A target tracking method, device and medium for intelligent AI glasses based on a transparent film UWB antenna

Abstract

The application relates to a target tracking method, device and medium of intelligent AI glasses based on a transparent film UWB antenna, the method comprising the following steps: hardware initialization and environment calibration; the antenna array adopts a transparent conductive film material of a metal copper grid, signal time domain sampling values of the calibrated UWB-MIMO antenna array are acquired, and a fused pose estimation is obtained through a loose coupling fusion algorithm; compressed sensing multipath suppression is performed on the signal time domain sampling values, and denoised UWB distance measurement values are obtained; a dynamic distribution bandwidth ratio is set according to a mixed scheduling strategy based on device priority weight and channel occupancy rate, and the fused pose estimation and the denoised UWB distance measurement values are transmitted according to the bandwidth ratio. Compared with the prior art, the application has the advantages of overcoming excessive Bluetooth occupation of UWB bandwidth in strong interference, and preventing lagging caused by excessive fluctuation of bandwidth due to interference.

Description

A target tracking method, device, and medium for smart AI glasses based on a transparent thin-film UWB antenna. Technical Field

[0001] This invention relates to the technical field of smart AI glasses, and in particular to a target tracking method for smart AI glasses based on a transparent thin-film UWB antenna. Background Technology

[0002] With the intelligent development of wearable devices, smart glasses, as the core terminal for AR / VR, navigation and positioning scenarios, urgently need to integrate UWB (ultra-wideband) technology to achieve centimeter-level positioning accuracy and low-latency data transmission.

[0003] Existing UWB antennas generally employ LDS / FPC technology or ceramic substrate structures (such as FR4 material), whose physical dimensions and profile height are difficult to adapt to the thin and light design requirements of eyeglasses. For example, traditional UWB patch antennas require at least 10mm × 15mm of layout space in the 8GHz band (referencing the IEEE 802.15.4z standard), forcing them to be concentrated at the ends of the temples in eyeglasses. This layout not only results in bulky and heavy temples (typically 8-12mm thick), but also significantly reduces antenna efficiency (< -3dBi) due to electromagnetic wave reflection losses from the metal conductor and head (measured return loss > -10dB). This makes it impossible to meet the radiation pattern and polarization requirements of UWB systems, leading to poor adaptability of traditional UWB antennas in eyeglasses.

[0004] Currently available glasses with integrated UWB are limited by the spatial arrangement conflict between optical lenses and circuit modules, and generally share the same area with the Bluetooth / WiFi module, resulting in severe signal interference, navigation failure, and a decline in user experience.

[0005] In pursuit of lightweight design, mainstream products employ flexible PCB antennas directly attached to the inside of the temples. However, the sheet resistance of transparent conductive materials (such as ITO film) increases dramatically when bent (sheet resistance > 5Ω / sq after a 180° bend), causing a sharp drop in signal transmission rate (measured < 10Mbps), making it impossible to support high-definition audio transmission (the minimum bandwidth required for encoding is 15Mbps). This contradiction means that existing technologies cannot simultaneously meet the core requirements of "lightweight design" and "high-performance UWB communication."

[0006] Current target tracking methods for smart AI glasses suffer from problems such as Bluetooth excessively consuming UWB bandwidth under strong interference and stuttering caused by excessive bandwidth fluctuations due to interference. Summary of the Invention

[0007] The purpose of this invention is to provide a target tracking method for smart AI glasses based on a transparent thin-film UWB antenna, which overcomes the problem of Bluetooth excessively occupying UWB bandwidth under strong interference and the stuttering caused by excessive bandwidth fluctuations due to interference, while achieving a thinner and lighter design, lower loss, and higher transmission efficiency.

[0008] The objective of this invention can be achieved through the following technical solutions:

[0009] A target tracking method for smart AI glasses based on a transparent thin-film UWB antenna, the method comprising the following steps:

[0010] Hardware initialization and environmental calibration are performed to obtain the calibrated UWB-MIMO antenna array for the AI ​​glasses. The antenna array uses a transparent conductive thin film material of metal copper mesh to form a double-layer cross array structure on the lens substrate of the AI ​​glasses.

[0011] The time-domain sampled values ​​of the calibrated UWB-MIMO antenna array are obtained, and the fused pose estimate is obtained through a loosely coupled fusion algorithm.

[0012] Compressed sensing multipath suppression is performed on the time-domain sampled values ​​of the signal to obtain the denoised UWB distance measurement values;

[0013] Based on device priority weights and channel occupancy rates, a hybrid scheduling strategy is adopted to set dynamically allocated bandwidth ratios, and the fused pose estimates and denoised UWB distance measurements are sent according to the bandwidth ratios.

[0014] Furthermore, the specific steps of the hardware initialization and environmental calibration include:

[0015] Calculate the deformation rate of the calibrated antenna substrate. If the deformation rate of the calibrated antenna substrate does not meet the requirements, redesign the UWB-MIMO antenna array.

[0016] Conversely, the position deviation is eliminated based on the phase difference compensation algorithm to obtain the calibrated UWB-MIMO antenna array.

[0017] Furthermore, the specific steps for obtaining the fused pose estimate using the loosely coupled fusion algorithm are as follows:

[0018] The IMU data and the time-domain sampled values ​​of the calibrated UWB-MIMO antenna array are acquired. The IMU data is zero-biased and the time-domain sampled values ​​of the signal are eliminated by wavelet thresholding. The calibrated IMU data and the time-domain sampled values ​​of the signal after eliminating multipath interference are loosely coupled and fused to obtain the fused pose estimate. The fused pose estimate is used to perform the positioning task.

[0019] Furthermore, the specific steps for performing compressed sensing multipath suppression on the time-domain sampled values ​​of the signal to obtain the denoised UWB distance measurement values ​​are as follows:

[0020] The time-domain sampled values ​​of the calibrated UWB-MIMO antenna array are sparsely reconstructed to extract the main path component. The main path component is then post-processed to output the denoised UWB distance measurement value, which is used for navigation.

[0021] Furthermore, the specific steps for setting the dynamically allocated bandwidth ratio using a hybrid scheduling strategy based on device priority weights and channel occupancy are as follows:

[0022] Obtain device priority weight, channel occupancy rate, location weight, interference level, and total bandwidth;

[0023] The base bandwidth is calculated based on the location weight and total bandwidth.

[0024] Calculate UWB bandwidth and Bluetooth bandwidth;

[0025] The bandwidth ratio is obtained based on UWB bandwidth and Bluetooth bandwidth.

[0026] Furthermore, the basic bandwidth is:

[0027]

[0028] in, To determine the weight, This represents the total weight of Bluetooth. Total bandwidth Interference level.

[0029] Furthermore, the UWB bandwidth is:

[0030]

[0031] in, This indicates the channel occupancy rate of UWB.

[0032] Furthermore, the Bluetooth bandwidth is:

[0033] .

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The exponential term introduced in step 2 of this invention can non-linearly attenuate the Bluetooth weight based on the UWB interference intensity, avoiding excessive Bluetooth bandwidth occupation during strong interference. This aligns with the practical logic that stronger interference necessitates ensuring positioning accuracy. The formula used for UWB bandwidth calculation can moderately adjust the impact of interference on bandwidth increments, preventing stuttering caused by excessive bandwidth fluctuations due to interference, and ensuring that the user's basic experience remains intact.

[0036] This invention utilizes a transparent conductive thin film material with a copper mesh to construct a transparent UWB antenna with a double-layer cross array structure on the lens substrate. It offers significant advantages: its light transmittance is ≥85% (meeting testing standards), maximizing the preservation of the lens's optical transparency and avoiding the light transmittance loss caused by metal wiring in traditional antennas, thus meeting the optical requirements of smart glasses; simultaneously, its sheet resistance is ≤0.5Ω / sq, effectively solving the problem of sheet resistance spikes when traditional transparent conductive materials are bent, ensuring the stability and efficiency of signal transmission. This achieves integrated antenna and curved lens design, overcoming the limitations of traditional UWB antennas in terms of layout and bulkiness on glasses, while also meeting the core requirements of UWB communication for low loss and high transmission efficiency. Attached Figure Description

[0037] Figure 1 is a flowchart of the present invention;

[0038] Figure 2 shows the architecture of AI glasses. Detailed Implementation

[0039] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0040] Figure 1 is a flowchart of the present invention.

[0041] The thin-film antenna substrate used in this invention is not limited to PET (polyethylene terephthalate, a common thermoplastic polyester) material, but also includes high-performance COP (cyclic olefin polymer) and CPI (transparent polyimide) materials. For high-precision optical requirements, COP is preferred (e.g., camera lenses, AR optical modules), while for extreme environments and flexibility requirements, CPI is essential (e.g., foldable screens, aerospace devices). The interface bonding technology between the transparent film and the CPI / COP substrate (e.g., plasma pretreatment to improve adhesion) and the use of a gradient material transition layer (e.g., gradient-doped magnesium fluoride film) for the lens suppress cracking caused by thermal deformation (cracking rate <0.1%).

[0042] Stress relief structure design based on shape memory polymer (SMP) and low temperature conformal packaging technology for CPI lenses (packaging thickness <30μm, bending fatigue life >100,000 cycles).

[0043] 2. Phase interference suppression method for multi-band composite antenna arrays.

[0044] Based on the wavelength characteristics of different frequency bands, non-uniformly distributed antenna element spacing (1 / 3-1 / 2 wavelength) is designed to avoid phase superposition distortion. The phase difference can be controlled within ±5° (measured according to NIST standards), reducing phase noise when multi-band signals are superimposed.

[0045] Meanwhile, by stacking multiple dielectric substrates, current paths of different frequency bands are separated, cross-frequency interference is suppressed, isolation is increased to more than 25dB, and cross-frequency crosstalk is reduced.

[0046] Based on advanced chip solutions and real-time environmental monitoring, the phase distribution of the antenna array is dynamically adjusted through digital signal processing (DSP). Within a temperature range of -20℃ to 60℃, the phase drift is less than 2° and the positioning accuracy is maintained at ±5cm (ETSI standard).

[0047] 3. Intelligent positioning and tracking algorithm architecture based on UWB-MIMO.

[0048] Through dynamic beamforming algorithms or hybrid positioning strategies, the algorithm architecture includes a loosely coupled UWB-IMU fusion algorithm and a multipath suppression method based on compressed sensing. Through hardware co-design of FPGA-accelerated pulse compression modules and dynamic bandwidth allocation protocols, and based on the path prediction model in AR navigation and the AoA device discovery protocol for social functions, the system can achieve centimeter-level positioning accuracy and millisecond-level response speed in wearable devices such as smart glasses, while taking into account the requirements of low power consumption and high reliability.

[0049] 4. Dynamic bandwidth allocation mechanism (Bluetooth / QoS priority scheduling).

[0050] The dynamic bandwidth allocation mechanism achieves seamless coordination between Bluetooth low-latency audio and UWB high-bandwidth data through intelligent priority scheduling, adaptive channel allocation, and interference suppression technologies. An adaptive weight allocation algorithm based on device type and service characteristics, along with an integrated design of an FPGA-implemented protocol stack accelerator and scheduling engine, ensures zero latency for navigation commands while maintaining positioning accuracy within ±5cm. The collaborative frequency hopping mechanism between Bluetooth FHSS and UWB DS-UWB prioritizes user-side audio calls, while other devices compete for remaining bandwidth via CSMA / CA, enabling an exceptional multi-user social interaction experience.

[0051] High-precision positioning and dynamic resource management are achieved through the synergy of multimodal sensor fusion, antenna design, and intelligent algorithms. The overall process consists of the following steps:

[0052] Step 1: Hardware initialization and environment calibration.

[0053] Objective: To establish basic hardware parameters and coordinate system benchmarks.

[0054] 1. Material compatibility design (corresponding to point 1 of this invention).

[0055] Input: CPI / COP substrate thermal expansion coefficient (α_COP=5×10) -6 / ℃), gradient doped magnesium fluoride film thickness (d_FMG=50nm).

[0056] Operation: Plasma pretreatment (power P=300W, time t=60s) enhances the adhesion between CPI and the transparent conductive film (peel force F≥15N / m).

[0057] Gradual refractive index model for graded-doped magnesium fluoride thin films:

[0058]

[0059] Wherein, nMgF2=1.38, nSiO2=1.46, dtotal=100μm.

[0060] Output: Deformation rate of the calibrated antenna substrate (cracking rate ≤ 0.1%).

[0061] 2. UWB-MIMO antenna array calibration.

[0062] Input: Coordinates of each antenna element (xi,yi,zi), center frequency fc=7.9872GHz.

[0063] Operation: Eliminate position deviation based on phase difference compensation algorithm:

[0064]

[0065] Output: The calibrated antenna phase matrix Φcalib.

[0066] Step 2: UWB-IMU loosely coupled fusion algorithm.

[0067] Objective: To achieve centimeter-level positioning and attitude synchronization.

[0068] Algorithm execution flow:

[0069] 1. Input parameters:

[0070] IMU data: acceleration ax / y / z[k], angular velocity ωx / y / z[k] (sampling period Ts=1ms);

[0071] UWB raw distance measurement d(i)UWB[k] (from the i-th anchor point).

[0072] 2. Preprocessing stage:

[0073] Zero bias correction is performed on the IMU data:

[0074]

[0075] Where ba is the static zero bias and Ka=0.1 is the adaptive filter coefficient.

[0076] UWB signal denoising: Wavelet thresholding method is used to eliminate multipath interference.

[0077] 3. Loosely coupled fusion:

[0078] Construct a state vector x[k]=[px,py,pz,qw,qx,qy,qz]T (position + quaternion pose);

[0079] Prediction step (IMU driven):

[0080]

[0081] Where u[k] is the IMU input and w[k] is the process noise (covariance matrix Q).

[0082] Update step (UWB constraint): z[k]=h(x[k])+v[k].

[0083] Where h(·) is the nonlinear observation function, and v[k] is the measurement noise (covariance matrix R).

[0084] Kalman gain Kk calculation:

[0085]

[0086] P k − ： The covariance matrix of the predicted state represents the uncertainty of the predicted state.

[0087] H k The observation matrix maps the state space to the observation space (e.g., converting system states into sensor measurements).

[0088] R k : Measure the noise covariance matrix to describe the uncertainty of sensor measurements;

[0089] ⊤: Matrix transpose symbol;

[0090] Output: Fusion pose estimate xfused[k] (accuracy ±5cm, ETSI standard).

[0091] Step 3: Compressed sensing multipath suppression module.

[0092] Objective: To eliminate multipath interference in UWB signals.

[0093] Algorithm execution flow:

[0094] 1. Input parameters:

[0095] Received signal time-domain sampled value s[n];

[0096] Measurement matrix Φ∈R M×N (M≪N, satisfying the RIP condition).

[0097] 2. Sparse Reconstruction:

[0098] Multipath components are modeled as sparse vectors x, and solved iteratively using the OMP algorithm:

[0099]

[0100] Extract the main path components and suppress the remaining reflection paths.

[0101] 3. Post-processing:

[0102] The reconstructed signal is output after passing through a matched filter:

[0103]

[0104] Where, τ l Let be the delay of the l-th path.

[0105] Output: Denoising UWB distance measurement value d clean [k].

[0106] Step 4: Dynamic bandwidth allocation and protocol coordination.

[0107] Objective: To optimize Bluetooth / UWB resource contention.

[0108] Algorithm execution flow:

[0109] Input parameters:

[0110] Equipment priority weight w device (Voice devices = 1.0, ordinary devices = 0.5);

[0111] Channel occupancy rate C used .

[0112] Hybrid scheduling strategy:

[0113] CSMA / CA-based backoff mechanism:

[0114] BackoffTime=CW min +Rand(0,CW max )×T slot .

[0115] PGA-accelerated pulse compression module:

[0116]

[0117] Where p(t) is the matched impulse function.

[0118] Output: Dynamically allocated bandwidth ratio B UWB :B BT (Prioritize positioning accuracy of ±5cm).

[0119] This invention employs a transparent UWB antenna design, utilizing a transparent conductive thin film material with a copper mesh to form a double-layer cross array structure on a lens substrate. The antenna transmittance is ≥85% (transmittance test conditions: AM 1.5 spectrum, 100mW / cm²). 2 (Illumination), sheet resistance ≤ 0.5Ω / sq (sheet resistance measurement method: four-probe method, 25℃ environment).

[0120] Operating frequency band CH9 (7.9872GHz±10MHz), 500MHz ultra-wide bandwidth design;

[0121] Multi-physics field integrated architecture - dual-lens symmetrical layout scheme: transparent antenna 1 / 2 is integrated on the inner surface of lens 1 / 2 respectively, and electromagnetic isolation design between UWB module circuit area and Bluetooth audio module (isolation degree > 20dB).

[0122] The antenna feed network adopts a coplanar waveguide structure with impedance matching of 50Ω±5%. Functional implementation system: Precision positioning and navigation system: based on UWB ToF ranging algorithm (accuracy ±10cm@10m); Intelligent tracking module: integrating IMU sensor and UWBAoA azimuth calculation (angle error <2°).

[0123] High-speed data transmission protocol: Custom MAC layer protocol, supporting 30Mbps adaptive modulation (adjustable to 50Mbps).

[0124] The structure of this invention follows the design requirements of UWB antennas and adopts an innovative centralized design structure---the UWB module circuit area is located in the middle of the lens frame (coordinate position: 3-5cm above the geometric center of the lens), and the transparent antenna 1 / 2 is embedded in the edge of the left and right lenses respectively (coordinate: the horizontal ±3cm area of ​​the lens). The lens 1 / 2 adopts a curved glass substrate (curvature radius R=300-500mm). The UWB module is connected to the FPC cable in the temple of the lens by ultrasonic welding (line width 0.15mm). The antenna feed point adopts a coplanar waveguide structure (impedance deviation <±5%). The antenna spacing is between 1 / 3 and 1 / 2 wavelength (the center distance between the two transparent antennas is only 14-18cm, which meets the baseline requirements for UWBAoA positioning).

[0125] Figure 2 shows the architecture of the AI ​​glasses. The AI ​​glasses include a wireless communication module and a processor (CPU, GPU). The wireless communication module includes a UWB module and Bluetooth 5.0+. The wireless communication module is located at the first antenna interface ANT1 and the second antenna interface ANT2. The processor includes a main chip and is connected to the battery, audio, tracking camera, sensors, USB / DP, and optical module.

[0126] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A target tracking method for smart AI glasses based on a transparent thin-film UWB antenna, characterized in that, The method includes the following steps: Hardware initialization and environmental calibration are performed to obtain a calibrated UWB-MIMO antenna array for the AI ​​glasses. This antenna array uses a transparent conductive thin film material with a copper mesh, forming a double-layer cross-array structure on the lens substrate of the AI ​​glasses. The time-domain sampling values ​​of the calibrated UWB-MIMO antenna array are acquired, and a fused pose estimate is obtained through a loosely coupled fusion algorithm. Compressed sensing multipath suppression is performed on the time-domain sampling values ​​to obtain denoised UWB distance measurements. A dynamically allocated bandwidth ratio is set using a hybrid scheduling strategy based on device priority weights and channel occupancy. The fused pose estimate and denoised UWB distance measurements are then transmitted according to this bandwidth ratio. The specific steps for setting the dynamically allocated bandwidth ratio using a hybrid scheduling strategy based on device priority weights and channel occupancy are as follows: Obtain device priority weights, channel occupancy, positioning weights, interference levels, and total bandwidth; calculate the base bandwidth based on the positioning weights and total bandwidth; calculate the UWB bandwidth and Bluetooth bandwidth; obtain the bandwidth ratio based on the UWB bandwidth and Bluetooth bandwidth; the base bandwidth is: in, To determine the weight, This represents the total weight of Bluetooth. Total bandwidth Interference level; UWB bandwidth: in, This indicates the channel occupancy rate of UWB; Bluetooth bandwidth is: 。 2. The target tracking method for smart AI glasses based on a transparent thin-film UWB antenna according to claim 1, characterized in that, The specific steps of the hardware initialization and environmental calibration include: calculating the deformation rate of the calibrated antenna substrate; if the deformation rate of the calibrated antenna substrate does not meet the requirements, the UWB-MIMO antenna array is redesigned; otherwise, the position deviation is eliminated based on the phase difference compensation algorithm to obtain the calibrated UWB-MIMO antenna array.

3. The target tracking method for smart AI glasses based on a transparent thin-film UWB antenna according to claim 1, characterized in that, The specific steps for obtaining the fused pose estimate using the loosely coupled fusion algorithm are as follows: acquire IMU data and the time-domain sampled values ​​of the calibrated UWB-MIMO antenna array; perform zero-bias correction on the IMU data and use wavelet thresholding to eliminate multipath interference on the time-domain sampled values ​​of the signal; perform loosely coupled fusion on the calibrated IMU data and the time-domain sampled values ​​of the signal after eliminating multipath interference to obtain the fused pose estimate, which is used to perform the positioning task.

4. The target tracking method for smart AI glasses based on a transparent thin-film UWB antenna according to claim 1, characterized in that, The specific steps for performing compressed sensing multipath suppression on the signal time-domain sampled values ​​to obtain denoised UWB distance measurement values ​​are as follows: sparse reconstruction is performed on the signal time-domain sampled values ​​of the calibrated UWB-MIMO antenna array, the main path components are extracted, and the main path components are post-processed to output denoised UWB distance measurement values, which are used for navigation.

5. A target tracking device for smart AI glasses based on a transparent thin-film UWB antenna, comprising a memory, a processor, and a program stored in the memory, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-4.

6. A storage medium having a program stored thereon, characterized in that, When the program is executed, it implements the method as described in any one of claims 1-4.

Images