of the first transimpedance amplifier. Its output pin is connected to the common input pin of the first analog multiplexer. The multiple channel selection output pins of the first analog multiplexer are connected in parallel and then connected to the signal input pin of the first programmable filter. Its channel address selection pin is connected to the general-purpose input / output pin on the processing system side of the main controller (1). The signal output pin of the first programmable filter is connected to the analog input positive pin of the first analog-to-digital converter. The serial clock pin, serial data input pin, and serial data output pin of the first analog-to-digital converter are respectively connected to the corresponding pins of the second serial peripheral interface on the programmable logic side of the main controller (1). Its conversion start control pin is driven by the general-purpose input / output pin of the main controller (1).

3. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The background field pickup antenna array (3) includes multiple ultra-wideband chip antennas, a high-frequency circuit board, multiple first low-noise amplifiers, multiple first mixers, a first frequency synthesizer, multiple first variable gain amplifiers, a first ultra-high-speed analog-to-digital converter, and a first field-programmable gate array. The signal output pin of each ultra-wideband chip antenna is soldered to the feed point of the high-frequency circuit board and connected to the RF input pin of the corresponding first low-noise amplifier through a microstrip transmission line. The RF output pin of each first low-noise amplifier is connected to the RF input pin of the corresponding first mixer through a microstrip transmission line. The local oscillator input pin of the first mixer is connected in parallel to the RF output pin of the first frequency synthesizer through a power distribution network. The clock pin and data pin of the serial configuration interface of the first frequency synthesizer are connected to the second serial peripheral interface on the processing system side of the main controller (1). The intermediate frequency output pin of each first mixer is connected to the second serial peripheral interface on the processing system side of the main controller (1). The differential input positive pin and differential input negative pin of the first variable gain amplifier are connected via coaxial cable or differential trace. The differential output positive pin and differential output negative pin of each of the first variable gain amplifiers are connected to the analog input differential channel positive pin and negative pin of the first ultra-high speed analog-to-digital converter via parallel differential pairs. All conversion channels of the first ultra-high speed analog-to-digital converter share an external reference clock input pin to receive the sampling clock signal. Its multiple pairs of transmit differential data positive pins and negative pins are connected to the corresponding multiple pairs of gigabit transceivers on the first field-programmable gate array via differential line pairs on the circuit board to receive differential data positive pins and negative pins. The synchronization control pin and power management pin of the first ultra-high speed analog-to-digital converter are connected to the general-purpose input and output pins of the first field-programmable gate array. Its digital output data stream is preprocessed by the first field-programmable gate array and then transmitted to the main controller (1) via a high-speed expansion interface.

4. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The metasurface wavefront sensing and phase conjugate plate (4) includes multiple liquid crystal metasurface units, a first high-voltage driving chip, a first integrated transceiver, an audio digital signal processing chip, a first high-voltage digital output driver, and at least one reference antenna. The upper electrode pin of each liquid crystal metasurface unit is connected to a corresponding current output pin of the first high-voltage driving chip, and the lower electrode pin of each liquid crystal metasurface unit is connected to the system analog ground. The serial clock pin and serial data pin of the first high-voltage driving chip are connected to the first internal integrated circuit interface on the processing system side of the main controller (1), and its multiple high-voltage output pins are respectively connected to the corresponding liquid crystal metasurface units. The upper electrode pin of the surface cell, the reference antenna is connected to the first receive channel input pin of the first integrated transceiver via a coaxial cable, the local oscillator and transmit sections of the first integrated transceiver are disabled, the data input port of the first integrated transceiver is connected to the audio digital signal processing chip, the multiple general-purpose input / output extension pins of the audio digital signal processing chip are connected to the parallel digital input pins of the first high-voltage digital output driver, the first high-voltage digital output driver receives the parallel digital control word from the audio digital signal processing chip and connects its corresponding multiple high-level digital output pins to the upper electrode pins of the corresponding liquid crystal metasurface cell via a second current-limiting resistor.

5. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The distributed noise lock-in amplifier array (5) includes multiple Josephson junction sensing units, a cryostat, multiple first high-speed current sources, multiple first instrumentation amplifiers, a first synchronous sampling analog-to-digital converter array, and a second field-programmable gate array. The two electrodes of each Josephson junction sensing unit are connected via superconducting wires inside the cryostat to the current output pin of the corresponding first high-speed current source and the non-inverting input of the first instrumentation amplifier through a first cryogenic coaxial connector. The shielding layer of each first cryogenic coaxial connector is connected to an independent low-noise reference ground. The digital control interface of each first high-speed current source is connected to the third serial peripheral interface on the processing system side of the main controller (1). The output pin of the first instrumentation amplifier is connected to the analog input positive pin of the analog-to-digital converter in the first synchronous sampling analog-to-digital converter array. The reference ground pin of the first instrumentation amplifier is connected to the same low-noise analog ground plane. The synchronous sampling control pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are interconnected and receive the sampling clock signal together. The serial data output pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are connected in parallel to the serial peripheral interface master input slave output pin of the second field-programmable gate array. The independent chip select pins of the analog-to-digital converters in the first synchronous sampling analog-to-digital converter array are controlled by different general-purpose input / output pins of the second field-programmable gate array.

6. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The nonlinear decoupling engine (6) includes a dedicated causal computing integrated circuit and a high-speed data buffer memory. The multi-channel high-speed parallel data input bus of the dedicated causal computing integrated circuit is connected to the general-purpose input / output pin group of the second field-programmable gate array in the distributed noise lock-in amplifier array (5) through the parallel bus on the circuit board. The dedicated causal computing integrated circuit is connected to the fourth serial peripheral interface on the processing system side of the main controller (1) through the configuration bus. The calculation result output of the dedicated causal computing integrated circuit is connected to the corresponding high-speed peripheral component interconnection interface pin on the programmable logic side of the main controller (1) through multiple pairs of transmit differential pins and receive differential pins of its high-speed peripheral component interconnection data link. The data pins, address pins and control pins of the high-speed data buffer memory are respectively connected to the corresponding external memory interface pins of the dedicated causal computing integrated circuit.

7. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The optical phase-locked thermodynamic imaging ring (7) includes a single-photon avalanche diode array, a matching readout chip, a direct digital frequency synthesizer, and an electro-optic phase modulator array. The single-photon avalanche diode array is integrated by multiple single-photon avalanche diode pixels in a two-dimensional matrix. The serial peripheral interface clock pin, serial data input pin, and serial data output pin of the matching readout chip are connected to the corresponding pins of the third serial peripheral interface on the programmable logic side of the first field-programmable gate array in the background field pickup antenna array (3). The multiple digital pulse output pins of the single-photon avalanche diode array are respectively connected to the multiple dedicated general-purpose input and output pins of the first field-programmable gate array in the background field pickup antenna array (3). The serial configuration interface of the direct digital frequency synthesizer is connected to the first internal integrated circuit interface on the processing system side of the main controller (1). The positive and negative ends of its differential analog output pins are connected to the driving electrodes of the electro-optic phase modulator array. The electro-optic phase modulator array is placed at the front end of the optical receiving path of the single-photon avalanche diode array.

8. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, The closed-loop topological insulator intervention module (8) includes a topological insulator film, a first power electrode, a second power electrode, an isolated gate driver, a full-bridge power amplifier circuit, a current monitoring amplifier, and a sampling resistor. The first edge of the topological insulator film is in ohmic contact with the first power electrode, and its second edge is in ohmic contact with the second power electrode. The first and second power electrodes are respectively welded to the first and second power output nodes of the full-bridge power amplifier circuit. The full-bridge power amplifier circuit is composed of four gallium nitride power transistors in a full-bridge topology. The connection point between the first high-side transistor and the first low-side transistor forms the first power output node, and the second high-side transistor... The connection point between the transistor on the first low side and the transistor on the second low side forms the second power output node. The gate drive pins of the four transistors in the full-bridge power amplifier circuit are respectively connected to the four independent output pins of the isolated gate driver. The input power supply and logic input pins of the isolated gate driver are connected to the programmable logic side general-purpose input / output pin group of the main controller (1). The sampling resistor is connected in series between the first power output node and the first power electrode. The voltage difference between its two ends is connected to the non-inverting input terminal and the inverting input terminal of the current monitoring amplifier. The voltage output pin of the current monitoring amplifier is connected to the second analog input channel pin of the first analog-to-digital converter in the perturbation sensor head array (2).

9. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 1, characterized in that, It also includes a system-level clock circuit (9), the output of which is connected to the main controller (1) and the clock input of each module to provide a unified synchronous clock reference.

10. The intelligent multimodal insomnia intervention system based on sleep monitoring according to claim 9, characterized in that, The system-level clock circuit (9) includes a system master clock generator and a clock buffer. The first clock output terminal of the system master clock generator is connected to the reference clock input pin of the processing system side of the main controller (1). The second clock output terminal of the system master clock generator is connected to the clock input of the clock buffer. The multiple clock output terminals of the clock buffer are respectively connected to the reference clock input pins of the first ultra-high-speed analog-to-digital converter in the background field pickup antenna array (3), the first synchronous sampling analog-to-digital converter array in the distributed noise phase-locked amplifier array (5), and the direct digital frequency synthesizer in the optical phase-locked thermodynamic imaging ring (7). The third clock output terminal of the system master clock generator is connected to the reference clock input pin of the first integrated transceiver in the metasurface wavefront sensing and phase conjugate plate (4).

Images