Wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm

Abstract

The present invention discloses a fully wireless ocular surface pressure monitoring system based on a multi-frequency phase-shift decoupling self-calibration algorithm, relating to the technical field of intelligent ophthalmic detection. The system comprises a wearable three-layer contact lens, wherein the outer layer is embedded with a capacitive eyelid pressure sensor and a first spiral coil, and the inner layer is embedded with a piezoresistive intraocular pressure sensor and a second spiral coil, which are respectively tuned to different center frequencies to modulate pressure changes into radio frequency signals. Externally powered receiving glasses or eye masks comprise a multi-frequency transmitting coil, a rectifier circuit and a receiving coil, which supply energy to the sensing structure in the contact lens and sense and obtain a composite signal formed by spatial superposition, which is transmitted to a processing circuit and then sent to a mobile computing terminal. The terminal performs multi-frequency decoupling, sparse matrix separation and temporal self-calibration processing to generate calibration curves of eyelid pressure and intraocular pressure, thereby realizing real-time perception of dual parameters in the eyes-closed state. The system has the advantages of compact structure, high signal separation accuracy and high degree of wirelessness.

Description

Technical Field

[0001] The present invention relates to the technical field of ocular surface pressure monitoring, and in particular to a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm. Background Art

[0002] In existing technologies, ocular surface pressure, an important indicator for assessing ocular pathologies such as glaucoma, dry eye, and eyelid diseases, is primarily measured using two physical quantities: intraocular pressure and eyelid pressure. Continuous dynamic monitoring of intraocular pressure typically relies on contact lens-based solutions based on strain gauges, such as the Triggerfish system, which indirectly reflects intraocular pressure fluctuations by sensing changes in corneal curvature. Eyelid pressure detection, on the other hand, is often performed using an adhesive or wired method, with the sensing structure often integrated onto the outside of the eyelid, acquiring pressure data through capacitance, resistance, and other methods. Some systems combine wireless transmission with flexible electronics design ideas to attempt to achieve remote monitoring and cloud-based analysis of ocular surface parameters.

[0003] However, the above technologies still face several bottlenecks in practical applications. On the one hand, current eyelid pressure measurements mostly rely on wired solutions, which are uncomfortable to wear and have low patient cooperation, making dynamic monitoring difficult. On the other hand, existing contact lens-based intraocular pressure detection methods require an external wired receiver, resulting in low system integration and the inability to achieve full wireless operation. In addition, since eyelid pressure and intraocular pressure act on the ocular surface simultaneously when the eyes are closed, existing methods have difficulty in effectively decoupling their signals, resulting in large measurement errors and a lack of high-resolution, low-interference dual-parameter independent recognition capabilities.

[0004] Therefore, there is an urgent need to propose a new ocular surface pressure measurement solution with higher integration, excellent decoupling performance, and suitable for dynamic monitoring in the closed eye state. Summary of the Invention

[0005] The present application provides a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm to improve the accuracy and anti-interference ability of ocular surface pressure monitoring.

[0006] The present application provides a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm, comprising:

[0007] A wearable contact lens having a three-layer structure, with a hard lens in the middle for support and shielding. The outer soft lens is embedded with a capacitive eyelid pressure sensor and a first spiral coil, and the inner soft lens is embedded with a piezoresistive intraocular pressure sensor and a second spiral coil. The first spiral coil and the second spiral coil are respectively tuned to different center frequencies and are used to modulate pressure changes into radio frequency signals of corresponding frequencies when coupled with their respective sensors.

[0008] Externally powered receiving glasses or eye masks, comprising a multi-frequency transmitting coil electromagnetically coupled to the first and second spiral coils, a power supply rectifier circuit, and a receiving coil, wherein the transmitting coil provides energy to the eyelid pressure sensor and the first spiral coil, and to the intraocular pressure sensor and the second spiral coil, respectively; the receiving coil is configured to sense and acquire a composite signal formed by respectively modulating and spatially superimposing the two sensor coil structures, and provide the composite signal to a processing circuit, which transmits the processed composite signal to a mobile computing terminal;

[0009] The mobile computing terminal is used to receive the processed composite signal, execute a multi-frequency phase shift decoupling algorithm to extract the corresponding frequency components, and separate the eyelid pressure signal and the intraocular pressure signal through sparse matrix decomposition; based on the continuous time series of the separated signals, a timing self-calibration algorithm is executed in combination with a preset corneal biomechanical model to correct the measurement errors caused by the closed eye state or interference, and generate a calibrated eyelid pressure curve and intraocular pressure curve.

[0010] The beneficial effects of this application mainly include: (1) By embedding a capacitive eyelid pressure sensor and a piezoresistive intraocular pressure sensor in a wearable contact lens, and configuring two spiral coils tuned to different center frequencies, the two types of pressure signals are modulated into radio frequency signals of different frequencies, thereby achieving synchronous and independent acquisition of two key physiological parameters of the ocular surface. (2) By setting up external power supply and receiving glasses or eye masks, which contain a multi-frequency transmitting coil, a power supply rectifier circuit and a receiving coil, it is possible to supply energy to the two sets of sensor coil structures in the contact lens respectively and sense the radio frequency signals emitted by them, thereby achieving a fully wireless energy and data path without batteries or wired connections at the sensing end. (3) The mobile computing terminal performs a multi-frequency phase shift decoupling algorithm and sparse matrix decomposition on the received composite signal, which can effectively extract the eyelid pressure signal and the intraocular pressure signal from the composite signal containing multiple frequency components, improve the decoupling accuracy, and avoid measurement deviations caused by signal aliasing. (4) By introducing a time-series self-calibration algorithm based on the corneal biomechanical model into the mobile computing terminal, the measurement errors caused by eye closure, eyelid contact changes or other interference factors can be automatically corrected according to the continuous time series data, significantly improving the stability of the system and the data credibility in real wearing scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a schematic diagram of a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm provided in the first embodiment of the present application. DETAILED DESCRIPTION

[0012] The following description sets forth many specific details to facilitate a thorough understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of the present application. Therefore, the present application is not limited to the specific implementations disclosed below.

[0013] The first embodiment of the present application provides a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm. Figure 1 , which is a schematic diagram of the first embodiment of this application. Figure 1 The first embodiment of the present application provides a wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm, which is described in detail.

[0014] The wireless ocular surface pressure monitoring system based on the multi-frequency phase shift decoupling self-calibration algorithm includes a wearable contact lens 101 , external power supply and receiving glasses or eye masks 102 , and a mobile computing terminal 103 .

[0015] The wearable contact lens 101 adopts a three-layer structure, with a hard lens in the middle for support and shielding. The outer soft lens is embedded with a capacitive eyelid pressure sensor and a first spiral coil, and the inner soft lens is embedded with a piezoresistive intraocular pressure sensor and a second spiral coil. The first spiral coil and the second spiral coil are tuned to different center frequencies respectively, and are used to modulate pressure changes into radio frequency signals of corresponding frequencies when coupled with their respective sensors.

[0016] In the fully wireless ocular surface pressure monitoring system described in the present invention, the wearable contact lens 101 constitutes the core of the front-end sensing structure. Its structural design, functional division, and material integration method must have a high degree of precision and biocompatibility to ensure that while achieving the ocular surface pressure collection function, it does not affect the stability of normal vision and corneal physiological environment.

[0017] The wearable contact lens 101 utilizes a three-layer composite structure, consisting of an outer soft lens, a middle hard lens, and an inner soft lens, arranged from outside to inside. The middle layer is a hard contact lens, typically made of polymethyl methacrylate (PMMA) or a silicone hard lens material with high oxygen permeability. Its primary function is to provide mechanical structural support while also shielding and isolating electromagnetic signals of varying frequencies. The peripheral contour of the hard lens must closely match the curvature of the wearer's cornea to prevent warping, slippage, or a foreign body sensation. Plasma or UV surface activation treatment is used to create a sufficient density of active groups on the surface to ensure interfacial bonding strength with the soft lens.

[0018] A capacitive eyelid pressure sensor and a first spiral coil are pre-embedded in the outer layer of the soft lens. The capacitive sensor is constructed with highly transparent materials, such as reduced graphene oxide (rGO) electrodes and Ecoflex 0030 elastomer to form a deformable membrane capacitor structure. The sensing unit is placed in the center or upper edge of the front surface of the soft lens, and can sense the mechanical pressure applied to the eye surface when the eyelids are closed or in contact. This pressure will cause the sensing capacitance value to change, and the modulation function is achieved through the resonant circuit formed by coupling with the first spiral coil. The first spiral coil is prepared from a patterned silver nanowire grid and transferred to a flexible polyimide (PI) substrate. Its coil layout is a compact spiral structure that surrounds the pupil opening of the soft lens to ensure that it does not block the visual light path. The resonant frequency of the coil is designed and tuned to a specific frequency , avoid the inner coil Interference occurs.

[0019] The inner soft lens is embedded with a piezoresistive intraocular pressure sensor and a second spiral coil. The piezoresistive sensor is usually constructed using a nanocomposite material with pressure-sensitive response characteristics, such as a composite structure of gold nanoflowers and two-dimensional titanium carbide aerogel. Its resistance changes with the pressure applied, and is used to sense the tension change from the inner side of the cornea to the outside, thereby indirectly reflecting the dynamic changes in intraocular pressure. The second spiral coil and the piezoresistive sensor together form a sensor tuned to the frequency The LC circuit, the frequency With the first coil There is sufficient frequency separation to allow subsequent signal decoupling through frequency domain analysis. The second coil is also made of flexible conductive material and surrounds the non-optical axis area of ​​the inner soft mirror to avoid interfering with visual light transmission while ensuring sufficient sensing area and coupling efficiency.

[0020] The three lens layers are compositely bonded together using methods such as light-curing UV adhesive, plasma treatment, or surface grafting, resulting in a structurally stable, uniformly thick, monolithic lens. To enhance wearing comfort and biocompatibility, the outer and inner soft lens layers are preferably made of silicone hydrogel with a high hydrogel content, ensuring an oxygen permeability of at least 80 Dk / t and meeting the physiological needs of the cornea during all-day wear.

[0021] The overall thickness of the wearable contact lens 101 is controlled within a range of approximately 150-300 microns, ensuring that the sensor and coil structures can be embedded without affecting the lens's flexibility and fit. The outer soft lens layer is typically 50-80 microns thick, the middle hard lens is 80-120 microns, and the inner soft lens is 40-70 microns thick. The electrical connection between the sensor and coil is achieved using nanowires or flexible microwiring processes, forming a complete sensing and modulation structure.

[0022] When worn, the contact lens 101 maintains a stable contact with the corneal surface without affecting visual function. Simultaneously, the dual-sensing structure detects eyelid pressure and intraocular pressure changes, modulating each signal into radio frequency signals of different frequencies. This signal is then sensed, decoded, and processed by an external receiving structure. This structure achieves a high degree of integration of dual-channel signal sources while maintaining wearer comfort, providing a foundation for high-resolution, continuous, and non-invasive ocular surface pressure monitoring.

[0023] Furthermore, the outer soft lens of the wearable contact lens is made of a highly translucent silicone hydrogel material; the capacitive eyelid pressure sensor is disposed in a non-central area of ​​the outer soft lens near the upper eyelid; the first spiral coil is formed of a grid structure composed of silver nanowires, which is patterned by laser and then transferred to a polyimide flexible film substrate, and is fixed in a ring-shaped manner in a non-optical axis area at the edge of the outer soft lens;

[0024] The non-optical axis annular region where the first spiral coil is located is provided with an array of micro-through holes that penetrate the thickness direction of the outer layer of the soft lens. The diameter of each micro-through hole is less than 50 microns. The through holes are filled with an array of compressible hydrophilic elastomer micro-pillars, which are used to form a buffer deformation layer when shear force is generated by eyelid movement, thereby reducing measurement errors caused by lateral friction of the eyelid.

[0025] The outer soft lens and the middle hard lens are subjected to surface plasma activation treatment and then interface bonded using a two-component UV copolymer photosensitive adhesive to form a stable multi-layer integrated structure. The structure maintains sensor signal drift below 5% during continuous wear for more than 72 hours, effectively improving long-term signal stability and repeatability under conditions of eyelid dynamic interference.

[0026] The outer soft lens of the wearable contact lens is preferably made of a highly translucent silicone hydrogel material. This material not only exhibits excellent softness and biocompatibility, but also possesses a high water content (preferably between 40% and 60%) and oxygen permeability (Dk value greater than 80), ensuring corneal respiration and comfort during prolonged wear. This high translucency helps reduce blurred vision caused by obstruction of the sensor structure and also facilitates the integration of embedded components within the transparent contact lens.

[0027] In terms of structural layout, the capacitive eyelid pressure sensor is placed in a non-central area of ​​the outer soft lens near the upper eyelid coverage area, which is usually located on the transparent zone outside the upper edge of the pupil. When the user's eyes are normally open, this area is within the pressure trajectory of the upper eyelid naturally contacting the lens. Therefore, this position can accurately sense the dynamic pressure changes on the lens caused by eyelid movement without blocking the user's normal visual axis, ensuring that signal acquisition and visual experience do not conflict. The sensor can be composed of a flexible electrode material and a compressible medium, and its capacitance value changes with the change in the membrane distance caused by external pressure.

[0028] The first spiral coil is constructed using a conductive mesh pattern of silver nanowires. Laser patterning is used to precisely etch a pre-defined coil path onto a substrate. This is then transferred to a polyimide flexible film substrate via transfer printing, forming a planar, stretchable, toroidal coil structure. Polyimide, as a flexible, high-temperature substrate, provides long-term, stable adhesion to the surface of the lens while maintaining the integrity of the coil structure and its inductive response. The coil is placed in a circular pattern at the edge of the outer lens, in a non-optical area surrounding the pupil but avoiding the light-transmitting area. This ensures a stable RF signal coupling path without interfering with the field of view.

[0029] To further enhance the outer layer's ability to resist interference when subjected to dynamic eyelid forces, particularly when eyelid movement generates lateral shear forces, an array of micro-through holes is provided in the non-optical-axis annular region where the first spiral coil resides. These micro-through holes extend through the thickness of the outer layer, each with a diameter less than 50 microns. The density of these holes is optimized based on force analysis of the annular region. Each through hole is filled with a compressible hydrophilic elastomeric material, forming a micro-columnar structure. These micro-columns can produce a buffering deformation in the tangential direction when the eyelid slides and applies pressure, absorbing some of the lateral shear stress. This makes the effective pressure applied to the capacitive sensor more stable and consistent, significantly reducing instantaneous capacitance drift or noise errors caused by eyelid displacement.

[0030] The compressible hydrophilic elastomer material utilizes a composite elastic hydrogel system composed of polyvinyl alcohol (PVA) and acrylamide (AAm). The components are prepared in the following proportions: 10 g of polyvinyl alcohol (molecular weight approximately 85,000-124,000, alcohol hydrolysis degree ≥99%) is added to 90 mL of deionized water. The mixture is stirred and heated in a 90°C water bath for 1 hour until completely dissolved, forming a transparent PVA solution. Separately, 3.3 g of acrylamide (AAm) monomer is dissolved in 10 mL of water to form a 33% AAm solution. This is then slowly added to the PVA solution to form a composite monomer solution with a PVA:AAm mass ratio of 3:1.

[0031] To achieve cross-linking, 0.05 g of N,N-methylenebisacrylamide (BIS) was added to the composite monomer solution as a chemical crosslinker, accounting for approximately 0.4% of the total weight. 0.03 g of ammonium persulfate (APS) was also added as an initiator, and 30 µL of N,N,N′,N′-tetramethylethylenediamine (TEMED) was added as a promoter. After preparation, the mixture was quickly poured into a PDMS or polytetrafluoroethylene micromold prefabricated with a through-hole array. The mold was placed in a low-temperature water bath (approximately 10–15°C) for 10 minutes to prevent bubble formation, and then transferred to a 60°C incubator for cross-linking and curing for 4 hours. After curing, the mold was removed and residual monomer was rinsed off, resulting in an elastic hydrogel structure with uniform micropillar array morphology and stable mechanical properties.

[0032] Tensile and compression testing revealed that the PVA-AAm composite elastomer microcolumns exhibit a compression modulus of approximately 35 kPa, good resilience at 20% deformation, and a moisture retention rate exceeding 65%. They also possess good surface affinity with silicone hydrogel lens materials and are non-toxic. During lens use, the material responds to eyelid shearing to produce a buffered deformation. This effectively absorbs the interference component of shear force without affecting vertical pressure transmission, significantly reducing noise and drift in the capacitive sensor and improving the stability of eyelid pressure measurement.

[0033] To achieve reliable bonding between the outer soft lens and the middle hard lens while maintaining the composite structure's transparency, flexibility, and mechanical consistency, the present invention utilizes plasma activation to modify the surfaces of both materials. This activation treatment creates a high density of polar functional groups at the interface, creating sufficient reactive sites for subsequent bonding. Furthermore, a two-component UV copolymer photosensitive adhesive is employed to bond the outer soft lens and the middle hard lens together, using a material that exhibits high transparency, strong adhesion, and UV-curing stability. This two-component UV copolymer photosensitive adhesive can be Norland UV-curing optical adhesive NOA81.

[0034] Multi-parameter test data of the composite bonding structure show that the sensor signal drift is controlled within 5% within a usage cycle of more than 72 hours of lens wearing, indicating that the structure can effectively withstand complex working conditions such as continuous movement of the user's eyelids, tear penetration and temperature fluctuations, and significantly improve the signal stability and repeatability in dynamic interference environments.

[0035] In summary, this embodiment not only provides a contact lens structural design with high reliability, high comfort and anti-interference ability, but also achieves structural stable support for capacitance measurement under eyelid shear dynamics by introducing a microstructure array with elastic response and a stable packaging method.

[0036] The externally powered receiving glasses or eye mask 102 include a multi-frequency transmitting coil, a power supply rectifier circuit, and a receiving coil that are electromagnetically coupled to the first spiral coil and the second spiral coil. The transmitting coil provides energy to the eyelid pressure sensor and the first spiral coil, and to the intraocular pressure sensor and the second spiral coil, respectively. The receiving coil is used to sense and obtain a composite signal formed by respectively modulating and spatially superimposing two groups of sensor coil structures, and provide the composite signal to a processing circuit. The processing circuit sends the processed composite signal to the mobile computing terminal.

[0037] In the fully wireless ocular surface pressure monitoring system described in the present invention, the external power supply and receiving glasses or eye masks 102 undertake the key relay function of energy transmission and signal sensing. Its structural design must not only ensure stable electromagnetic coupling with the two spiral coils in the wearable contact lens, but also ensure high power supply efficiency, accurate signal acquisition, good wearing comfort and reasonable spatial layout, so as to be suitable for dynamic data acquisition in long-term wearing scenarios.

[0038] The externally powered receiving glasses or eye mask 102 typically adopts a dual-arm frame or wrap-around design. The main frame integrates multiple functional components, including a multi-frequency transmitting coil, a power supply rectifier circuit, a receiving coil, and a processing circuit. The multi-frequency transmitting coil comprises at least two independently wound excitation coils, which are spatially aligned with the first and second spiral coils in the contact lens 101. The transmitting frequency corresponds to the resonant frequency of the spiral coils, enabling energy to be transferred to the dual sensing structure within the contact lens through magnetic field coupling without physical contact. This energy transfer process utilizes the principle of resonant induction power supply. An external coil generates an alternating magnetic field, which the spiral coil within the contact lens absorbs at its tuned frequency. This energy is converted to DC power by the lens' internal rectifier circuit, providing low-power, stable power to the corresponding sensors.

[0039] The power supply rectifier circuit is typically integrated into the sides or bridge area of ​​the eyeglass or eye mask frame. Its input is connected to the output of the transmitting coil, and its output provides a constant voltage or current-limited DC power supply in a closed magnetic coupling path. This circuit uses a synchronous rectification structure to improve conversion efficiency and reduce system heat dissipation. Key components include high-frequency, low-loss MOSFETs, high-Q magnetic devices, and flexible PCB packaging. To prevent multi-frequency interference, an electromagnetic isolation structure or shielding plate is installed between the two transmitting channels, and a time-division or frequency-division multiplexing strategy is used for excitation scheduling on the driving signal source.

[0040] The signal receiving coil is placed on the inside of the eyeglasses or eye mask, near the lens area. The coil diameter and mounting position must be precisely aligned with the spatial position of the two modulation sources in the contact lens to ensure maximum sensing of the RF signal modulated by the two sensor coils inside the contact lens. Because the two sets of sensors modulate the pressure signals to different frequencies, the two signals naturally superimpose in space as electromagnetic waves, and the receiving coil senses a composite signal containing two frequency components. The output of the receiving coil is connected to a processing circuit that includes filtering, amplification, analog-to-digital conversion, and preprocessing functions. This circuit can initially organize the composite signal to form a stable digital intermediate signal stream, ready for transmission to the mobile computing terminal for further decoupling and analysis.

[0041] The processing circuitry can be housed within the eyeglass or eye mask frame. It can be implemented using an on-chip microcontroller (MCU), a dedicated demodulation chip, a low-power analog-to-digital converter (ADC), and a Bluetooth or near-field communication module. After completing signal shaping and data regularization, this circuitry transmits the processed composite signal uplink to a mobile computing device via wireless communication. BLE 5.2 is the preferred transmission protocol for its low power consumption, long range, and stability. If the system is designed for near-field wearable scenarios, an NFC module can also be used for instantaneous point-to-point communication.

[0042] The entire external energy-supplying glasses or eye mask 102 should be designed to be as lightweight as possible and conform closely to the facial contours, ensuring no pressure or risk of falling off during prolonged wear. The glasses should feature adjustable nose pads and elastic temples, while the eye mask should preferably be constructed from flexible materials such as memory foam or medical silicone coatings. The internal module should utilize a flexible printed circuit board to adapt to the curvature of the face. All power supply and signal coils and associated circuitry should be enclosed in a magnetically impermeable housing to ensure radiation stability and user safety.

[0043] In summary, the externally powered receiving glasses or eye mask 102, acting as a signal and energy relay platform, play a key role in transmitting information from the contact lens to the mobile terminal. Its functions encompass dual-frequency magnetic coupling power supply, spatial composite signal sensing, signal preprocessing, and communication output, providing essential technical support for the system's fully wireless, dual-parameter decoupled monitoring. This structure addresses multiple design objectives, including wearing comfort, electromagnetic efficiency, and system integration, and is crucial for achieving practical clinical usability.

[0044] Furthermore, the receiving coil in the externally powered receiving glasses or eye mask includes two flexible coil arrays arranged in a spatially staggered manner, which correspond to the preset spatial coupling positions of the first spiral coil and the second spiral coil in the wearable contact lens, respectively. A high-magnetic permeability isolation sheet is provided between the two spatially staggered flexible coil arrays. The high-magnetic permeability isolation sheet is made of a high-performance nanocrystalline alloy material or a flexible soft magnetic composite silicon steel sheet with a magnetic permeability of not less than 2000, and is inserted in a normal direction perpendicular to the two spatially staggered flexible coil arrays to form a magnetic flux shielding path to limit magnetic flux leakage and reduce magnetic coupling crosstalk between the two spatially staggered flexible coil arrays.

[0045] Each of the spatially staggered flexible coil arrays forms an independent resonant receiving circuit with a set of adjustable matching capacitors. The adjustable matching capacitors are varactor devices with voltage-controlled capacitance characteristics. The control input signals are dynamically calculated and issued by the mobile computing terminal based on the amplitude and phase indicators of the current received signal. The control input signals are used to tune the impedance characteristics of the resonant circuit in real time to improve the selectivity of the frequency response curve and compensate for non-ideal coupling offsets in actual wearing conditions.

[0046] A flexible controllable fine-tuning component is arranged between the high magnetic permeability isolation plate and the two spatially staggered flexible coil arrays. The flexible controllable fine-tuning component includes a miniature flexible driving arm made of shape memory alloy material and a repeatedly deformable supporting connecting plate. The flexible controllable fine-tuning component is used to drive the two spatially staggered flexible coil arrays to achieve relative rotation fine-tuning along the normal direction or the lateral direction within a preset angle range, so that when the wearable contact lens is positionally offset or the wearing posture of the external power supply receiving glasses or eye masks changes, the mechanical posture linkage and the capacitance parameter are coordinated to achieve adaptive recovery of the spatial coupling alignment state of the first spiral coil and the second spiral coil, and improve the decoupling accuracy and signal stability of the corresponding frequency components of the first spiral coil and the second spiral coil in the composite signal in their respective receiving channels.

[0047] In this embodiment, the externally powered receiving glasses or eye mask are responsible for sensing, receiving, and outputting the composite radio frequency signal modulated by the dual-parameter sensing structure within the wearable contact lens. To enable reliable reception and decoupling analysis of two independent sensor signals, the external structure incorporates two spatially offset flexible coil arrays. These two flexible coil arrays utilize a planar structure and are adhered to the inner side of the glasses or eye mask frame, corresponding to the spatial coupling region where the first and second spiral coils within the wearable contact lens reside. Their placement is pre-calibrated and fine-tuned to ensure that, when worn, the two spiral coils form an effective magnetic field coupling channel while maintaining the independence of their respective receiving paths.

[0048] To reduce the magnetic flux interference generated during operation between two spatially offset flexible coil arrays, a high-permeability spacer is introduced into the system. This spacer is installed between the two coil arrays and inserted along their normal direction. The spacer is preferably made of a high-performance nanocrystalline alloy material with a magnetic permeability of not less than 2000, such as Fe-based nanocrystalline ribbon, or a flexible soft magnetic silicon steel composite sheet, to achieve good magnetic flux guidance and shielding capabilities under high-frequency operating conditions. Its function is to establish a low-magnetic resistance leakage path between the two independent channels, so that the modulated signals of the first spiral coil and the second spiral coil will not cause crosstalk due to spatial field overlap when received, thereby improving the signal-to-noise separation required for frequency domain decoupling.

[0049] Each flexible coil array arranged in a spatially staggered manner is connected to its own independent resonant receiving circuit through a wire, and a set of adjustable matching capacitors is configured in the circuit. The adjustable matching capacitor uses a variable capacitor device with voltage-controlled characteristics, such as a GaAs or MOS type variable capacitor chip, and its capacitance value can be adjusted with high precision and continuously under the action of an external control voltage. In order to adapt to different users' wearing angles, facial geometry differences and frame assembly errors, the mobile computing terminal dynamically calculates the required capacitance adjustment range based on the amplitude, phase or mismatch indicators received in each frequency channel, and sends a control signal in real time to accurately adjust the matching capacitance parameters. This adjustment process not only optimizes the impedance matching relationship of each receiving channel, but also provides electrical parameter compensation capabilities during the spatial coupling deviation period, so that the receiving efficiency and frequency filtering characteristics are maintained in a high-sensitivity state.

[0050] In addition, to further enhance the positional alignment capabilities of the two spatially offset flexible coil arrays in dynamic usage environments, the system also features a flexible, controllable fine-tuning component designed to achieve active adjustment of spatial coupling deviations at the structural level. This component is composed of a micro-actuating arm made of shape memory alloy (SMA) material. The SMA can achieve a controllable bending response by inducing a lattice phase transition through electric current heating, and forms a reversible deformation structure in conjunction with an elastic connecting piece. When the glasses slip or the head rotates while being worn, the mobile computing terminal drives the SMA actuator to fine-tune the angle of the corresponding coil array based on the abnormal trend of the received signal, achieving normal rotation or lateral micro-offset within a range of ±5°, ensuring that the two receiving coils can continuously align with the spiral coil in the wearable contact lens.

[0051] In summary, the receiving coil, high-permeability isolator, voltage-controlled matching capacitor, and fine-tuning mechanism form a three-dimensional, coordinated dynamic receiving system. It not only spatially isolates the magnetic field of the dual-frequency signal channels, but also establishes a linkage mechanism for wearing error compensation, electrical matching adjustment, and automatic structural posture recovery. This ensures that the modulated signals of the first and second spiral coils maintain sufficient frequency resolution and time-domain stability after composite reception, laying a high-quality data foundation for subsequent multi-frequency phase shift decoupling and self-calibration modeling. This structure is suitable for monitoring ocular surface pressure during long-term wear and multi-activity scenarios, and is particularly well-suited for continuous signal acquisition with eyes closed.

[0052] The mobile computing terminal 103 is used to receive the processed composite signal, execute a multi-frequency phase shift decoupling algorithm to extract the corresponding frequency components, and separate the eyelid pressure signal and the intraocular pressure signal through sparse matrix decomposition; based on the continuous time series of the separated signals, a timing self-calibration algorithm is executed in combination with a preset corneal biomechanical model to correct the measurement errors caused by the closed eye state or interference, and generate a calibrated eyelid pressure curve and intraocular pressure curve.

[0053] In the fully wireless ocular surface pressure monitoring system described herein, the mobile computing terminal 103 undertakes the key tasks of back-end data processing, signal separation, intelligent calibration, and result visualization. This terminal is typically a smartphone, tablet, or portable processor with sufficient computing power. Its hardware must include at least one multi-core low-power processor, a communication interface (such as a BLE 5.2 Bluetooth module), memory resources, and a local storage module to support the real-time execution of the signal decoupling and dynamic modeling algorithms.

[0054] The first function of the mobile computing terminal 103 is to receive a composite signal transmitted by the processing circuitry of the externally powered receiving glasses or eye mask 102. This composite signal is the sum of two RF signals sensed by the receiving coil, modulated at different center frequencies. After filtering, amplification, and analog-to-digital conversion by the pre-stage circuitry, it is transmitted to the terminal via Bluetooth or near-field communication. The terminal's built-in signal receiving module performs frame synchronization and preliminary analysis on this composite signal, extracting the raw frequency domain data stream to provide effective input for the subsequent frequency analysis module.

[0055] After receiving the composite signal, the terminal inputs it into the multi-frequency phase shift decoupling algorithm module. This algorithm, based on the differences in phase response and time delay between different frequency components, uses a fast Fourier transform (FFT) or sliding window Fourier analysis to extract the frequency components corresponding to the first and second spiral coils. This stage requires not only clearly detecting the amplitude variation trends of each center frequency but also dynamically modeling amplitude perturbations and phase drift to enhance the algorithm's robustness to environmental interference and wearer errors. The basic processing of the decoupling algorithm is as follows: First, a fast Fourier transform (FFT) is performed on the input composite signal to obtain its frequency domain amplitude and phase information. Based on the two spiral coil tuning center frequencies preset in the system hardware design, for example, 13.56 MHz and 15.20 MHz, the terminal locates the energy peaks near these two frequencies in the spectrum. To reduce errors caused by crosstalk between adjacent frequency bands, the terminal sets a bandwidth filtering window for each frequency and introduces a phase recovery mechanism to reconstruct the time domain phase consistency of the original signal through phase shift compensation. Finally, after filtering and phase matching processing, two independent frequency components representing the changes in eyelid pressure and intraocular pressure are obtained respectively.

[0056] After extracting the frequency components, the system further decouples the two signal channels using a sparse matrix decomposition algorithm. Based on the principles of sparse coding, this algorithm constructs a set of eigenvectors in two orthogonal or non-overlapping support domains. Combined with L1 regularization constraints, it iteratively solves the composite signal into two sparse components. The decoupled signals from each channel are then mapped to the original sensor output values, corresponding to the time-series electrical changes in eyelid pressure and intraocular pressure, respectively.

[0057] To eliminate measurement errors caused by factors such as posture, eye closure, eyelid slippage, or changes in corneal humidity, the mobile computing terminal also utilizes a built-in self-calibration algorithm. Based on a corneal biomechanical response model, this module establishes a nonlinear mapping relationship between sensor response and actual physiological pressure. It then performs sliding averaging, trend fitting, and outlier identification based on data trends within a time window. The model inputs include not only the signal itself but also dynamic characteristic indicators such as wear time and output stability during the previous phase. Parameter adjustment allows for automatic compensation of signal drift and amplitude deviation.

[0058] To dynamically calibrate signals collected from wearable contact lenses, the mobile computing terminal's built-in self-calibration algorithm incorporates a response model based on corneal biomechanics to simulate the nonlinear response of corneal deformation to the combined effects of eyelid pressure and intraocular pressure. After signal separation, this model establishes a mapping between the sensor's raw signal and the actual corneal deformation state, compensating for local measurement errors caused by posture changes, eye closure, or lens decentration.

[0059] The basic data for the corneal biomechanical response model comes from measurements of the central displacement of the human cornea under controlled transient pressure, as documented in public literature and experimental databases. Current clinical studies and numerous publications have shown that when a certain range of eyelid pressure or intraocular pressure disturbance is applied, the central cornea exhibits a concave response of a certain magnitude. This response exhibits a nonlinear trend with increasing pressure, exhibiting a certain degree of hysteresis and recovery delay. The system utilizes the relationship curve between the maximum central corneal concave displacement and applied pressure from known experimental data, and employs a piecewise fitting method to construct a standard template library for initial fitting.

[0060] In terms of model structure, the system represents the corneal pressure response relationship as a family of nonlinear functions, specifically quadratic functions with pressure as the independent variable and corneal displacement per unit time as the dependent variable. For example, the system uses the following approximate expression to model the corneal deformation trend by default: It is the deformation displacement of the center point of the cornea, that is, the response at the physiological level; the displacement of the center of the cornea and instantaneous pressure Satisfy between:

[0061] ;

[0062] in, It represents the corneal equivalent stiffness coefficient, reflecting the linear anti-deformation ability of the cornea to the basic intraocular pressure. The nonlinear recovery factor of the cornea is used to describe the accelerated growth and delayed recovery of the cornea's deformation under rapid loading, such as eye closure. In practice, these two parameters are not fixed constants but rather dynamically adjustable fitting parameters in the system.

[0063] During initial system startup, the mobile computing terminal first loads a set of standard model parameters from a template library that most closely matches the user's physiological characteristics, such as age and corneal thickness, as initial values. Subsequently, after receiving and decoupling raw eyelid pressure and intraocular pressure signals, the system automatically identifies the time period of maximum corneal stress during the eye closure cycle, extracts characteristics such as the signal change rate, pressure duration, and intraocular pressure change amplitude during this period, and uses a minimum residual fitting algorithm to locally adjust the aforementioned nonlinear function. By performing a rolling fit on the parameters of this function, the system can update the corneal model in real time based on individual differences, making it more closely aligned with the user's actual corneal response behavior in a dynamic state.

[0064] In order to avoid the amplification deviation of the calibration results caused by the model error, the system also adopts the window sliding weighting strategy to fuse the fitting results in multiple time periods. At the end of each time window, the system records the fitting residuals, sensor signal stability and the rate of change of the estimated parameters in this window, constructs a set of weighted coefficients, and and The two core parameters are updated smoothly to prevent the model from experiencing drastic jumps during short-term abnormal events. Furthermore, after model fitting, the system feeds the calibration results back to the signal processing module to normalize the amplitude and correct relative drift of the original pressure curve, thereby outputting eyelid and intraocular pressure curves that are more consistent with physiological reality.

[0065] After calibration, the terminal generates standardized eyelid pressure curves and intraocular pressure curves, and displays them visually on the local interface. Curve display formats can include time series graphs, pressure fluctuation graphs, daily average trend graphs, etc. If the system design has networking capabilities, the results can be uploaded to a remote database or cloud platform for remote monitoring by doctors or access to electronic medical record systems. The terminal can also set pressure alarm thresholds. Once the intraocular pressure fluctuation exceeds the safe range, an alarm prompt will automatically pop up, or a notification will be sent to the monitoring device.

[0066] From a system implementation perspective, mobile computing terminals can deploy the entire algorithm process through a mobile app, performing regular backend data collection, analysis, and storage, keeping the frontend interface lightweight and user-friendly. App development can be based on Android or iOS platforms, invoking the system-level Bluetooth stack and integrating FFT, sparse decomposition, and model optimization code using C++ or Python modules, achieving efficient and portable data processing.

[0067] In summary, mobile computing terminal 103 is more than just a simple data receiving and display device; it also serves as an intelligent analysis core that implements frequency decoupling, time correction, signal reconstruction, and trend assessment. Its functional completeness and processing accuracy directly determine the system's overall medical application value and clinical usability, making it a key component supporting a fully wireless, dual-channel, and highly reliable ocular surface pressure monitoring system.

[0068] The following is a reference implementation code of the mobile computing terminal 103:

[0069] import numpy as np

[0070] from scipy.fft import fft, ifft

[0071] from sklearn.linear_model import Lasso

[0072] import matplotlib.pyplot as plt

[0073] # Simulate the received composite signal (demodulated and filtered digital signal)

[0074] def receive_composite_signal():

[0075] # Assuming the signal sampling rate is 1kHz and the length is 5000 points, simulate the superposition of two frequency components

[0076] t = np.linspace(0, 5, 5000)

[0077] signal1 = 1.0 np.sin(2 np.pi 13.56 t) # first frequency component

[0078] signal2 = 0.8 np.sin(2 np.pi 15.20 t + np.pi / 3) #second frequency component

[0079] noise = 0.1 np.random.randn(len(t)) # add a small amount of noise

[0080] composite = signal1 + signal2 + noise

[0081] return composite, t

[0082] # Perform multi-frequency phase shift decoupling to extract two frequency components

[0083] def frequency_domain_separation(composite_signal, sample_rate=1000):

[0084] # Use FFT to extract frequency components

[0085] spectrum = fft(composite_signal)

[0086] freq_axis = np.fft.fftfreq(len(composite_signal), d=1 / sample_rate)

[0087] # Set the frequency window and extract two known center frequencies (simplify to Hz in MHz as an example)

[0088] band1 = (13.5, 13.7)

[0089] band2 = (15.1, 15.3)

[0090] # Construct two empty spectrum channels

[0091] channel1 = np.zeros_like(spectrum, dtype=complex)

[0092] channel2 = np.zeros_like(spectrum, dtype=complex)

[0093] for i, f in enumerate(freq_axis):

[0094] if band1[0]<= abs(f)<= band1[1]:

[0095] channel1[i] = spectrum[i]

[0096] elif band2[0]<= abs(f)<= band2[1]:

[0097] channel2[i] = spectrum[i]

[0098] # Perform IFFT on the two frequency components to obtain the time domain signal

[0099] signal1_time = np.real(ifft(channel1))

[0100] signal2_time = np.real(ifft(channel2))

[0101] return signal1_time, signal2_time

[0102] # Sparse matrix decomposition for signal decoupling (eyelid pressure & intraocular pressure)

[0103] def sparse_decomposition(signal, dictionary_type='gaussian'):

[0104] n = len(signal)

[0105] # Build a sparse dictionary: based on Gaussian or piecewise linear

[0106] if dictionary_type =='gaussian':

[0107] dictionary = np.array([

[0108] np.exp(-((np.arange(n) - i) twenty two 20 2)) for i in range(0, n, 100) ])

[0110] else:

[0111] dictionary = np.array([

[0112] np.maximum(0, 1 - abs(np.arange(n) - i) / 50) for i inrange(0, n, 100) ])

[0114] # Using L1 regularization (Lasso) to implement sparse coding

[0115] model = Lasso(alpha=0.01, max_iter=1000)

[0116] model.fit(dictionary.T, signal)

[0117] coeffs = model.coef_

[0118] reconstructed = dictionary.T @coeffs

[0119] return reconstructed

[0120] #Corneal biomechanics self-calibration model

[0121] def biomechanical_self_calibration(pressure_curve, calibration_type='eyelid'):

[0122] # Simulate nonlinear correction: perform amplitude compression or trend correction based on the current pressure value

[0123] calibrated_curve = []

[0124] for p in pressure_curve:

[0125] if calibration_type =='eyelid':

[0126] if p>1.5:

[0127] p = p / (1 + 0.2 p) # Simulate corneal hysteresis compression

[0128] else:

[0129] if p>15:

[0130] p -= 0.5 # simulates the intraocular pressure shift caused by corneal deformation

[0131] calibrated_curve.append(p)

[0132] return np.array(calibrated_curve)

[0133] # Generate a graph and output the results

[0134] def visualize_and_export(time_axis, eyelid_curve, iop_curve):

[0135] plt.figure(figsize=(12, 5))

[0136] plt.plot(time_axis, eyelid_curve, label='Calibrated EyelidPressure (kPa)')

[0137] plt.plot(time_axis, iop_curve, label='Calibrated IntraocularPressure (mmHg)')

[0138] plt.xlabel('Time (s)')

[0139] plt.ylabel('Pressure')

[0140] plt.title('Calibrated Eye Surface Pressure Curves')

[0141] plt.legend()

[0142] plt.grid(True)

[0143] plt.tight_layout()

[0144] plt.savefig("calibrated_eye_pressure_curve.png")

[0145] plt.show()

[0146] # Main function: combine all steps and execute a complete calculation process

[0147] def main():

[0148] # Step 1: Receive composite signal

[0149] composite_signal, time_axis = receive_composite_signal()

[0150] # Step 2: Multi-frequency phase-shift decoupling

[0151] eyelid_component, iop_component = frequency_domain_separation(composite_signal)

[0152] # Step 3: Sparse reconstruction

[0153] eyelid_sparse = sparse_decomposition(eyelid_component, dictionary_type='gaussian')

[0154] iop_sparse = sparse_decomposition(iop_component, dictionary_type='linear')

[0155] # Step 4: Self-calibration

[0156] eyelid_calibrated = biomechanical_self_calibration(eyelid_sparse,calibration_type='eyelid')

[0157] iop_calibrated = biomechanical_self_calibration(iop_sparse,calibration_type='iop')

[0158] # Step 5: Generate images and export results

[0159] visualize_and_export(time_axis, eyelid_calibrated, iop_calibrated)

[0160] Furthermore, before performing the sparse matrix decomposition, the mobile computing terminal performs dynamic time-domain modeling on the signal morphologies of the eyelid pressure signal and the intraocular pressure signal based on the trend of changes in the spectral energy distribution of the extracted multi-frequency signal within a sliding time window. During the modeling process, the eyelid pressure signal is regarded as a short-duration, bursty, high-amplitude impact signal, and the intraocular pressure signal is regarded as a low-frequency, slowly varying, steady trend signal. The modeling parameters are initialized in combination with historical monitoring data of users wearing contact lenses and a standard ocular surface pressure physiological model to improve the individual adaptability of the model.

[0161] After completing the time-domain modeling, the mobile computing terminal constructs a Gaussian function dictionary for reconstructing the eyelid pressure signal and a piecewise linear basis dictionary for reconstructing the intraocular pressure signal, respectively, and performs the sparse matrix decomposition process through L1 norm minimization and joint regularization to perform unsupervised sparse deconstruction and structural reconstruction of the eyelid pressure signal and the intraocular pressure signal in a context of partial overlap of frequency components and significant differences in amplitude dynamic range. The Gaussian function dictionary is used to characterize local impact disturbances generated during eyelid closure and blinking cycles, and the piecewise linear basis dictionary is used to characterize the slow variation of intraocular pressure during circadian rhythms and resting states.

[0162] After completing the sparse matrix decomposition, the mobile computing terminal performs an automatic feedback residual evaluation mechanism based on the time-frequency error index between the reconstructed signal and the original separated signal. When the reconstructed residual exceeds a threshold or the sparsity index does not meet a preset condition, a local update process of the sparse dictionary is dynamically triggered. In this process, a joint feature offset vector between the user's historical data and the standard template is constructed, the feature atomic vector used in the current sliding time window is corrected, and the regularization weight coefficient and decoupling-related parameters are updated;

[0163] The mobile computing terminal is further used to send the updated eyelid pressure signal and the intraocular pressure signal to the input end of the timing self-calibration algorithm respectively, and perform a calibration operation in combination with a preset corneal biomechanical model. The calibration result is used to correct the measurement error caused by the closed eye state, eyeglass posture deviation or soft lens eccentricity in real time, and generate dynamically updated eyelid pressure curves and intraocular pressure curves to achieve continuous and stable ocular surface pressure monitoring and subsequent abnormality detection.

[0164] In order to achieve accurate separation and stable output of eyelid pressure signals and intraocular pressure signals in a composite state in wearable contact lenses, this embodiment provides a complete data processing flow. Relying on the high-frequency signal processing capabilities and model-driven algorithm structure of the mobile computing terminal, it gradually completes signal modeling, sparse decomposition, dictionary update and self-calibration deduction, and finally obtains a physically meaningful ocular surface pressure change curve.

[0165] First, the mobile computing terminal receives a composite signal output by the receiving coil in the externally powered receiving glasses or eye mask, shaped by a processing circuit. This signal is a frequency-domain composite waveform formed by spatially coupling two RF signals modulated by the first and second spiral coils, respectively. It carries the original coded information from the capacitive eyelid pressure sensor and the piezoresistive intraocular pressure sensor in response to pressure changes. Through front-end electromagnetic demodulation and dual-frequency channel decoupling, the mobile computing terminal extracts two primary frequency channels, corresponding to the initial separation of the eyelid pressure channel and the intraocular pressure channel. However, at this point, the two channel signals still exhibit some spectral overlap and amplitude mixing.

[0166] To further enhance signal decoupling quality and accommodate individual differences, the mobile computing terminal first performs time-domain modeling of the eyelid pressure signal and the intraocular pressure signal within each sliding time window, based on the energy distribution and dynamic change curves of the multi-frequency signals extracted during the current time period. During the modeling process, the eyelid pressure signal is defined as a non-stationary, high-amplitude, impact signal with significant temporal mutation characteristics. Its manifestation is often associated with blinking, eye closure, or sudden changes in eyelid tension, with a short duration and a sharp peak. The intraocular pressure signal, on the other hand, is defined as a relatively stable, low-frequency, varying signal whose changing trend is often associated with circadian rhythms, body position adjustments, or slow fluctuations in intraocular pressure, exhibiting strong trend and predictability. During the modeling process, the mobile computing terminal utilizes historical pressure curve data from the contact lens wearer and a pre-set standard ocular surface pressure physiological model to initialize the modeling parameters, ensuring that the dictionary construction and deconstruction path are more individually adaptable and accurate in real time.

[0167] After the modeling is completed, the system enters the sparse matrix decomposition stage. Based on the above signal model, the mobile computing terminal constructs a set of Gaussian function dictionaries for reconstructing the eyelid pressure signal and a set of piecewise linear basis dictionaries for reconstructing the intraocular pressure signal. In the process of constructing the dictionary, the system adjusts the parameters according to the spectrum envelope width, waveform duration and change rate, so that the Gaussian function family is suitable for short-period pulse detection, while the linear basis group is suitable for fitting the stable growth or slow decline of the trend term. After the dictionary is constructed, the mobile computing terminal performs sparse coding on the two channel signals respectively through the L1 norm minimization joint regularization method, so that it can restore the signal prototype with the simplest basis function combination in different dimensions. This process not only realizes the structural separation of sudden disturbances and smooth trend signals, but also effectively eliminates the mixed components caused by amplitude overlap or frequency domain crosstalk, so that the two sets of decoupled signals have good mutual difference and time consistency.

[0168] Next, to ensure robustness under long-term wear conditions, the system also sets up an automatic feedback mechanism based on signal reconstruction residuals. When the reconstruction error within a sliding window (i.e., the mean squared difference or frequency offset between the original signal and the deconstructed and reconstructed signal) exceeds the system-set threshold, or when the sparsity of the decoded signal does not meet the set constraints, the mobile computing terminal will automatically trigger the local dictionary update process. In this process, the system extracts the joint feature offset vector between the ocular surface pressure signal morphology of the wearable contact lens user in the current time period and the historical time period, and dynamically matches this vector with a preset template to select an atomic basis that better fits the current individual state and reassign the regularization weight factor. This process not only improves the stability of deconstruction, but also allows the dictionary structure to adaptively fine-tune individual changes, forming a semi-supervised self-learning mechanism.

[0169] After completing the sparse deconstruction and generating the separated signals, the mobile computing terminal feeds the updated eyelid pressure signal and the updated intraocular pressure signal into the back-end timing self-calibration algorithm module. This module pre-sets a multi-factor correction model based on corneal biomechanical characteristics, including the feedback relationship between eyelid pressure and corneal pressure change, the impact path of soft lens decentration on signal amplitude jitter, and the coupling perturbation coefficient of wearing posture changes (such as glasses slipping or head posture deviation) on the signal. After receiving the two sets of raw pressure signals, the system derives calibration coefficients based on the biomechanical inversion model and then performs timing correction on the original signal sequence to ensure that the signal output more accurately reflects the wearer's actual ocular surface physiological state.

[0170] Finally, the system will generate continuous eyelid pressure curves and intraocular pressure curves for the two calibrated signals for visual viewing by the user, and can also be used for subsequent abnormality detection, eye closure status monitoring or long-term trend analysis. The entire data processing process starts with the reception of the composite signal, and through frequency separation, modeling initialization, sparse deconstruction, dictionary feedback update and time series self-calibration deduction, it gradually extracts a high-reliability pressure curve with physical stability and physiological relevance, which constitutes the key support path for the system of the present invention to continuously and accurately monitor dual ocular surface parameters under high dynamic interference conditions. This process can not only adapt to the differences in intraocular pressure structure between different individuals, but also has the ability to adjust to dynamic wearing changes. It is widely applicable to multiple application scenarios such as daily monitoring, postoperative tracking, fatigue detection, etc.

[0171] Furthermore, the time-series self-calibration algorithm performs parameter fitting based on a preset corneal biomechanical response model. The corneal biomechanical response model is based on the nonlinear pressure-displacement curve of the central area of ​​the cornea under different transient pressures. The model fits and dynamically updates two response parameters: the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor. The corneal equivalent stiffness coefficient is used to describe the average anti-deformation ability of the cornea under the intraocular pressure background, and the corneal nonlinear recovery factor is used to reflect the deformation recovery characteristics of the cornea during the rapid change stage of the eyelid pressure signal.

[0172] The timing self-calibration algorithm identifies the time interval between the rising edge and the falling edge of the eyelid pressure signal in the eye-closed state, and is used to calculate the eyelid pressure application cycle within the current time window. The algorithm combines the local change trend of the intraocular pressure signal to determine the maximum response interval of the cornea under the combined action of the eyelid pressure signal and the intraocular pressure signal. Based on this interval, the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor are dynamically adjusted to adapt to the composite force state under different eyelid dynamic and intraocular pressure trend conditions during the wearing of the wearable contact lens.

[0173] After each sliding time window ends, the time series self-calibration algorithm combines the eyelid pressure signal change range and the intraocular pressure signal change rate within multiple time windows to construct a window weighted fitting result, performs a rolling update on the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor, and generates a fitting calibration factor for current window correction;

[0174] The mobile computing terminal is used to apply the fitting calibration factor output by the timing self-calibration algorithm to the timing correction process of the eyelid pressure signal and the intraocular pressure signal, and output the calibrated eyelid pressure curve and the intraocular pressure curve to improve the pressure monitoring stability and signal consistency of the wearable contact lens under conditions of closed eyes, eccentric wearing or posture changes.

[0175] In this invention, the key to the mobile computing terminal's timing correction of eyelid pressure and intraocular pressure signals lies in correctly acquiring and applying fitting calibration factors. These fitting calibration factors are derived from a dynamic fitting process of the corneal biomechanical response model. Their core function is to convert the electrical signal output by the sensor into a physical quantity reflecting the actual corneal deformation state, thereby improving measurement accuracy under non-ideal conditions such as closed eyes, posture changes, and wearing eccentricity.

[0176] Specifically, the system first establishes a functional mapping relationship between pressure and corneal center displacement through a preset corneal biomechanical response model. This model is based on the nonlinear deformation data of the central corneal area under different pressure conditions and is described by a quadratic function by default, that is, the corneal displacement d is related to the instantaneous applied pressure. The relationship can be expressed as:

[0177] ;

[0178] in, It represents the corneal equivalent stiffness coefficient, which is used to characterize the cornea's linear compressive deformation resistance under normal intraocular pressure. The nonlinear recovery factor of the cornea is used to characterize the nonlinear deformation growth and hysteretic recovery characteristics of the cornea during rapid loading, such as eye closure. The two parameters in this function are adjustable and are dynamically obtained by fitting data within a sliding time window during actual system operation.

[0179] During the signal correction process, the timing self-calibration algorithm first identifies the rising and falling edges of the eyelid pressure signal within the current window, thereby inferring the position range of the eye closure pressure cycle. The system will detect the fluctuation amplitude and slope changes of the two signals within this cycle, and determine the time period when the corneal response is most intense based on their coupling trend. Then, this data is used to and Least squares fitting is performed to ensure that the quadratic function optimally approximates the relationship between the true sensor response and corneal deformation. This fitting process can be achieved using standard linear regression methods, which simply construct linear and square terms whose inputs are the raw pressure data and whose outputs are the observed corneal equivalent deformation values ​​(e.g., calibrated values ​​from a simulation model or historical template library).

[0180] After the fitting is completed, the system will integrate the fitting results of the current window with the historical results of the previous windows to build a window weighted fitting model. The weighting coefficient is automatically adjusted according to the residual size, signal stability and sensor environment parameters of each window to prevent sudden changes in fitting parameters due to a single abnormal window. In this way, the system outputs a pair of stable and physiologically meaningful fitting calibration factors (i.e., the current window's and value).

[0181] Then, the mobile computing terminal applies these two fitting calibration factors to the original eyelid pressure signal and intraocular pressure signal in the current time window. For each sampling point, the system reads the original output value of the sensor , and then substitute this value into the function , calculate the corneal equivalent displacement value at that moment , which is the pressure value after physical calibration at that moment. The signals in the entire time window are nonlinearly mapped according to this method, and finally a continuous pressure curve that conforms to the law of corneal mechanical response is output. The eyelid pressure signal and the intraocular pressure signal are processed separately in this correction process, but the same and To ensure that their relative relationship is consistent in timing and avoid drift deviation caused by decoupling processing.

[0182] After each window correction is completed, the system will also compare the change range, trend offset and response delay between the current corrected curve and the uncalibrated curve to further evaluate whether the dictionary matching parameters, regularization factors or filter settings need to be readjusted to maintain the dynamic adaptability of the entire signal chain.

[0183] Through the above operations, the present invention not only achieves nonlinear physical correction of the sensor's raw signal but also ensures the continuity and reliability of the fitting parameters under different usage environments through model fitting and a sliding weighting mechanism. The resulting eyelid pressure and intraocular pressure curves are highly stable, continuous, and physiologically reliable, providing a solid data foundation for subsequent dynamic trend modeling, anomaly detection, and remote diagnosis.

[0184] The following is the reference implementation code corresponding to the above description of this embodiment:

[0185] import numpy as np

[0186] import matplotlib.pyplot as plt

[0187] from sklearn.linear_model import Lasso

[0188] #Simulation signal generation: including eyelid pulse + slowly changing intraocular pressure + noise

[0189] def simulate_composite_signal():

[0190] t = np.linspace(0, 5, 5000)

[0191] eyelid_signal = 1.2 np.exp(-((t - 2.5) 2) / 0.01) # Eye-closing burst pulse

[0192] iop_signal = 0.2 np.sin(2 np.pi 0.2 t) + 1.0 # Slowly changing intraocular pressure

[0193] noise = 0.05 np.random.randn(len(t)) # system noise

[0194] return eyelid_signal + iop_signal + noise, eyelid_signal, iop_signal, t

[0195] # Build sparse dictionary (Gaussian / piecewise linear)

[0196] def build_dictionary(length, mode='gaussian'):

[0197] dictionary = []

[0198] centers = np.linspace(0, length, length / / 50)

[0199] for c in centers:

[0200] if mode == 'gaussian':

[0201] atom = np.exp(-((np.arange(length) - c) 2) / (2 (length / / 20) 2))

[0202] else:

[0203] atom = np.maximum(0, 1 - abs(np.arange(length) - c) / (length / / 10))

[0204] dictionary.append(atom)

[0205] return np.array(dictionary).T

[0206] # Sparse decomposition + residual evaluation + regularization factor update mechanism

[0207] def adaptive_sparse_decomposition(signal, dictionary, alpha_init=0.05, max_iter=3):

[0208] alpha = alpha_init

[0209] for _ in range(max_iter):

[0210] model = Lasso(alpha=alpha, max_iter=1000)

[0211] model.fit(dictionary, signal)

[0212] reconstruction = dictionary @ model.coef_

[0213] residual = np.linalg.norm(signal - reconstruction)

[0214] sparsity = np.count_nonzero(model.coef_)

[0215] # Dynamically adjust the regularization coefficient based on the residual

[0216] if residual>0.5:

[0217] alpha = 1.5

[0218] elif sparsity<5:

[0219] alpha = 0.7

[0220] else:

[0221] break

[0222] return reconstruction, residual, sparsity, model.coef_

[0223] # Eye closure detection + identification of the maximum corneal stress area

[0224] def detect_eyelid_closure_and_response_zone(eyelid_signal, iop_signal, t):

[0225] threshold = 0.5

[0226] mask = eyelid_signal>threshold

[0227] start_idx = np.argmax(mask)

[0228] end_idx = len(mask) - np.argmax(mask[::-1])

[0229] if end_idx<= start_idx or not np.any(mask):

[0230] return 0, len(eyelid_signal) # No eye closure detected, return the full interval

[0231] return start_idx, end_idx

[0232] # Simulated corneal parameter fitting + adaptive correction

[0233] def biomechanical_calibration(eyelid_signal, iop_signal, start, end):

[0234] k = 0.85 # Equivalent stiffness

[0235] alpha = 0.25 # non-linear recovery factor

[0236] eyelid_corr = np.copy(eyelid_signal)

[0237] iop_corr = np.copy(iop_signal)

[0238] for i in range(start, end):

[0239] eyelid_corr[i] = eyelid_signal[i] / (1 + alpha eyelid_signal[i] / k)

[0240] iop_corr[i] = iop_signal[i]- 0.05 (eyelid_signal[i] / k)

[0241] return eyelid_corr, iop_corr

[0242] # Abnormal detection: alarm if high intraocular pressure > 2.0 mmHg

[0243] def detect_abnormal_events(iop_signal, threshold=2.0):

[0244] anomalies = []

[0245] for i, val in enumerate(iop_signal):

[0246] if val>threshold:

[0247] anomalies.append(i)

[0248] return anomalies

[0249] # Visualize the results

[0250] def visualize_results(t, eyelid, iop, anomalies):

[0251] plt.figure(figsize=(12, 5))

[0252] plt.plot(t, eyelid, label='Eyelid pressure after calibration')

[0253] plt.plot(t, iop, label='Calibrated intraocular pressure')

[0254] for a in anomalies:

[0255] plt.axvline(t[a], color='red', linestyle='--', alpha=0.6)

[0256] plt.title("Ocular surface pressure monitoring results")

[0257] plt.xlabel("Time(s)")

[0258] plt.ylabel("pressure")

[0259] plt.legend()

[0260] plt.grid()

[0261] plt.tight_layout()

[0262] plt.show()

[0263] # Main process

[0264] def full_pipeline():

[0265] composite, raw_eyelid, raw_iop, t = simulate_composite_signal()

[0266] # Dictionary construction

[0267] dict_len = len(composite)

[0268] dict_eyelid = build_dictionary(dict_len,'gaussian')

[0269] dict_iop = build_dictionary(dict_len,'linear')

[0270] # Sparse decomposition + regularization adaptation + reconstruction

[0271] rec_eyelid, res_e, sp_e, _ = adaptive_sparse_decomposition(raw_eyelid, dict_eyelid)

[0272] rec_iop, res_i, sp_i, _ = adaptive_sparse_decomposition(raw_iop,dict_iop)

[0273] # Eye closure detection + maximum response interval identification

[0274] start_idx, end_idx = detect_eyelid_closure_and_response_zone(rec_eyelid, rec_iop, t)

[0275] # Self-calibration based on corneal response model

[0276] eyelid_calibrated, iop_calibrated = biomechanical_calibration(rec_eyelid, rec_iop, start_idx, end_idx)

[0277] # Anomaly detection (e.g. intraocular pressure spikes)

[0278] anomaly_indices = detect_abnormal_events(iop_calibrated)

[0279] # Display results

[0280] visualize_results(t, eyelid_calibrated, iop_calibrated, anomaly_indices)

[0281] full_pipeline()

[0282] Furthermore, before performing the sparse matrix decomposition, the mobile computing terminal constructs a frequency decoupled path cost function , for dynamically selecting an optimal channel mapping path between the output frequency responses of the first spiral coil and the second spiral coil, wherein the path cost function is defined by the following formula:

[0283] ;

[0284] in, and are the dominant frequency components decoupled by the first and second spiral coils at the current moment respectively; and are the theoretical center frequencies of the first and second channels, respectively, which come from the initial resonant frequency set by the system; Indicates the current and The corresponding frequency domain overlapping power spectral density in the composite signal; 、 、 It is an adjustable weight coefficient, which is dynamically set according to the current signal decoupling residual, channel crosstalk level and wearing status.

[0285] The path cost function The frequency decoupling optimization module is input as the objective function, and the system uses gradient descent to find a pair of and , making Min. and It is used to remap the original spectrum to a dynamically adapted center frequency, thereby improving the decoupling robustness and signal reconstruction accuracy when attitude disturbances, coupling anomalies, or coil deformation occur.

[0286] The wearable smart contact lens system of the present invention is based on a dual-channel resonant structure. The outer and inner soft lens layers are integrated with a first spiral coil and a second spiral coil, respectively. These coils are coupled to a capacitive eyelid pressure sensor and a piezoresistive intraocular pressure sensor, respectively, to modulate the two independent physiological pressure signals into radio frequency signals with different center frequencies. A mobile computing terminal receives these mixed signals via an external receiving coil and uses a frequency decoupling algorithm to split them into two independent signal streams.

[0287] In order to improve the decoupling stability and robustness in the case of soft lens wearing deviation, glasses micro deformation or posture disturbance, the system will first perform a frequency component based path evaluation process before performing sparse matrix decomposition. The core of this process is to construct a frequency decoupling path cost function , which is used to find the optimal decoupling path pair from the two dominant frequency components extracted at the current moment, thereby reducing signal crosstalk or separation failure caused by frequency drift.

[0288] The mathematical expression of the cost function is as follows:

[0289] ;

[0290] The expression consists of three weighted terms, representing the penalty for frequency drift, the suppression of the degree of frequency coupling overlap, and the control of the accuracy of the signal decoupling results by their weighted combination.

[0291] first, and They represent the two dominant frequency peaks calculated by the system's current intra-frame detection algorithm. The system uses Fast Fourier Transform (FFT) or Sliding Window Short-Time Fourier Transform (STFT) to process the received signal in the current period, extract the maximum peak frequency within the preset frequency band, and assign them to and These two values ​​reflect the dominant energy frequency of the RF signal currently output by the first spiral coil and the second spiral coil, in MHz, and are generally in the range of [13.0, 14.0] MHz.

[0292] and They represent the theoretical resonant frequencies set during system initialization, i.e., the center frequencies assigned to the first and second spiral coils during the design phase. For example, can be 13.56 MHz, The frequency difference (280 kHz) is sufficient to ensure theoretical frequency domain separability. These parameters are obtained by tuning the sensor's physical parameters and measuring them with a network analyzer. They are static configuration parameters.

[0293] Item 1 Indicates the square of the amplitude of the current frequency deviating from its theoretical value, which is used to penalize frequency drift. The deviation term for the second channel is defined in the same way as the first. These two terms ensure that the frequency selection is as close to the designed center frequency as possible, thus preventing the system from selecting the wrong frequency at low signal-to-noise ratios.

[0294] Item 3 Indicates frequency and The power spectrum overlap intensity of the corresponding channel in the composite signal spectrum is calculated as follows: and Centered and expanding to both sides For example, a bandwidth of ±50 kHz is used to calculate the integrated overlap area of ​​the two channel spectra within this bandwidth. The larger the overlap area, the more difficult it is to distinguish the two frequency components on the power spectrum, and the larger the decoupling residual will be. The unit is dBHz or normalized power unit.

[0295] There are three weight coefficients, which can be set dynamically in the actual system. The recommended initial values ​​are:

[0296] , a square penalty term used to balance frequency drift;

[0297] , used to enhance the sensitivity adjustment of power spectrum overlap.

[0298] When the power spectrum overlaps seriously or the signal attenuation is obvious, the system can Dynamically increase it to above 10.0 to give priority to frequency pairs with better decoupling effects.

[0299] The entire cost function is used as the objective function input to the frequency decoupling path optimization module. In the local frequency space with the initial value, the minimization operation based on the gradient descent method or the quasi-Newton iterative algorithm is performed. and Find a set of continuous intervals and , making Take the minimum value. The final selected and It will serve as the updated frequency mapping point for subsequent signal reconstruction, matrix sparse decomposition and physiological signal re-extraction.

[0300] Through the dynamic programming mechanism of the above-mentioned path cost function, the present invention can effectively alleviate the decoding distortion problem caused by frequency drift, coil offset or cross-talk during the decoupling process of multi-source RF signals, and ensure that the stability of signal separation and decoupling quality can still be maintained at a high level under actual clinical conditions such as mobile wearing, posture changes, and short-term interference, thereby further improving the credibility of subsequent pressure identification and pathological feature judgment.

[0301] Although the present application is disclosed as above with the preferred embodiments, it is not intended to limit the present application. Any person skilled in the art may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. A wireless ocular surface pressure monitoring system based on a multi-frequency phase shift decoupling self-calibration algorithm, characterized in that: include: A wearable contact lens having a three-layer structure, with a hard lens in the middle for support and shielding. The outer soft lens is embedded with a capacitive eyelid pressure sensor and a first spiral coil, and the inner soft lens is embedded with a piezoresistive intraocular pressure sensor and a second spiral coil. The first spiral coil and the second spiral coil are respectively tuned to different center frequencies and are used to modulate pressure changes into radio frequency signals of corresponding frequencies when coupled with their respective sensors. Externally powered receiving glasses or eye masks, comprising a multi-frequency transmitting coil electromagnetically coupled to the first and second spiral coils, a power supply rectifier circuit, and a receiving coil, wherein the transmitting coil provides energy to the eyelid pressure sensor and the first spiral coil, and to the intraocular pressure sensor and the second spiral coil, respectively; the receiving coil is configured to sense and acquire a composite signal formed by respectively modulating and spatially superimposing the two sensor coil structures, and provide the composite signal to a processing circuit, which transmits the processed composite signal to a mobile computing terminal; a mobile computing terminal, configured to receive the processed composite signal, execute a multi-frequency phase shift decoupling algorithm to extract corresponding frequency components, and separate the eyelid pressure signal from the intraocular pressure signal by sparse matrix decomposition; Based on the continuous time series of the separated signals and combined with the preset corneal biomechanical model, a time series self-calibration algorithm is executed to correct the measurement errors caused by the closed eye state or interference, and generate the calibrated eyelid pressure curve and intraocular pressure curve.

2. The wireless ocular surface pressure monitoring system based on the multi-frequency phase shift decoupling self-calibration algorithm according to claim 1 is characterized in that: The outer soft lens of the wearable contact lens is made of highly translucent silicone hydrogel material; the capacitive eyelid pressure sensor is arranged in a non-central area of ​​the outer soft lens close to the upper eyelid; the first spiral coil is formed by a grid structure composed of silver nanowires, which is transferred to a polyimide flexible film substrate after laser patterning and fixed in a ring-shaped manner in a non-optical axis area at the edge of the outer soft lens.

3. The wireless ocular surface pressure monitoring system based on the multi-frequency phase shift decoupling self-calibration algorithm according to claim 2 is characterized in that: The non-optical axis annular region where the first spiral coil is located is provided with an array of micro-through holes that penetrate the thickness direction of the outer layer of the soft lens. The diameter of each micro-through hole is less than 50 microns. The through holes are filled with an array of compressible hydrophilic elastomer micro-pillars, which are used to form a buffer deformation layer when shear force is generated by eyelid movement, thereby reducing measurement errors caused by lateral friction of the eyelid. The outer soft lens and the middle hard lens are subjected to surface plasma activation treatment and then interface bonded using a two-component ultraviolet copolymer photosensitive adhesive to form a stable multi-layer integrated structure.

4. The wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm according to claim 1 is characterized in that: The receiving coil in the externally powered receiving glasses or eye mask includes two flexible coil arrays arranged in a spatially staggered manner, which correspond to the preset spatial coupling positions of the first spiral coil and the second spiral coil in the wearable contact lens, respectively. A high-magnetic permeability isolation sheet is provided between the two spatially staggered flexible coil arrays. The high-magnetic permeability isolation sheet is made of a high-performance nanocrystalline alloy material or a flexible soft magnetic composite silicon steel sheet with a magnetic permeability of not less than 2000, and is inserted in a normal direction perpendicular to the two spatially staggered flexible coil arrays to form a magnetic flux shielding path to limit magnetic flux leakage and reduce magnetic coupling crosstalk between the two spatially staggered flexible coil arrays. Each of the spatially staggered flexible coil arrays constitutes an independent resonant receiving circuit with a group of adjustable matching capacitors. The adjustable matching capacitors are varactor devices with voltage-controlled capacitance characteristics. Their control input signals are dynamically calculated and issued by the mobile computing terminal based on the amplitude and phase indicators of the current received signal, and are used to perform real-time tuning of the impedance characteristics of the resonant circuit to improve the selectivity of the frequency response curve and compensate for the non-ideal coupling offset under actual wearing conditions.

5. The wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm according to claim 4 is characterized in that: A flexible controllable fine-tuning component is arranged between the high magnetic permeability isolation plate and the two spatially staggered flexible coil arrays. The flexible controllable fine-tuning component includes a miniature flexible driving arm made of shape memory alloy material and a repeatedly deformable supporting connecting plate. The flexible controllable fine-tuning component is used to drive the two spatially staggered flexible coil arrays to achieve relative rotation fine-tuning along the normal direction or the lateral direction within a preset angle range, so that when the wearable contact lens is positionally offset or the wearing posture of the external power supply receiving glasses or eye masks changes, the mechanical posture linkage and the capacitance parameter are coordinated to achieve adaptive recovery of the spatial coupling alignment state of the first spiral coil and the second spiral coil, and improve the decoupling accuracy and signal stability of the corresponding frequency components of the first spiral coil and the second spiral coil in the composite signal in their respective receiving channels.

6. The wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm according to claim 1 is characterized in that: Before performing the sparse matrix decomposition, the mobile computing terminal performs dynamic time-domain modeling on the signal morphologies of the eyelid pressure signal and the intraocular pressure signal based on a trend of change in the spectral energy distribution of the extracted multi-frequency signal within a sliding time window. During the modeling process, the eyelid pressure signal is regarded as a short-duration, bursty, high-amplitude impact signal, and the intraocular pressure signal is regarded as a low-frequency, slowly varying, steady trend signal. The modeling parameters are initialized based on historical monitoring data of users wearing contact lenses and a standard ocular surface pressure physiological model to improve individual adaptability of the model. After completing the time domain modeling, the mobile computing terminal respectively constructs a Gaussian function dictionary for reconstructing the eyelid pressure signal and a piecewise linear basis dictionary for reconstructing the intraocular pressure signal, and performs the sparse matrix decomposition process through L1 norm minimization and joint regularization to perform unsupervised sparse deconstruction and structural reconstruction of the eyelid pressure signal and the intraocular pressure signal in the context of partial overlap of frequency components and significant differences in amplitude dynamic range. The Gaussian function dictionary is used to characterize the local impact disturbance generated during the eyelid closure and blinking cycle, and the piecewise linear basis dictionary is used to characterize the slow change process of intraocular pressure under circadian rhythm and resting state.

7. The wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm according to claim 6, characterized in that: After completing the sparse matrix decomposition, the mobile computing terminal performs an automatic feedback residual evaluation mechanism based on the time-frequency error index between the reconstructed signal and the original separated signal. When the reconstructed residual exceeds a threshold or the sparsity index does not meet a preset condition, a local update process of the sparse dictionary is dynamically triggered. In this process, a joint feature offset vector between the user's historical data and the standard template is constructed, the feature atomic vector used in the current sliding time window is corrected, and the regularization weight coefficient and decoupling-related parameters are updated; The mobile computing terminal is also used to send the updated eyelid pressure signal and the intraocular pressure signal to the input end of the timing self-calibration algorithm respectively, and perform a calibration operation in combination with a preset corneal biomechanical model. The calibration result is used to correct the measurement error caused by the closed eye state, eyeglass posture deviation or soft lens eccentricity in real time, and generate a dynamically updated eyelid pressure curve and the intraocular pressure curve to achieve continuous and stable ocular surface pressure monitoring and subsequent abnormality detection.

8. The wireless ocular surface pressure monitoring system based on multi-frequency phase shift decoupling self-calibration algorithm according to claim 1 is characterized in that: The time-series self-calibration algorithm performs parameter fitting based on a preset corneal biomechanical response model. The corneal biomechanical response model is based on the nonlinear pressure-displacement curve of the central area of ​​the cornea under different transient pressures. The algorithm fits and dynamically updates two response parameters: the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor. The corneal equivalent stiffness coefficient is used to describe the average anti-deformation ability of the cornea under the intraocular pressure background, and the corneal nonlinear recovery factor is used to reflect the deformation recovery characteristics of the cornea during the rapid change stage of the eyelid pressure signal. The timing self-calibration algorithm identifies the time interval between the rising edge and the falling edge of the eyelid pressure signal in the eye-closed state, and is used to calculate the eyelid pressure application cycle within the current time window. The algorithm combines the local change trend of the intraocular pressure signal to determine the maximum response interval of the cornea under the combined action of the eyelid pressure signal and the intraocular pressure signal. Based on this interval, the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor are dynamically adjusted to adapt to the composite force state under different eyelid dynamic and intraocular pressure trend conditions during the wearing of the wearable contact lens. After each sliding time window ends, the time series self-calibration algorithm combines the eyelid pressure signal change range and the intraocular pressure signal change rate within multiple time windows to construct a window weighted fitting result, performs a rolling update on the corneal equivalent stiffness coefficient and the corneal nonlinear recovery factor, and generates a fitting calibration factor for current window correction; The mobile computing terminal is used to apply the fitting calibration factor output by the timing self-calibration algorithm to the timing correction process of the eyelid pressure signal and the intraocular pressure signal, and output the calibrated eyelid pressure curve and the intraocular pressure curve to improve the pressure monitoring stability and signal consistency of the wearable contact lens under conditions of closed eyes, eccentric wearing or posture changes.

Images