An adaptive intelligent glasses complete system

Abstract

This invention discloses an adaptive smart glasses system, belonging to the fields of smart wearable devices and adaptive optics technology. The system uses a polarization-independent gradient refractive index liquid crystal lens as its core, integrating a multimodal sensing unit, an AI intelligent decision-making unit, a polymer nano-transmission unit, a modular overall control unit, and a multi-scene adaptation unit, forming an integrated closed-loop system of "perception-decision-transmission-execution-feedback-self-learning." This invention addresses the shortcomings of similar products worldwide (Apple's modular glasses, Finland's IXI eye-tracking glasses, etc.) and related research, innovatively achieving a deep integration of polarization-independent adaptive focusing, multimodal contextual awareness, modular hardware adaptation, and low-loss nano-transmission. This solves the pain points of existing smart glasses, such as polarization dependence, low focusing accuracy, functional fragmentation, poor adaptability, and inability to achieve overall coordinated control. The system can achieve fully automatic refraction in 10 seconds, adaptive focusing within 180ms, light transmittance ≥82%, focusing error ≤±0.10D, and supports myopia and hyperopia correction from -10.00D to +6.00D and astigmatism correction up to ±6.00D. It also integrates AR display, eye health monitoring, multi-scene mode switching, and AI voice interaction functions. The system is compact, mass-producible, and highly versatile. This invention features outstanding overall synergy, clear innovations, and performance superior to similar international products. It can be widely applied in daily vision correction, AR / VR interaction, medical eye health monitoring, outdoor travel, and other scenarios, possessing extremely high engineering application value and patent protection value.

Description

Technical Field

[0001] This invention relates to the fields of intelligent wearable devices, adaptive optics, AI intelligent control and nanotechnology, specifically an adaptive intelligent glasses system suitable for various scenarios such as daily vision correction, AR / VR interaction, medical eye health monitoring, outdoor travel, and office meetings. It belongs to the field of intelligent glasses system integration and collaborative control technology. Background Technology

[0002] With the rapid development of smart wearable devices, adaptive smart glasses have become a research hotspot in the fields of vision correction and intelligent interaction, with global shipments increasing year by year, reaching 14.773 million units in 2025. The Chinese market saw an even higher growth rate of 87.1%. However, existing adaptive smart glasses generally suffer from fatal flaws such as "functional fragmentation, poor coordination, polarization dependence, and limited adaptability," making it impossible to achieve integrated adaptive control of the entire system. Specifically: 1) Poor overall coordination: The modules such as lens focusing, environmental adaptation, AI function, and health monitoring are independent of each other and lack a unified decision-making and transmission system, resulting in problems such as response delay and asynchronous control, leading to a poor user experience; 2) Severe polarization dependence: Most smart glasses use traditional liquid crystal lenses. Changes in incident polarization can lead to a decrease in focusing accuracy, with errors reaching more than 0.5D. The light transmittance is extremely low in low light environments (generally <50%), which cannot meet the needs of use in all scenarios. 3) High transmission loss and weak anti-interference: Using traditional metal wires or ordinary wireless transmission results in a transmission loss of ≥10%, making it susceptible to electromagnetic interference and affecting control accuracy and signal stability. 4) Limited adaptability and scalability: Most products have a fixed structure and cannot change functional modules according to user needs, making it difficult to adapt to the personalized needs of different face shapes and different usage scenarios. 5) Insufficient intelligence: It lacks multimodal context awareness capabilities, cannot achieve automatic scene switching, requires manual intervention and adjustment, and has no self-learning ability, making it unable to adapt to users' long-term usage habits; 6) Low functional integration: It is difficult to integrate multiple functions such as vision correction, AR display, eye health monitoring, and AI interaction. Either the functions are limited or they interfere with each other, affecting the user experience. Comparison of existing international technologies (referencing relevant products and publications worldwide): Apple has published multiple patents for modular smart glasses that support the replacement of modules such as temples and batteries. However, they lack polarization-independent lens design, focus relies on single distance perception, lack multimodal context perception and closed-loop collaborative control of the whole device, and do not integrate eye health monitoring functions. Finland's IXI: Eye-tracking focusing glasses have issues such as polarization dependence, short battery life, and high transmission latency (≥500ms). They lack nanometer transmission design and their functions are limited to focusing, making it impossible to achieve multi-scenario adaptation and AI function linkage. Zeiss (Germany): The polarization-independent optimized lens, ΔT=6.8%, still has a significant polarization dependence and has not formed a whole system integration, focusing only on the lens itself, lacking AI decision-making and multimodal perception capabilities; Berkeley, USA: Metasurface liquid crystal lens, polarization-independent but with only 60% light transmittance and a focusing response of ≥600ms, not integrated with overall control, nano-transmission, and AI interaction; Domestic brands include Fengyuzhu: AI glasses with dual-mode navigation system, which focuses on scene switching and voice interaction, but lacks polarization-independent focusing and eye health monitoring functions, and the overall synergy of the device is insufficient; Quark AI glasses: have the ability to adjust the image distance and handle low light, but have not solved the polarization-dependent problem and have limited module scalability. Related publications (such as "Adaptive Liquid Crystal Lenses for Smart Glasses" and "Modular Design of AI-Powered Wearable Devices") only focus on the research of a single module (lens or AI interaction) and do not propose an integrated closed-loop system of "polarization-independent lens + multimodal sensing + nanometer transmission + modular device", which cannot solve the pain points of synergy and adaptability of existing technologies. To address these shortcomings, this invention, drawing upon the technical approaches of relevant products and publications worldwide, proposes an integrated closed-loop system comprising "polarization-independent optical lenses + multimodal perception + AI decision-making + nanometer transmission + modular assembly." This system enables coordinated control of the entire adaptive smart glasses system, filling a technological gap in similar products globally and enhancing the product's practicality and market competitiveness. Summary of the Invention

[0003] Technical solution The present invention discloses an adaptive smart glasses system, the core of which lies in realizing a closed-loop collaborative mechanism of "perception-decision-transmission-execution-feedback-self-learning" throughout the entire system. The various units are interconnected and seamlessly connected, and the specific structure is as follows: Polarization-independent optical lens unit: As the core execution component of the whole machine, it adopts a four-layer composite structure of substrate layer-metasurface layer-liquid crystal layer-protective layer, with built-in bipolar polarizer lens template. The angle between the long axis of the metasurface layer and the projection angle of the molecular director of the liquid crystal layer is strictly controlled at 45°±0.3°, eliminating the need for a polarizer. The light transmittance is ≥82%, the polarization independence ΔT≤3.2%, and the steady-state focusing error is ≤±0.10D. It can achieve stepless continuous focusing from -10.00D to +6.00D myopia, hyperopia, and ±6.00D astigmatism, with a focusing response time ≤180ms, no jumps, and no distortion. The lens edge integrates a nano-transmission interface, which can be directly connected to the whole machine's nano-transmission unit. The multimodal perception unit integrates an optical environment perception module (collecting light intensity, color temperature, polarization state, and glare intensity), an ocular physiological perception module (collecting intraocular pressure, axial length, pupil diameter, corneal curvature, and refractive state), a user behavior perception module (collecting viewing distance, eye movement trajectory, and head posture), and a scene semantic perception module (collecting scene type, user task, and usage status). It adopts a synchronous sampling mechanism to ensure the temporal consistency of data from each modality, providing comprehensive and accurate data support for AI decision-making. AI Intelligent Decision Unit: Employing a lightweight deep learning model, it incorporates five sub-models: optometry and focusing decision, ambient light adaptation decision, eye health monitoring decision, AI application scheduling, and multi-scene mode switching. It can normalize, filter noise, align and weight fusion multimodal perception data, generate adaptive control commands in real time, and achieve a decision response time of ≤10ms. It supports multi-task parallel processing and can simultaneously realize focusing, environmental adaptation, and AI function linkage without interference. Polymer nanotransmission unit: Adopting a carbon nanotube-polyimide-graphene composite transmission structure, the transmission loss is ≤3%, and the anti-electromagnetic interference capability is improved by ≥80%. It realizes the integrated low-loss transmission of control signals, energy and optometry data, and connects the multimodal sensing unit, AI intelligent decision-making unit, modular whole machine control unit and polarization-independent optical lens unit to ensure the real-time performance and stability of signal transmission. Modular control unit: Adopting a detachable modular design, it includes a core control module, power module, display module, and voice interaction module. The temples, battery, and core function modules can be replaced and expanded independently. Standardized interfaces are reserved, and peripherals such as cameras, dedicated NPUs, and motion monitoring modules can be added as needed. After the addition, the system automatically adapts and allocates computing power without modifying the internal core components, adapting to the personalized needs and scenario requirements of different users. Multi-scene adaptation unit: Preset daily mode, outdoor strong light mode, indoor reading mode, AR interaction mode, and eye health monitoring mode, which can automatically switch according to multimodal perception data, with a switching response time of ≤50ms; In outdoor strong light mode, it automatically reduces lens transmittance and enhances glare suppression; In indoor reading mode, it increases transmittance and activates eye protection mode; In AR interaction mode, it realizes accurate overlay of virtual information and real field of vision, and supports 1080P resolution projection; In eye health monitoring mode, it monitors eye health parameters in real time and issues abnormal warnings. Closed-loop feedback unit: Real-time acquisition of lens focusing effect, user visual feedback and changes in environmental parameters, calculation of control error and dynamic correction of control commands, closed-loop response time ≤180ms; at the same time, it records the user's historical optometry data, usage habits and scenario preferences, realizes the self-learning optimization of AI decision model, so that the consistency error of repeated optometry is ≤±0.02D, and the accuracy continues to improve with long-term use. 1) The overall machine control method of the present invention includes the following steps: After the system is started, the multimodal perception unit synchronously collects multi-dimensional data on optical environment, eye physiology, user behavior and scene semantics, with the sampling frequency controlled at 15Hz to ensure data timing consistency. Multimodal data is transmitted to the AI ​​intelligent decision-making unit, and after normalization, noise filtering, spatiotemporal alignment and weighted fusion processing, a unified context feature vector is constructed. Based on the context feature vector, the AI ​​intelligent decision-making unit generates adaptive control commands such as optometry and focusing, environmental adaptation, functional linkage, and scene switching through parallel computation of five sub-models. The polymer nanotransmission unit sends control commands to the polarization-independent optical lens unit and the modular overall control unit with low loss, with a transmission delay of ≤10ms and a transmission loss of ≤3%. The polarization-independent optical lens unit executes focusing commands, while the modular whole-machine control unit executes environmental adaptation, function linkage, and scene switching commands to complete adaptive adjustment; The closed-loop feedback unit collects execution results and changes in environmental and physiological parameters in real time, calculates control errors and dynamically corrects control commands, while recording user data and iteratively optimizing the AI ​​decision-making model. The system continuously cycles through the above steps to achieve adaptive closed-loop control across all scenarios, without requiring manual intervention from the user. 2) Innovation Points 1) World's first: Constructing an integrated closed-loop system of "polarization-independent optical lenses + multimodal perception + AI decision-making + nanometer transmission + modular whole machine" to achieve coordinated control of the whole machine for focusing, environmental adaptation, AI functions, health monitoring and scene switching, solving the pain points of existing products' fragmented functions and poor coordination, and differentiating itself from single-function-oriented products such as Apple's modular design and IXI eye-tracking focus. 2) Co-optimization of polarization independence and high transmittance: The polarization-independent gradient refractive index liquid crystal lens is deeply integrated with the whole system. Through the co-design of bipolar polarizer and metasurface, the polarization dependence is eliminated in principle (ΔT≤3.2%), and the transmittance is ≥82%, which is significantly better than Berkeley metasurface liquid crystal lens (transmittance 60%) and Zeiss polarization-independent lens (ΔT=6.8%). 3) Multimodal context awareness and AI self-learning collaboration: Integrating four types of perception modalities, it achieves comprehensive perception of scene, physiology, and behavior. Combined with self-learning algorithms, it can dynamically adapt to user habits. The consistency error of repeated refraction is ≤±0.02D. Its intelligence level far exceeds that of existing similar products, filling the technical gap of multimodal perception and whole-machine collaborative self-learning. 4) Modular and nano-transmission integrated design: It adopts a detachable modular structure, supports the independent replacement and expansion of functional modules, and combines polymer nano-transmission technology to achieve low-loss transmission of signals and energy (loss ≤3%), solving the problems of high transmission loss, weak anti-interference and poor module expandability of traditional transmission. It is different from the defects of Apple's modular design without nano-transmission and traditional smart glasses without modular expansion. 5) Multi-scene fully automatic adaptation: It has five preset scene modes and can automatically switch according to multimodal data with a switching response time of ≤50ms. It integrates AR display, eye health monitoring, AI voice interaction and other functions to achieve "one device for multiple uses", which is superior to the design of Fengyuzhu dual-mode navigation and Quark AI glasses single-function optimization. 6) Closed-loop control system for the whole machine: Construct a complete closed loop of "perception-decision-transmission-execution-feedback-self-learning", with a closed-loop response time ≤180ms and a focusing error ≤±0.10D. Its performance surpasses that of similar international products in all aspects, and it has significant innovation and practicality. 3) Beneficial effects Strong overall coordination: The units are seamlessly connected, enabling synchronous control of focusing, environmental adaptation, AI functions, and health monitoring, with no delay or interference, resulting in a significantly improved user experience; Polarization independent and high light transmittance: No polarizer required, light transmittance ≥82%, consistent focusing accuracy under any polarization state (ΔT≤3.2%), suitable for use in all scenarios such as strong light and weak light; Precise focusing and fast response: 10-second fully automatic refraction, focusing response ≤180ms, steady-state error ≤±0.10D, adaptable to various types of refractive errors, and high visual comfort; Stable transmission and low loss: Nanoscale transmission loss ≤3%, strong anti-electromagnetic interference capability, ensuring the real-time performance and stability of signal and energy transmission; Highly scalable and adaptable: The modular design supports the independent replacement and expansion of functional modules, which can adapt to the personalized needs of different users and different scenarios, and has strong versatility. High level of intelligence: Multimodal perception + AI self-learning enables automatic scene switching and personalized adaptation without manual user intervention, and the accuracy continues to improve with long-term use; High integration of functions: It takes into account functions such as vision correction, AR display, eye health monitoring, AI voice interaction, and multi-scene adaptation, making it a versatile and highly practical device. Highly producible: Each unit has a mature structure, controllable manufacturing process, and modular design reduces production costs, making it suitable for large-scale mass production and possessing extremely high commercial value. Attached Figure Description Figure 1. Block diagram of the adaptive smart glasses system Figure 2. Schematic diagram of polarization-independent optical lens unit structure Figure 3. Closed-loop control flowchart of the whole system Figure 4 Schematic diagram of modular overall structure Figure Labels 100 Adaptive Smart Glasses System; 101 Polarization-Independent Optical Lens Unit; 102 Multimodal Sensing Unit; 103 AI Intelligent Decision-Making Unit; 104 Polymer Nanotransmission Unit; 105 Modular System Control Unit; 106 Multi-Scene Adaptation Unit; 107 Closed-Loop Feedback Unit; 108 Nanotransmission Interface; 201 Substrate Layer; 202 Metasurface Layer; 203 Bisbian Polarizing Template; 204 Liquid Crystal Layer; 205 Protective Layer; 1051 Core Control Module; 1052 Power Module; 1053 Display Module; 1054 Voice Interaction Module; 1055 Expansion Interface. Detailed Implementation Example 1: Fabrication and Control of Standard Adaptive Smart Glasses System Assemble the polarization-independent optical lens unit: Use 0.45mm high-transmittance optical glass as the substrate layer, deposit a 0.01mm metasurface layer by electron beam evaporation, and control the angle between its long axis and the liquid crystal molecules to be 45.0°; use a wash-refill process to stack bipolar polarizer templates, inject polarization-independent liquid crystal material, cure at 85℃ for 25min to form a liquid crystal layer, attach a 7H wear-resistant and oil-resistant protective layer, and fabricate a nano-transmission interface at the edge of the lens; Assemble the multimodal sensing unit: integrate ambient light sensor, intraocular pressure sensor, eye axis measurement module, ToF distance sensor, and scene recognition camera, and calibrate the sampling frequency to 15Hz to ensure synchronous data acquisition from each module; Assemble an AI intelligent decision-making unit: equipped with a lightweight deep learning model, integrating five sub-models, with a debugging decision response time of ≤10ms and supporting multi-task parallel processing; Assemble the polymer nanotransmission unit: Use carbon nanotube-polyimide-graphene composite transmission circuit to connect each unit, and adjust the transmission loss to ≤3% and the transmission delay to ≤10ms. Assemble the modular complete unit control unit: install the core control module, power module, AR display module, and AI voice interaction module, and reserve expansion interfaces to ensure that the modules are detachable and replaceable; It integrates a multi-scenario adaptation unit and a closed-loop feedback unit, with a debugging scenario switching response time of ≤50ms and a closed-loop response time of ≤180ms. System overall debugging: Upon system startup, the multimodal sensing unit collects data such as light intensity of 300 lux, color temperature of 4000K, viewing distance of 0.35m, and indoor reading scenario. The AI ​​decision unit generates control commands for +2.00D focusing, improving light transmittance, and activating eye protection mode. These commands are then sent to each execution unit via the nano-transmission unit. The closed-loop feedback unit performs real-time calibration, ultimately achieving a focusing response of 150ms, an error of 0.06D, a light transmittance of 82.3%, and ΔT=3.2%. System performance: transmittance 82.3%; polarization independence ΔT=3.2%; focusing response ≤180ms; focusing error ±0.10D; transmission loss 2.7%; scene switching response 45ms; repeat refraction consistency error ±0.02D; false wake-up rate 0.3%; voice recognition latency 280ms. Example 2: Outdoor High-Light Scene Adaptation System Based on Example 1, the strong light adaptation algorithm of the multi-scene adaptation unit is optimized. A ±0.1° dynamic fine-tuning mechanism is added to the metasurface layer. In strong light environments (light intensity ≥10000 lux), the lens transmittance is automatically reduced by 20% and glare suppression is enhanced (glare suppression ≥85%). The AI ​​decision unit prioritizes the allocation of computing power to focusing and environmental adaptation to ensure visual clarity and comfort under strong outdoor light. After debugging, the transmittance fluctuation under strong light is ≤1%, the refraction error is ≤±0.06D, and the focusing response is 165ms, meeting the needs of outdoor travel. Example 3 Modular Expandable Complete System Based on Example 1, an ultra-miniature camera and a dedicated NPU module are added through a modular expansion interface. The system automatically identifies the new module and allocates computing power. The camera supports 3K high-definition shooting and eye tracking, and the dedicated NPU module improves the AI ​​decision-making speed by 30%. At the same time, the eye health monitoring function is expanded, which can take real-time photos of the retina, analyze blood vessel morphology, and assist in the diagnosis of diabetic retinopathy, realizing flexible expansion of functions without modifying the core control unit.

Claims

1. An adaptive smart glasses system, characterized in that, include: Polarization-independent optical lens unit, multimodal sensing unit, AI intelligent decision-making unit, polymer nano-transmission unit, modular whole machine control unit, multi-scene adaptation unit and closed-loop feedback unit; Each unit achieves low-loss integrated signal and energy transmission through a nano-transmission unit, forming a closed-loop collaborative control, and realizing overall machine collaboration with adaptive focusing, environmental adaptation, functional linkage, and self-learning optimization.

2. The adaptive smart glasses system according to claim 1, characterized in that, The polarization-independent optical lens unit adopts a four-layer composite structure of substrate layer-metasurface layer-liquid crystal layer-protective layer, with a built-in bipolar polarizer lens template. The angle between the major axis of the metasurface layer and the projection angle of the molecular director of the liquid crystal layer is 45°±0.3°. No polarizer is required, the transmittance is ≥82%, the polarization independence ΔT≤3.2%, the steady-state focusing error is ≤±0.10D, and it is suitable for myopia and hyperopia of -10.00D to +6.00D and astigmatism of ±6.00D.

3. The adaptive smart glasses system according to claim 1, characterized in that, The multimodal perception unit includes an optical environment perception module, an eye physiology perception module, a user behavior perception module, and a scene semantic perception module. It simultaneously collects multi-dimensional data such as light intensity, color temperature, intraocular pressure, axial length, viewing distance, and scene type. The sampling frequency is 10–30Hz, the timing deviation is ≤5ms, and the transmission delay is ≤10ms.

4. The adaptive smart glasses system according to claim 1, characterized in that, The AI ​​intelligent decision-making unit incorporates five sub-models: optometry and focusing decision, ambient light adaptation decision, eye health monitoring decision, AI application scheduling, and multi-scenario mode switching. It adopts a lightweight deep learning algorithm, supports multi-task parallel processing, and can generate adaptive control commands in real time based on multimodal perception data, with a decision response time of ≤10ms.

5. The adaptive smart glasses system according to claim 1, characterized in that, The polymer nanotransmission unit adopts a carbon nanotube-polyimide-graphene composite transmission structure, with a transmission loss of ≤3% and an electromagnetic interference resistance of ≥80%. It can realize the integrated low-loss transmission of control signals, energy and optometry data, and connects to the nanotransmission interface between the lens unit and the whole machine control unit.

6. The adaptive smart glasses system according to claim 1, characterized in that, The modular control unit adopts a detachable modular design, which supports the independent replacement and expansion of the temples, batteries and core functional modules. It has reserved standardized interfaces and can add peripherals such as cameras, dedicated NPUs and motion monitoring modules as needed. After the addition, the system automatically adapts and allocates computing power without modifying the internal core components.

7. The adaptive smart glasses system according to claim 1, characterized in that, The multi-scene adaptation unit supports automatic switching between daily mode, outdoor strong light mode, indoor reading mode, AR interaction mode, and eye health monitoring mode, with a switching response time of ≤50ms. It can automatically adjust the lens optical parameters, display mode, and function priority according to the scene type. The AR mode supports 1080P resolution projection, equivalent to a 120-inch virtual screen.

8. The adaptive smart glasses system according to claim 1, characterized in that, The closed-loop feedback unit collects the lens focusing effect, user visual feedback and environmental parameter changes in real time, dynamically corrects control commands, and has a closed-loop response time of ≤180ms. At the same time, it records user habits and refraction data to achieve self-learning optimization of the AI ​​decision model, so that the consistency error of repeated refraction is ≤±0.02D.

9. A control method for an adaptive smart glasses system, characterized in that, Includes the following steps: 1) The multimodal sensing unit simultaneously collects optical environment, eye physiology, user behavior, and scene semantic data; 2) The AI ​​intelligent decision-making unit performs feature fusion and analysis on the collected data to generate adaptive control commands; 3) The polymer nano-transmission unit sends the control commands to the polarization-independent optical lens unit and the modular whole-machine control unit with low loss; 4) Each execution unit executes the control commands to complete focusing, environmental adaptation, and functional linkage; 5) The closed-loop feedback unit collects the execution results, dynamically corrects the commands, and achieves self-learning optimization.

10. The adaptive smart glasses system according to any one of claims 1-9, characterized in that, It also integrates an eye health monitoring module and an AI voice interaction module, which can monitor pupil response and retinal vascular morphology in real time to assist in the diagnosis of early eye diseases. It also supports functions such as real-time simultaneous interpretation, voice control, and meeting minutes generation, with a false wake-up rate of ≤0.5% and a voice recognition latency of ≤300ms.

Images