Indoor positioning method based on optical AOA and ultrasonic-assisted ranging

Abstract

The invention discloses an indoor positioning method based on optical AOA and ultrasonic auxiliary wave distance measurement, a plurality of groups of acousto-optic combined positioning systems are deployed in each indoor direction, each positioning system comprises an ultrasonic transmitter, an ultrasonic receiver, an LED array, a detector array and a microcontroller, and each microcontroller is connected to a terminal; the LED array performs frequency modulation on the optical signals, the ultrasonic transmitter performs pulse modulation on the ultrasonic signals, and the detector array receives the optical signals from the LED array, records the arrival angle when the optical signals arrive and corrects the arrival angle; the ultrasonic receiver receives the ultrasonic signal, and the distance R between the target point location and the ultrasonic receiver is determined by measuring and calculating the time difference of signal propagation; and the position of the target point location in the three-dimensional space is determined according to the corrected arrival angle and the distance R, a self-adaptive dynamic weighting hybrid filtering algorithm is operated, the space coordinate point location of the target point location is obtained, and indoor positioning of the target point location is achieved.

Description

Technical Field

[0001] The present invention relates to the field of indoor positioning technology, and in particular to an indoor positioning method based on optical AOA and ultrasonic auxiliary wave ranging. Background Art

[0002] Conventional optical positioning includes geometric optics, time optics, and light intensity positioning. Among them, light intensity positioning, namely RSS (Received Signal Strength), is highly dependent on the stability of the ambient light intensity for its accuracy; TOF (Time Of Flight) and pulsed laser positioning, representatives of time optics, both use lasers for ranging. Although they ensure accuracy, they have high requirements for high-precision clocks and are relatively expensive. In addition, lasers can cause certain damage to the human body and are not suitable for indoor positioning. Although geometric optics can avoid the above problems to a certain extent, it also has its own defects. One of its representatives is the multi-base station AOA (Angle Of Arrival) positioning technology.

[0003] The traditional AOA positioning process first receives signals, then uses an AOA measurement algorithm, then employs a geometric positioning algorithm, and finally integrates the data to arrive at an estimated target position. These two steps can lead to measurement and positioning errors, respectively. Furthermore, light propagation is easily blocked by obstacles, resulting in significant discrepancies between the final estimated result and the actual result.

[0004] In addition to geometric optical positioning, ultrasonic positioning is also one of the mainstream positioning methods. However, due to the limited propagation distance of ultrasonic waves in air, current ultrasonic ranging positioning is generally only suitable for measurement ranges of several to more than ten meters. As the propagation distance increases, the signal attenuates severely, positioning accuracy decreases, and long-distance positioning performance is insufficient. Furthermore, the slow propagation speed of sound waves (approximately 340m / s) and the large positioning delay make it unsuitable for accurately locating the direction of moving targets. Furthermore, ultrasonic waves are sensitive to environmental factors such as underwater and wind noise, making positioning easily interfered with.

[0005] All of the above technologies have more or less problems in terms of accuracy, breadth and applicability. Therefore, the emergence of an indoor positioning method that can minimize the impact of the external environment and has high measurement accuracy is particularly important. Summary of the Invention

[0006] In view of this, the present invention provides an indoor positioning method based on optical AOA and ultrasonic-assisted wave ranging, which combines the characteristics of light propagation, such as fast speed, wide range, and little environmental impact, with the ability of ultrasonic waves to penetrate most non-metallic obstacles (such as cloth and smoke), reducing the impact of obstacles to a certain extent, thereby achieving high-precision and high-efficiency indoor positioning.

[0007] The present invention achieves the above technical objectives through the following technical means.

[0008] Indoor positioning method based on optical AOA and ultrasonic assisted ranging:

[0009] Multiple groups of acoustic and optical positioning systems are deployed at various locations indoors, and the microcontrollers in each acoustic and optical positioning system are connected to the same terminal; the acoustic and optical positioning system includes an ultrasonic transmitter, an ultrasonic receiver, an LED array, a detector array, and a microcontroller;

[0010] The LED array frequency modulates the light signal so that the detector array can identify and decode it. At the same time, the ultrasonic transmitter pulse modulates the ultrasonic signal so that the ultrasonic receiver can accurately measure the propagation time of the signal.

[0011] The detector array receives the light signal from the LED array, records the arrival angle of the light signal when it reaches the detector array, and corrects the arrival angle; the ultrasonic receiver receives the ultrasonic signal, determines the distance R between the target point and the ultrasonic receiver by measuring the time difference of signal propagation; and determines the position of the target point in three-dimensional space based on the corrected arrival angle and distance R;

[0012] Based on the position of the target point in three-dimensional space, an adaptive dynamic weighted hybrid filtering algorithm is run to obtain the spatial coordinates of the target point, thereby realizing the indoor positioning of the target point.

[0013] Furthermore, the LED array adopts an intrinsic polarization excitation mechanism. Specifically, under the action of a forward pulse of 0-5mA, carriers will preferentially recombine along the normal direction of the quantum well, generating linearly polarized light with a polarization axis parallel to the lattice; under the action of a reverse pulse of -2-0mA, the symmetry of the carrier distribution is destroyed, thereby generating elliptically polarized light with the long axis at 45° to the lattice direction.

[0014] Furthermore, the LED array adopts a honeycomb layout, with every seven LED nodes forming a regular hexagon forming a group. Within the group, the central LED node and the six adjacent nodes on all sides form a retractable regular hexagonal unit. At this time, the distance between the central LED node and any of the surrounding nodes is the reference spacing; the area where the regular hexagonal unit formed by the seven nodes in the middle of the array is located is the central area, a circle of regular hexagonal units located at the edge of the array forms an edge area, and the rest is the middle area; the distance between the central LED node of the regular hexagonal unit in the central area and any of the surrounding nodes is extended to 1.2 times the reference spacing, while the distance between the central LED node of the regular hexagonal unit in the edge area and any of the surrounding nodes is compressed to 80% of the reference spacing; the height of the LED array is controlled in layers by a retractable bracket, the LED nodes in the central area are raised, the height of the LED nodes in the middle area remains unchanged, and the height of the LED nodes in the edge area is lowered, and the ratio of the LED node raising, height remaining unchanged, and height lowering is 6:5:4.

[0015] Furthermore, the ultrasonic transmitter and the ultrasonic receiver constitute an ultrasonic ranging module, and the ultrasonic ranging module adopts an intelligent micro-vibration cleaning mechanism:

[0016] When the ultrasonic receiver detects an increase in broadband noise caused by particle scattering, the microcontroller identifies it as dust interference. The ultrasonic transmitter then uses a high-frequency pulse blasting mode, initially exciting micron-sized cavitation bubbles with a high-frequency vibration of 150kHz, and using the shock wave from the bubble rupture to peel off particles. The ultrasonic transmitter increases its operating frequency by 5kHz every 10ms to expand the operating frequency band. At the same time, the ultrasonic receiver dynamically monitors the reflected wave signal-to-noise ratio. If the reflected wave signal-to-noise ratio falls below 15dB, the microcontroller triggers a 5ms short pulse to increase the energy density of the ultrasonic transmitter, gradually reducing the frequency to 120kHz.

[0017] When the microcontroller detects the attenuation of the high-frequency reflected wave caused by surface tension, it determines that there is interference from water droplets on the surface and adopts the standing wave shear control mode. The microcontroller tracks the impedance phase angle change rate Δθ' / Δt of the ultrasonic transmitter's drive circuit in real time, causing Δθ' to suddenly increase the resonant frequency by 30%. At this frequency, the microcontroller controls the ultrasonic transmitter to apply 3.0μm standing wave vibration, forming a periodic shear stress field. The imbalance of liquid surface tension causes the water film to break into droplets and roll off.

[0018] When the microcontroller detects an increase in the low-frequency resonance peak caused by viscosity, it determines that it is an oil-related interference. Using a low-frequency mechanical shock and thermal-assisted mode, the ultrasonic transmitter initially vibrates at a low frequency of 40kHz and a large amplitude to generate an impact acceleration exceeding 500g. At the same time, the piezoelectric inverse effect is used to generate Joule heat to reduce the viscosity of the oil. The microcontroller controls the ultrasonic transmitter to decrease the frequency in steps of 1kHz until the residual oil is controlled below 2%.

[0019] Furthermore, the acousto-optic combined positioning system is placed in a protective shell, which adopts a double-layer composite structure, including a gradient functional layer and a fractal resonance layer, and the fractal resonance layer is located inside the gradient functional layer; the outer surface of the gradient functional layer is engraved with a Koch snowflake fractal groove, whose fractal dimension is 1.26-1.58, the number of levels is 5, and the groove depth is 0.2-0.5mm; the overall depth of the groove is gradient distributed, decreasing from the center to the edge at a gradient of 0.05mm / mm, so that the extension direction of the groove is at an angle of 10°-20° with the incident direction of the sound wave; a microhole with a diameter of 20μm is provided at the bottom of the groove, and the depth of the microhole is 1 / 5 of the groove depth.

[0020] Furthermore, a filter cover is provided on the outside of the ultrasonic receiver, and the outer surface of the filter cover is provided with a leading edge serration array made of ceramic. The leading edge serration array is designed by integrating bionics and quantum biotechnology: the leading tooth inclination angle of the serrations in the leading edge serration array is 45°, the rear tooth inclination angle is 30°, the root fillet radius of the serrations is ≤0.05mm, and a quantum biofilm is coated on the surface of the serrations. The biofilm is composed of a spider silk protein network synthesized by gene editing and a cadmium sulfide quantum dot array.

[0021] Furthermore, the LED array performs frequency modulation on the light signal, specifically:

[0022] Design a nonlinear energy efficiency ratio evaluation function:

[0023]

[0024] Among them, the intermediate quantity s(t)=[SNR(t),e dec (t)], SNR(t) is the signal-to-noise ratio of the average power of the sound wave signal received by the ultrasonic receiver (2) at time t to the noise power, e dec (t) is the decoding error rate at time t, ΔSNR(t) = SNR(t) - SNR ref , SNR ref is the reference signal-to-noise ratio, K1 is the smoothing coefficient, λ is a constant, (1-f / f max ) l is a concave function, f max is the upper limit of the sending frequency, l is the adjustment attenuation rate;

[0025] When SNR<20dB and e dec When >0.05, the main constraint of the evaluation function is the error term e -λedec(t) ; When SNR>30dB and e dec <0.01, the frequency term (1-f / f max ) l Become the main constraint;

[0026] Energy consumption hierarchical control is achieved through the parameter l: when l = 2, it is superlinear penalty, when l = 0.5, it is sublinear attenuation, and when l = 1, it is a linear energy consumption model.

[0027] Furthermore, the ultrasonic transmitter performs pulse modulation on the ultrasonic signal so that the ultrasonic emission power satisfies:

[0028]

[0029] Among them, P max and P min are the maximum power and minimum power of ultrasonic emission respectively, E is the environmental index, and:

[0030]

[0031] Wherein, L is the ambient light intensity measured during the calculation of the LED array transmission frequency, SNR is the signal-to-noise ratio measured during the calculation of the LED array transmission frequency, T is the ambient temperature measured by the temperature sensor, and H is the humidity measured by the humidity sensor.

[0032] Furthermore, the position of the target point in three-dimensional space is:

[0033] x=x i +R·cos(θ correct,y )·sin(θ correct,x )

[0034] y=y i +R·cos(θ correct,y )·cos(θ correct,x )

[0035] z=z i +R·sin(θ correct,y )

[0036] Among them, x i 、y i 、z i is the spatial three-dimensional coordinate of the i-th acousto-optic positioning system, θ correct，x and θ correct，y is the corrected arrival angle.

[0037] Furthermore, the adaptive dynamic weighted hybrid filtering algorithm includes:

[0038] (1) Dual-modal noise perception mechanism: real-time monitoring of the angular noise σ of the optical AOA θ Distance noise from ultrasonic waves (σ R ), establish the dynamic noise covariance matrix:

[0039] Q k=diag([σ θ 2 ,σ R 2 ])

[0040] in:

[0041] σ θ =α·(1+e -β·SNR )

[0042]

[0043] Where α, β, and δ are calibration coefficients, SNR is the signal-to-noise ratio, ΔT is the temperature change, ΔH is the humidity change, ΔT0 and ΔH0 are the normalized reference values of temperature and humidity, and ΔT and ΔH are the actual changes in temperature and humidity.

[0044] (2) Nonlinear geometric constraint fusion: For each acousto-optic positioning system, calculate the theoretical measurement value:

[0045]

[0046] Among them, θ ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) is the theoretical angle between the position of the positioning system and the target point, R ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) the theoretical distance between the position of the positioning system and the target point calculated;

[0047] A geometric position verification layer is introduced to construct a cost function using the physical constraint relationship between AOA angle and ultrasonic distance:

[0048] J(X)=w optical ·(θ corrected -θ ideal (X)) 2 +w ultrasonic ·(R-||XX sensor ||) 2 +ε·||XX prior || 2

[0049] Among them, ||XX sensor || is the current position estimate X and the ultrasonic receiver position X sensor The Euclidean distance between ||XX prior || 2Estimated current position X and previous position X prior The squared distance, weight w optical 、w ultrasonic Dynamic changes are made based on the data measured by the AOA angle measurement device and the ultrasonic ranging module, respectively. ε is the regularization term.

[0050] (3) Sliding window adaptive weighting:

[0051] Use time sliding window to analyze historical data consistency, weight w optical and w ultrasonic They are:

[0052]

[0053] w ultrasonic =correlation(R window ,R motion_model )

[0054] Among them, θ window 、R window are the AOA angle measurement value sequence and ultrasonic ranging value sequence of the latest N frames, R motion_model is the "expected distance" sequence estimated based on the kinematic model and the state at the previous moment. entropy() is used to calculate the discrete entropy of N groups of angle values, and correlation() is used to obtain the Pearson correlation coefficient of the two sequences. N represents the window length.

[0055] When entropy() increases, the weight w optical When <0.5, increase the weight w ultrasonic The value of w optical +w ultrasonic =1;

[0056] The adaptive dynamic weighted hybrid filtering algorithm updates the state and covariance in the following way:

[0057]

[0058] Among them, x k|k-1 is the predicted state at the current moment k, P k|k-1 is the prior estimation error covariance matrix of k at the current moment, the improved Kalman gain H i is the Jacobian matrix, R θ is the covariance matrix of the noise values measured in the horizontal direction of the ultrasonic receiver installation angle, R R is the covariance matrix of the noise values measured in the straight line between the ultrasonic receiver and the target point, and the global Kalman gain K golbal =(∑Ki )+κ·I, κ is the regularization coefficient, I is the unit matrix, z residual is the residual.

[0059] The indoor positioning method proposed in the present invention has the following advantages:

[0060] (1) Combining acoustic and optical technologies to achieve high robustness: Optical AOA angle measurement equipment is greatly affected by environmental factors such as light changes and obstacles, while ultrasonic ranging modules are less dependent on ambient light and obstructions. The combination of the two can effectively reduce the weaknesses of a single system and improve the robustness of the positioning system in complex environments. Optical AOA angle measurement equipment may be interfered with by factors such as light changes and reflections, while ultrasonic ranging modules are relatively less affected by these interferences. The combination of the two can enhance the overall anti-interference ability of the system.

[0061] (2) The LED array of the present invention utilizes an intrinsic polarization excitation mechanism, which can suppress background light interference. This eliminates the need for external optical components, streamlining the equipment while reducing equipment costs. Furthermore, the current pulse control employed by the present invention achieves nanosecond polarization switching speeds, compared to the millisecond response of liquid crystals in conventional technologies (using polarizers or liquid crystal controllers). Furthermore, while conventional polarization devices consume relatively high energy, the present invention only consumes negligible energy due to current waveform adjustment.

[0062] (3) Improved Layout: This invention breaks through the rigid constraints of the traditional rectangular grid layout. The LED array adopts a honeycomb layout. By simulating the elastic mechanical properties of the honeycomb structure, the nodes form scalable hexagonal units, achieving blind-zone coverage for optical signals. At the same time, by controlling the layer height, light utilization and coverage uniformity are optimized.

[0063] (4) Ultrasonic cleaning mechanism: Ultrasonic micro-vibration is used to actively remove dust, water droplets, and oil stains without the need for additional equipment. It achieves autonomous cleaning while continuously positioning, avoiding unnecessary equipment redundancy and cost consumption, and ensuring overall equipment efficiency. The system can also intelligently identify the type of contamination and dynamically adjust cleaning parameters to achieve efficient decontamination.

[0064] (5) Assistance of external hardware equipment: The protective shell of the present invention adopts Koch snowflake fractal grooves to improve ultrasonic absorption and noise suppression; the filter cover combines an arched reflective structure, low-impedance cork material and a bionic owl feather serrated array to effectively reduce environmental noise, optimize target sound wave transmission, and ensure measurement accuracy.

[0065] (6) Dynamic adjustment of transmission signal parameters: The present invention adjusts the transmission signal parameters according to the signal-to-noise ratio (SNR) and the bit error rate (BER). dec) optimizes the LED array's light frequency (100Hz-1kHz), balancing signal stability, anti-interference capabilities, and energy consumption. The ultrasonic component also synchronizes pulse modulation and adaptively adjusts the transmit power (10mW-100mW) based on environmental indices to ensure effective signal transmission. This method combines optical signal modulation with ultrasonic power regulation, improving system robustness and energy efficiency.

[0066] (7) Improvement of the sound speed formula: The present invention is based on the classical sound speed formula An improved sound speed calculation model is proposed to accurately describe the influence of temperature and humidity on the sound speed of air. This formula improves the accuracy of sound speed calculation and provides optimized support for ultrasonic ranging and sensing applications in complex environments such as high humidity and high temperature.

[0067] (8) The corresponding curvature compensation is calculated using the measured ultrasonic propagation path distance. The curvature correction can partially restore the signal propagation information in the blocked area, reduce the path deviation caused by signal reflection, further optimize and correct the optical AOA angle, and improve positioning accuracy.

[0068] (9) Improved data fusion algorithm: Compared with the traditional Kalman filtering algorithm, the adaptive dynamic weighted hybrid filtering algorithm proposed in the present invention can further automatically downgrade abnormal nodes and has stronger anti-occlusion capabilities. At the same time, its dynamic weighting based on data characteristics and perception of dynamic noise enable it to perform dynamic positioning more accurately and quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a block diagram of the overall implementation of the method of the present invention;

[0070] Figure 2 This is a honeycomb collaborative layout diagram of the LED array of the present invention;

[0071] Figure 3 This is a structural diagram of a protective shell engraved with a Koch snowflake fractal pattern according to the present invention;

[0072] Figure 4 A comparison chart of the sound absorption effect of the device housing when using the device housing;

[0073] FIG5( a ) is a structural diagram of an owl feather bionic filter cover according to the present invention;

[0074] FIG5( b ) is a three-dimensional model diagram of the sawtooth of the present invention;

[0075] Figure 6 This is a working principle diagram of the indoor positioning device combining optics and ultrasonic waves of the present invention;

[0076] Figure 7 Schematic diagram of the position of the optical AOA positioning and ultrasonic ranging equipment of the present invention in the reference coordinate system;

[0077] Figure 8 This is a three-dimensional spatial distribution diagram of the indoor positioning results based on AOA and ultrasound in the present invention;

[0078] Figure 9 This is a comparison of the indoor positioning errors based on pure AOA and AOA combined with ultrasound in the present invention.

[0079] Explanation of the numbers in the figure: 1. Ultrasonic transmitter; 2. Ultrasonic receiver; 3. LED array; 4. Detector array; 5. Microcontroller; 6. Protective shell; 7. Gradient functional layer; 8. Fractal resonance layer; 9. Filter cover; 10. Top of the filter cover; 11. Leading edge serration array; 12. Installation position of the ultrasonic receiver; 13. Bracket; 14. Temperature sensor; 15. Humidity sensor; 16. LED node; 17. Edge area; 18. Middle area; 19. Center area; 20. Retractable micro bracket. DETAILED DESCRIPTION

[0080] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0081] The present invention proposes an indoor positioning method based on AOA and ultrasonic assisted ranging, and its overall implementation block diagram is as follows: Figure 1 As shown, it includes the following steps:

[0082] Step 1: The ultrasonic transmitter 1, ultrasonic receiver 2, LED array 3, and detector array 4 are respectively connected to the same microcontroller 5 (i.e., MCU) with an ADC module (specifically, an STM32 single-chip microcomputer is used in this embodiment). These five devices together constitute a basic sound and light combined positioning system. Multiple sets of this system are deployed in various directions indoors, and the microcontrollers 5 of each system are then connected to the same terminal system, which records, processes, and analyzes the data obtained by all microcontrollers 5 in real time.

[0083] For the LED array 3, the present invention proposes an intrinsic polarization excitation mechanism, in which the microcontroller 5 controls the LED's intrinsic polarization state via the direction of carrier injection. This mechanism utilizes the asymmetric waveform of the LED drive current to alter the recombination direction of electron-hole pairs in the quantum well, inducing the intrinsic polarization characteristics of the spontaneously emitted light. Specifically, under a 0-5mA forward pulse (rise time 10ns), carriers preferentially recombine along the quantum well normal, generating linearly polarized light with a polarization axis parallel to the crystal lattice. Conversely, under a -2-0mA reverse pulse (fall time 50ns), the carrier distribution symmetry is disrupted, resulting in elliptically polarized light with a major axis at a 45° angle to the crystal lattice. This technology utilizes purely electrically controlled excitation of the LED's intrinsic polarization, eliminating external optical components (such as polarizers) and relying solely on algorithms to suppress background light interference.

[0084] To ensure that the LED array 3 can achieve optical signal coverage of the target area without blind spots, the present invention adopts a honeycomb layout method to break through the rigid constraints of the traditional rectangular grid. By simulating the elastic mechanical properties of the honeycomb structure, every seven LED nodes that can form a regular hexagon are grouped together. Within this group, the central LED node and the six adjacent nodes on the periphery form a retractable regular hexagonal unit. At this time, the distance between the central LED node and any of the surrounding nodes is the reference spacing; Figure 2 As described above, the area formed by the regular hexagonal units composed of the seven nodes in the middle of the array is the central area 19, the circle of regular hexagonal units located at the edge of the array forms the edge area 17, and the remaining middle part is the middle area 18. The spacing is dynamically adjusted using virtual tension feedback: the distance between the central LED node in the regular hexagonal unit located in the central area 19 and any node on the surrounding area is automatically expanded to 1.2 times the reference spacing to reduce redundant energy consumption, while the distance between the central LED node in the regular hexagonal unit in the edge area 17 and any node on the surrounding area is compressed to 80% of the reference spacing to increase density and eliminate geometric blind spots. The height of the LED array is controlled in layers by the retractable bracket 20 ( Figure 2 ), the LED nodes in the central area are raised, and the wide-angle LED light covers the ceiling. The height of the LED nodes in the middle area remains unchanged, maintaining the vertical projection of the LED light. The height of the LED nodes in the edge area is lowered, and the narrow-angle LED light is focused on the corner. The ratio of the LED node raising, the height remaining unchanged, and the height lowering is 6:5:4 (the specific height is adjusted according to the upper limit of the indoor height to be measured), forming a three-dimensional parabolic light field.

[0085] The total number of LED arrays is estimated based on the formula:

[0086]

[0087] Among them, A is the room area, A LEDis the effective coverage area of a single LED array, and η is the overlap ratio of the effective coverage areas between adjacent LED arrays, ranging from 0.1 to 0.2.

[0088] According to the total number of LED arrays and the spatial range in which they are to be arranged, the LED arrays 3 are evenly arranged in the space, and then the detector array 4, the ultrasonic transmitter 1, the ultrasonic receiver 2 and the microcontroller 5 are arranged in sequence.

[0089] The detector array 4 uses photodiodes. Due to the micro-stress of the packaging lens, the photodiodes have inherent anisotropy (polarization sensitivity ratio ≈ 1.2:1). By calibrating the polarization response matrix of each detector array, the anisotropic response is modeled:

[0090]

[0091] Among them, r i,u is the corresponding coefficient of detector i in the polarization direction of angle u.

[0092] The received signal of the photodiode is:

[0093]

[0094] Among them, P i is the polarization state of the i-th LED in the LED array corresponding to the detector array, and M is the external noise interference.

[0095] For this photodiode, the OMP (Orthogonal Matching Pursuit) algorithm is used to invert the original polarization encoding from the multi-detector data, suppress the noise M, and improve the signal-to-noise ratio of the photodiode by more than 20dB.

[0096] The ultrasonic transmitter 1 and ultrasonic receiver 2 together form the ultrasonic ranging module. The ultrasonic receiver is equipped with a highly sensitive ultrasonic sensor with a sensitivity better than -80dB, enabling precise detection of echo signals. Each ultrasonic ranging module maintains a consistent signal propagation path with its LED array 3 within the same acousto-optic positioning system, enabling a more compact design by sharing a microcontroller 5 and computing resources. After installation, the ultrasonic propagation delay and power distribution are calibrated using test signals to ensure that the signal coverage area matches that of the optical signal.

[0097] In addition, the ultrasonic transmitter 1 is also equipped with a temperature sensor 14 and a humidity sensor 15 to transmit data in real time to dynamically sense changes in the external environment.

[0098] The present invention improves the algorithm of the aforementioned ultrasonic ranging module and designs an intelligent micro-vibration cleaning mechanism. This mechanism, based on the principle of an ultrasonic transmitter, utilizes the micro-vibrations generated by the ultrasonic transmitter 1 to remove dirt and other interfering substances from the surface of the ultrasonic sensor. When an interfering object appears, the microcontroller 5 connected to the ultrasonic ranging module first detects the type of interfering object based on the characteristics of the currently received reflected wave. Under normal indoor conditions, interfering objects that may occur are mainly categorized as dust, water droplets, and oil stains. When the ultrasonic receiver 2 detects an increase in broadband noise caused by particle scattering, the microcontroller 5 determines it as dust interference and transmits a signal to the ultrasonic transmitter 1. The ultrasonic transmitter 1 adopts a high-frequency pulse blasting mode, initially exciting micron-sized cavitation bubbles with a high-frequency vibration of 150kHz (amplitude 1.0μm), and using the shock wave (>10MPa) of the bubble rupture to peel off the particles; the effective frequency band is expanded by 5kHz every 10ms. At the same time, the ultrasonic receiver 2 dynamically monitors the reflected wave signal-to-noise ratio (SNR) (transmitting data to the microcontroller 5 every 100ms). If the reflected wave SNR is lower than 15dB, the microcontroller 5 triggers the ultrasonic transmitter 1 to emit a 5ms short pulse (1.5μm) to enhance the energy density and gradually reduce the frequency to 120kHz to cover a larger polluted area. When the attenuation of the high-frequency reflected wave caused by surface tension is detected, the microcontroller 5 determines that there is water droplet interference on the surface and adopts the standing wave shear control mode. The microcontroller 5 quickly sweeps the frequency within the range of 80-100kHz, and accurately locks the resonant frequency (usually 85-95kHz) that makes Δθ' suddenly increase by 30% by tracking the impedance phase angle change rate Δθ' / Δt of the driving circuit of the ultrasonic transmitter 1 in real time (that is, the rate of change of the electrical impedance phase angle θ with time when the ultrasonic transmitter 1 is working). At this frequency, the microcontroller 5 controls the ultrasonic transmitter 1 to apply 3.0μm standing wave vibration, forming a periodic shear stress field (>0.2N / m 2 ), causing the liquid's surface tension to unbalance, breaking the water film into droplets. When the microcontroller 5 detects an increase in the low-frequency resonance peak caused by viscosity, it identifies it as an oil-related disturbance. Using a low-frequency mechanical shock and thermal-assisted mode, the ultrasonic transmitter 1 initially vibrates at a low frequency (5μm) of 40kHz, generating an impact acceleration exceeding 500g. This, combined with the piezoelectric inverse effect, automatically generates Joule heating to reduce the oil's viscosity. The microcontroller 5 controls the ultrasonic transmitter 1 to decrease its frequency in 1kHz steps. Based on the changes in the low-frequency resonance peak, the microcontroller dynamically adjusts the amplitude to control the oil residue below 2% within 200ms. The entire cleaning process utilizes the STM32 microcontroller's TIM1 advanced timer to achieve nanosecond-level switching of the 20-150kHz PWM frequency, combined with a DAC to output a 12-bit precision amplitude control signal.

[0099] At the same time, in order to further enhance the overall intelligence and adaptability of the ultrasonic ranging module, the present invention integrates intelligent environmental perception technology on it. The data of the humidity sensor installed on the ultrasonic transmitter 1 is collected in real time through the built-in ADC module of the STM32 microcontroller, and the hyperbolic tangent humidity response function is constructed:

[0100] RH_threshold=75%+10%·tanh(0.1·(dRH / dt))

[0101] When the ambient humidity RH>RH_threshold and the humidity change rate dRH / dt>5% / min, it indicates that the current humidity environment may cause the risk of condensation of water vapor on the surface of the humidity sensor 14. The microcontroller 5 will activate the standing wave shear mode of the intelligent micro-vibration cleaning mechanism in advance and start preventive cleaning with parameters 90kHz / 3.0μm. The running time of the standing wave shear mode is earlier than the actual water droplet formation time, which can effectively curb the water droplet interference naturally formed by external humidity changes.

[0102] The 12-bit ADC module built into the STM32 microcontroller collects data from the temperature sensor installed on the ultrasonic transmitter 1 in real time. When the external temperature T>60°C is detected, the microcontroller 5 compresses the working cycle of the ultrasonic transmitter 1 and switches the continuous vibration mode of the ultrasonic transmitter 1 to the pulse mode. The duty cycle of the ultrasonic transmitter 1 pulse signal is adjusted according to the gradient:

[0103] D(t)=D0·(1-0.02·∫(T(t)-60)dt)

[0104] Where D0 is the initial duty cycle, T(t) is the time-varying trend of the current temperature, and the duty cycle D(t) of the pulse signal is set to a minimum of 30%. In addition, the frequency and temperature are coupled and adjusted, and the base frequency is automatically lowered by 3kHz for every 5°C increase in temperature to reduce the heat generated by the piezoelectric element and achieve a cooling effect.

[0105] like Figure 3 As shown, the present invention designs a protective shell 6 with acoustic adaptation characteristics on the outside of the acousto-optic combined positioning system. The shell adopts a double-layer composite structure, including a gradient functional layer 7 and a fractal resonance layer 8. The fractal resonance layer 8 is located inside the gradient functional layer 7.

[0106] The gradient functional layer 7 is made of an EPDM rubber matrix, which has excellent sound insulation properties and can shield high-frequency, medium- and low-frequency noise. Its exterior is designed with a black matte coating, which effectively absorbs and isolates external light interference. In addition, the present invention uses a laser to engrave Koch snowflake fractal grooves on the outer surface of the gradient functional layer 7. The fractal dimension is 1.26-1.58, the number of levels is 5, the groove depth is 0.2-0.5mm, and the groove density is 30 grooves / cm 2 The overall depth of the groove is distributed in a gradient, decreasing from the center to the edge at a gradient of 0.05 mm / mm, so that the extension direction of the groove forms an angle of 10°-20° with the incident direction of the sound wave; a microhole with a diameter of 20 μm is provided at the bottom of the groove, and the depth of the microhole is 1 / 5 of the groove depth.

[0107] The complexity of the Koch snowflake fractal pattern means it effectively resonates or scatters sound waves across a wide range of frequencies. Furthermore, due to the highly non-uniform nature of the fractal geometry, it produces varying degrees of sound absorption in different directions. This ensures that sound waves are highly absorbed as they propagate from multiple directions, thereby enhancing the overall sound absorption effect. Figure 4 The comparison of the sound absorption effect of the protective shell 6 engraved with the Koch snowflake fractal pattern at different frequencies and when not in use shows that the sound absorption effect of this design is significantly enhanced.

[0108] The primary material of the fractal resonance layer 8 is matte, open-pore, sound-absorbing foam. Its surface also features a matte black coating, which absorbs interfering light and prevents reflection. It also provides excellent internal acoustic transparency, preventing ultrasonic signal attenuation. The pore absorption further reduces external noise from entering the device, minimizing interference. Furthermore, the pore surface features a tapered microporous expansion structure (entry aperture 0.5mm, terminal aperture 0.1mm), ensuring a sound absorption coefficient of ≥0.65 from 500Hz to 20kHz.

[0109] The protective shell 6 is fixed by the bracket 13 ( Figure 6 ), the materials of the inner and outer layers have certain elasticity and impact resistance, which can protect the internal hardware equipment and reduce the probability of equipment damage caused by falling from the air, external impact, etc.

[0110] The measuring angle of the ultrasonic wave sent by the ultrasonic transmitter 1 is 15°, so the overall design of the protective housing 6 is 15° outward to maximize the measuring range and avoid internal interference caused by ultrasonic wave reflection as much as possible.

[0111] The present invention also incorporates a filter cover 9 for the ultrasonic receiver 2. As shown in FIG5(a), the filter cover 9 has an overall arched structure, which can reflect sound waves from non-target directions to a certain extent. Furthermore, the filter cover top 10 is made of cork, which has a low acoustic impedance and effectively conducts sound waves reflected from the target direction. It also reduces noise transmission, ensuring that the target sound waves reach the ultrasonic receiver's mounting position 12 within the filter cover 9 in a timely and accurate manner. In Figure 5(b), the leading edge serration array 11 made of ceramic and arranged on the outer surface of the filter cover 9 is designed by integrating bionics and quantum biotechnology; the tooth height of the serrations in the leading edge serration array 11 is 0.3mm, the tooth pitch is 0.8mm, the front tooth inclination angle of the serrations is 45°, the rear tooth inclination angle is 30°, and the root fillet radius of the serrations is ≤0.05mm, which can suppress the generation of secondary vortices; the surface of the serrations is coated with a quantum biofilm, which is composed of a genetically edited synthetic spider silk protein (MaSp2) network and a cadmium sulfide (CdS) quantum dot array. The quantum dots absorb ambient light energy (wavelength 400-800nm) to excite the local electromagnetic field, triggering the dynamic folding and reconstruction of the spider silk protein chain, so that the leading edge serration array 11 imitating owl feathers has nanoscale deformation ability: the serrations can change the angle between them and the surface of the filter cover 9 according to the airflow velocity to decompose the initial vortex. The self-healing properties of the quantum biofilm (fracture recovery rate > 95%) ensure the long-term stability of the serrated structure under the impact of high-speed airflow.

[0112] Step 2: LED array 3 frequency-modulates the optical signal so that detector array 4 can identify and decode it. Microcontroller 5 controls LED array 3 to select an appropriate transmission frequency based on ambient lighting conditions and the detector's response speed, ensuring signal stability and reliability. Simultaneously, ultrasonic transmitter 1 pulse-modulates the ultrasonic signal, enabling ultrasonic receiver 2 to accurately measure the signal's propagation time.

[0113] In order to improve the robustness and signal transmission quality of the system while minimizing energy consumption, the present invention proposes a dynamic frequency adaptive modulation method: First, the digital signal collected by the ADC module of the microcontroller 5 is used to calculate the average power and noise power of the signal, thereby obtaining its signal-to-noise ratio, i.e., the SNR value; at the same time, with the help of the photodiodes of the detector array 4, during the signal decoding process, the decoding error ratio is calculated as the decoding error rate, i.e., e dec Based on the above two basic parameters, a nonlinear energy efficiency ratio evaluation function is designed, and its mathematical expression is:

[0114]

[0115] Where s(t)=[SNR(t),e dec (t)]; ΔSNR(t) = SNR(t) - SNRref , SNR ref The reference signal-to-noise ratio is the historical sliding window mean; K1 is the smoothing coefficient, which is initially set to 1. Due to the large fluctuation of ΔSNR(t), K1 is set to the value of the fluctuation amplitude; K1 and ΔSNR(t) are added as the denominator to suppress the oscillation caused by the sudden change of SNR; λ is used to control the sensitivity of the bit error rate, which is initially 2. When e dec When λ > 0.1, the value of λ decays rapidly to zero, forcing the system to prioritize communication reliability. max ) l , when the LED array's transmission frequency approaches the upper limit f max When =1000Hz, the concave function increases sharply, guiding the model to balance between energy consumption and resolution, and l is the adjustment attenuation rate.

[0116] When SNR<20dB and e dec When the value of the evaluation function is greater than 0.05, the error term e -λedec(t) Dominant, drive system frequency reduction to improve robustness; when SNR>30dB and e dec <0.01, the frequency term (1-f / f max ) l It becomes the main constraint, motivating the system to increase frequency and optimize positioning accuracy;

[0117] Energy consumption is controlled by the parameter l: when l = 2, it is a super-linear penalty, which suppresses high-frequency heating; when l = 0.5, it is a sub-linear attenuation, which allows short-term high-frequency bursts; when l = 1, it is a linear energy consumption model, which provides stable low-power power supply.

[0118] According to the above formula, when the transmission frequency of the LED array satisfies SNR stability, low bit error rate and energy efficiency optimization at the same time, it can be used as a suitable transmission frequency.

[0119] In addition, before each light signal is emitted by the LED array 3, the photodiode first collects the photocurrent signal. The operational amplifier within the microcontroller 5 performs current-to-voltage conversion and signal amplification, and then digitizes the signal using the ADC module to obtain high-precision light intensity data. The data collected by each microcontroller 5 is then transmitted to the terminal for comprehensive calculation and analysis. The terminal estimates the average and peak characteristic parameters of the ambient light intensity. Within 60 seconds, an average light intensity of ≤100 lux is considered low light, 100–2000 lux is considered medium light, and 2000 lux or above or an instantaneous peak value of >3000 lux (corresponding to direct sunlight at noon in summer) is considered high light. This is used to determine the current ambient light level. However, to account for the ±5% measurement error of the sensor, a 10 lux hysteresis interval is implemented in practice (for example, to determine the transition from low light to medium light, the light intensity must reach 110 lux). Dynamically adjust the LED drive current: When the ambient light is strong (high light), the LED luminous intensity is increased to improve signal recognition ability and prevent background light interference; when the light is weak (low light), the LED luminous intensity is reduced to prevent energy waste and equipment overheating, thereby achieving efficient, stable and precise light intensity control of the system.

[0120] While the LED array 3 transmits the light signal, the ultrasonic transmitter 1 pulse-modulates the ultrasonic signal so that the ultrasonic receiver 2 can accurately measure the signal's propagation time. The ultrasonic transmission power has a significant impact on ultrasonic transmission; typically, it should be between 10mW and 100mW. Based on the measurement environment, the present invention constructs an environmental index to dynamically adjust the ultrasonic transmission power, ensuring effective signal propagation.

[0121]

[0122] Wherein, E is the environmental index to be constructed, L and SNR are respectively the ambient light intensity and signal-to-noise ratio measured in the above-mentioned LED array transmission frequency calculation process, T and H are respectively the ambient temperature measured by the temperature sensor 14 and the humidity measured by the humidity sensor 15. Used to highlight the impact of high light intensity, ln(1+T+H) maps the sum of temperature and humidity to a smoother curve, and finally takes the average to ensure the balanced contribution of each parameter; P max and P min The maximum power of ultrasonic emission is 100mW and the minimum power is 10mW. The whole formula ensures that when the environmental index is low, the power is small, and when the environmental index increases, the power rises rapidly, but it will never exceed P maxLower power is suitable for short-range measurements in low-noise environments, such as small indoor spaces like kitchens and bathrooms, while higher power is suitable for longer distances or complex environments, such as larger indoor spaces like basketball courts. Different targets have different environments, and the corresponding environmental index varies. Choosing the appropriate power ensures signal effectiveness.

[0123] Step 3: The detector array 4 receives the light signal from the LED array 3 and records the angle of arrival of the light signal when it reaches the detector array 4; the ultrasonic receiver 2 receives the ultrasonic signal and determines the distance between the target and the ultrasonic receiver 2 by measuring the time difference of signal propagation; and determines the position of the target point in three-dimensional space.

[0124] like Figure 7 As shown, D is the target point, A, B, and C are detectors, where B and C are adjacent to A in the horizontal direction and vertical direction respectively, E and F are the projection points of the target point D on the xoy plane and the z axis respectively, θ x and θ y are the horizontal and vertical deflection angles of the light signal sent by the LED array reflected by the target point D to reach the detector array, G is the foot of the perpendicular of C on AE, and H is the foot of the perpendicular of B on AF. The length of line segment AC is set to d x , line segment AB is set to d y The length of AG is the horizontal path difference from D to points A and C (i.e. the difference between AE and GE), which is Δd in the figure. x The length of AH is the vertical path difference between D and points A and B (the difference between AF and HF), which is Δd in the figure. y . Find d x , Δd x d y , Δd y , put it into the formula, and we can get the horizontal angle θ x and vertical angle θ y .

[0125] Phase difference Δd between two adjacent detectors A and C x It can be expressed by geometric relations as:

[0126] Δd x =d x ·sin(θ x )

[0127] The phase difference Δd between two adjacent detectors A and B y It can be expressed by geometric relations as:

[0128] Δd y =d y ·sin(θ y)

[0129] According to the above two sets of formulas, the path difference (Δd x and Δd y ) reversely to obtain θ x and θ y The formula can be expressed as:

[0130]

[0131] Ultrasonic receiver 2 receives the echo signal and calculates the time difference of signal propagation. Assuming the distance between the target point and ultrasonic receiver 2 is R, the distance calculation formula is:

[0132]

[0133] Among them, t ultrasound is the time difference between the transmitted and received signals, v sound The speed of ultrasonic waves in air is usually about 340 m / s. However, considering the influence of temperature and humidity in the air on the speed of ultrasonic waves, a specific algorithm formula can be used to compensate for it. This paper proposes a sound speed compensation algorithm for temperature and humidity. The improved sound speed formula is as follows:

[0134]

[0135] Wherein, T is the ambient temperature (°C) and H is the relative humidity (%RH).

[0136] In this formula, 331.45m / s is the theoretical value of the speed of sound under standard conditions (0℃, dry air). The improved speed of sound formula is Convert ambient temperature from Celsius to Kelvin (T K =T+273.15), accurately describing the effect of temperature on the thermal motion of air molecules. This form is derived from the classical sound speed formula (γ is the adiabatic index, R is the gas constant, and M is the molar mass), which is simplified into an engineering practical form through parameter fitting. The 0.606T in the main humidity term is a linear term, which describes the positive enhancement effect of humidity on the speed of sound; the coefficient 0.606 is obtained by fitting experimental data, reflecting the gain in the speed of sound per unit temperature change (1% RH). 0.0124H characterizes the sound wave enhancement effect under high humidity. The molar mass of water molecules (H2O) (18 g / mol) is lower than the average molar mass of dry air (about 28.97 g / mol). An increase in the proportion of water vapor in the air will reduce the gas density, thereby increasing the speed of sound; 1+0.00309T is the temperature-humidity coupling term. Due to the intensified thermal motion of water molecules at high temperatures, the molecular kinetic energy increases, which enhances its modulation effect on the speed of sound; at the same time, when the temperature rises, the maximum amount of water vapor that can be accommodated in the air (saturated humidity) increases exponentially, and the absolute humidity (g / m 3 ) is higher, so the effect of humidity on the speed of sound is amplified nonlinearly with increasing temperature; the slope of 0.00309 indicates that the contribution of the humidity term increases by about 0.309% for every 1°C increase in temperature.

[0137] The indoor environment is considered as a local non-Euclidean space. Based on the distance R, the curvature K is calculated in the microcontroller 5 using a two-dimensional projection:

[0138]

[0139] Where, ΔR ideal is the ideal straight-line distance of the ultrasonic path (i.e., the path length directly measured ignoring external environmental interference), and L' is the environmental characteristic scale (the indoor length in the horizontal direction of the ultrasonic wave measured by the ultrasonic ranging module).

[0140] The terminal collects the actual ultrasonic path length ΔR in different directions measured by each ultrasonic ranging module i , environmental characteristic scale L' i And the ideal straight line distance ΔR in the direction of the ultrasonic path ideal,i , construct the path bending data set, and use the least squares method to get the formula:

[0141]

[0142] Substitute the curvature K obtained by each ultrasonic ranging module into the above formula respectively, and find the K value that minimizes the result of the above formula. This is the optimal global curvature K' that can best explain the curvature of the path in all directions.

[0143] According to Fermat's principle, the bending of the optical signal path causes the AOA measurement deviation Δθ, which is related to the optimal global curvature K' as follows:

[0144]

[0145] Where dI is a tiny length element on the ultrasonic path, and the integral represents the effect of the cumulative curvature along the entire ultrasonic propagation path.

[0146] Substituting the optimal global curvature K' into the above formula, the corrected AOA angle θ can be obtained correct，x and θ correct，y for:

[0147] θ correct,x =θ x +Δθ

[0148] θ correct,y =θ y +Δθ

[0149] It is known that the spatial three-dimensional coordinates of the i-th acousto-optic positioning system are (x i ,y i ,z i ), then combine the corrected AOA angle and distance R, and use geometric principles to calculate the position (x, y, z) of the target point in three-dimensional space:

[0150] x=x i +R·cos(θ correct,y )·sin(θ correct,x )

[0151] y=y i +R·cos(θ correct,y )·cos(θ correct,x )

[0152] z=z i +R·sin(θ correct,y )

[0153] The above formula combines the advantages of optical and ultrasonic information to quickly determine the three-dimensional position of the target point. However, actual measurements are still inevitably affected by certain noise, systematic errors, and environmental influences, so filtering algorithms are needed to further optimize the measurement results.

[0154] Step 4: Kalman filtering is a linear minimum mean square error estimation method suitable for state estimation problems in dynamic systems. However, the performance of traditional Kalman filtering is limited in nonlinear or non-Gaussian noise scenarios. This paper proposes an adaptive dynamic weighted hybrid filtering algorithm designed specifically for multimodal sensors for fusion processing, which can significantly reduce the impact of noise on measurement results and improve positioning accuracy. The content of the adaptive dynamic weighted hybrid filtering algorithm includes:

[0155] (1) Dual-mode noise sensing mechanism: real-time monitoring of the angular noise of optical AOA (σ θ ) and ultrasonic distance noise (σR ), establish the dynamic noise covariance matrix:

[0156] Q k =diag([σ θ 2 ,σ R 2 ])

[0157] in:

[0158] σ θ =α·(1+e -β·SNR )

[0159]

[0160] Among them, α, β, and δ are calibration coefficients; SNR is the signal-to-noise ratio, ΔT is the temperature change, ΔH is the humidity change, and ΔT0 and ΔH0 are normalized reference quantities, which are used to normalize the actual changes in temperature and humidity ΔT and ΔH to a dimensionless scale. The specific values are determined according to the indoor temperature and humidity fluctuation data measured on site (for example, if the current indoor temperature fluctuation is generally within ±5°C, then ΔT0 = 5°C, and the same applies to ΔH0).

[0161] (2) Nonlinear geometric constraint fusion: For each acousto-optic positioning system, calculate the theoretical measurement value:

[0162]

[0163] Among them, θ ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) is the theoretical angle between the position of the positioning system and the target point, R ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) is the theoretical distance between the positioning system position and the target point calculated.

[0164] A geometric position verification layer is introduced to construct a cost function using the physical constraint relationship between AOA angle and ultrasonic distance:

[0165] J(X)=w optical ·(θ corrected -θ ideal (X)) 2 +w ultrasonic ·(R-||XX sensor ||) 2 +ε·||XX prior || 2

[0166] Among them, ||XX sensor || is the current position estimate X and the ultrasonic receiver position X sensor The Euclidean distance between ||XX prior || 2 Estimated current position X and previous position X prior The squared distance, weight w optical 、w ultrasonic The regularization term ε is used to prevent sudden changes in the target position (1≤ε≤5 is used in high-noise environments to achieve strong regularization to suppress sudden changes, and 0.1≤ε≤1 is used in low-noise environments to achieve weak regularization and give priority to responding to measurement data).

[0167] (3) Sliding window adaptive weighting:

[0168] Using a time sliding window (length N = 5) to analyze the consistency of historical data, then w optical and w ultrasonic The two weights are:

[0169]

[0170] w ultrasonic =correlation(R window ,R motion_model )

[0171] Among them, θ window 、R window are the AOA angle measurement value sequence and ultrasonic ranging value sequence of the latest N frames, R motion_model It is the "expected distance" sequence estimated based on the kinematic model and the state at the previous moment (that is, using the optimal state at the previous moment + discrete constant speed model to push forward N steps). entropy() is used to calculate the discrete entropy of N groups of angle values, and correlation() is used to obtain the Pearson correlation coefficient of the two sequences.

[0172] When entropy() increases, the weight w optical <0.5, the Kalman gain should be based on the ultrasonic ranging data, and the microcontroller 5 increases w ultrasonic The value of w optical +w ultrasonic =1.

[0173] When implementing the adaptive dynamic weighted hybrid filtering algorithm, the initialization is first performed to calibrate the initial position of each positioning system (x s ,y s ,zs ), collect environmental baseline parameters (ambient temperature T, ambient humidity H, ambient light intensity L), initialize the sliding window; then perform state prediction:

[0174] x k|k-1 =F k ·x k-1|k-1 +B k ·u k

[0175] P k|k-1 =F k ·P k-1|k-1 ·F k T +Q k

[0176] Among them, x k|k-1 is the predicted state at the current moment k (based on the estimate and control input at the previous moment); F k is the state transition matrix, which describes the state transition rule from time k-1 to time k; x k-1|k-1 is the state estimate at the previous moment k-1; u k is the control input (i.e., the speed of the target point, which is calculated from the distance the target point moves from the previous moment to the current moment); B k is the control input matrix, which is used to convert the control input u k Mapping to state space; P k|k-1 is the prior estimation error covariance matrix of the current time k, which represents the uncertainty of the current state estimation based on the information at time k-1, P k-1|k-1 is the posterior estimation error covariance matrix at time k-1, which represents the uncertainty of the state estimation after the k-1 measurement update, Q k is the process noise covariance matrix, which describes the noise characteristics in the motion model and is used to quantify the uncertainty of the system model.

[0177] Further multi-source data fusion is performed, that is, weight analysis is performed using the cost function, and the Jacobian matrix H is obtained. i (θ ideal and R ideal The matrix of partial derivatives of the position coordinate parameters x, y, and z), combined with the adaptive weight w optical 、w ultrasonic , thus obtaining the improved Kalman gain:

[0178]

[0179] Among them, R θ is the covariance matrix of the noise values measured in the horizontal direction of the ultrasonic receiver installation angle, R RIt is the covariance matrix of the noise values measured in the straight line direction between the ultrasonic receiver and the target point.

[0180] Then the global Kalman gain is obtained:

[0181] K global =(∑K i )+κ·I

[0182] Among them, κ is the regularization coefficient to prevent the matrix from being singular (the initial value of κ is set to 10 -4 If the measured noise is large, making SNR < 0.5 or the data dispersion is high, making w optical <0.5, can be increased to 10 -1 If the noise is too small, so that SNR>0.9, it can be reduced to 10 -6 , to retain more measurement information); I is the identity matrix, which is used to ensure numerical stability.

[0183] Update the state and covariance as follows:

[0184]

[0185] Among them, the residual z residual (i.e. the position obtained by measuring, calculating and optimizing the acoustic-optical combined positioning system and the theoretical estimated position x k|k-1 The difference) is calculated before the update.

[0186] After the hybrid filtering algorithm is optimized, the spatial coordinate point x is obtained k|k .

[0187] Compared with the traditional Kalman filter algorithm, the adaptive dynamic weighted hybrid filter algorithm achieves the best optical angle noise σ θ and ultrasonic distance noise σ R Real-time monitoring; the cost function J(x) is constructed, combining optical angle error, ultrasonic distance error, and regularization terms to achieve collaborative optimization of multi-source data. Based on these methods, the improved filtering algorithm significantly improves noise modeling accuracy in dynamic environments and avoids estimation bias caused by fixed noise parameters.

[0188] like Figure 8 As shown in the figure, a MATLAB simulation shows the three-dimensional spatial distribution of indoor positioning results based on AOA (angle of arrival) and ultrasonic waves. The spatial coordinates obtained after optimization using the hybrid filtering algorithm are significantly closer to the target point than those calculated directly. This shows that the hybrid filtering algorithm not only enhances the system's anti-interference capabilities but also effectively improves the accuracy and robustness of the entire positioning method. Especially in complex environments, leveraging the complementary advantages of optical and ultrasonic data can significantly reduce the errors caused by a single positioning device.

[0189] Step 5: The spatial coordinate point x after the above hybrid filtering algorithm is optimized k|k Conduct positioning tests to verify its accuracy. Analysis of test results includes:

[0190] (1) Error calculation: In order to evaluate the positioning accuracy of the entire positioning method, it is first necessary to calculate the error between each positioning result and the actual position. The error calculation formula is:

[0191]

[0192] Among them, x actual 、y actual 、z actual is the real position coordinate, x estimated 、y estimated 、z estimated is the positioning coordinate after optimization by hybrid filtering algorithm.

[0193] (2) Performance evaluation: By statistically analyzing the error data from multiple measurements, the positioning performance of the system under different environmental conditions can be evaluated and its applicability can be verified. The main performance evaluation indicators include average error, error variance, and maximum error. The average error represents the overall accuracy of the system and is calculated as the average of all measurement errors. The error variance reflects the degree of dispersion of the system error distribution. The smaller the variance, the more stable the system. The maximum error indicates the maximum deviation in the system positioning results and is used to evaluate the performance in the worst-case scenario.

[0194] To further verify the advantages of the system, a comparative experiment was designed. Under controlled variables and in the same ambient light, temperature and other factors as possible, the coordinates obtained by using the optical AOA positioning method alone and the optical AOA combined with ultrasonic sound and light were measured multiple times, and the error values of the two were calculated and summarized to obtain a discount comparison chart. Figure 9 , compared with the statistical data, the error of the sound and light combined measurement results is significantly smaller than the results measured using the optical geometry method alone in most cases. This shows that in complex environments, optical AOA positioning alone is limited by the obstruction of light signals and environmental noise, resulting in large positioning errors. In addition, in complex environments, the error distribution of the optical AOA method has higher discreteness and poor system stability. By utilizing the complementary advantages of optical and ultrasonic data, the errors caused by environmental factors can be significantly reduced. In summary, the present invention is far superior to traditional optical geometry positioning methods in indoor positioning.

[0195] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An indoor positioning method based on optical AOA and ultrasonic assisted ranging, characterized by: Multiple groups of acoustic and optical positioning systems are deployed at various directions indoors, and the microcontrollers in each acoustic and optical positioning system are connected to the same terminal; the acoustic and optical positioning system comprises an ultrasonic transmitter (1), an ultrasonic receiver (2), an LED array (3), a detector array (4), and a microcontroller (5); The LED array (3) performs frequency modulation on the optical signal so that the detector array (4) can recognize and decode it. At the same time, the ultrasonic transmitter (1) performs pulse modulation on the ultrasonic signal so that the ultrasonic receiver (2) can accurately measure the propagation time of the signal. The detector array (4) receives the light signal from the LED array (3), records the arrival angle of the light signal when it reaches the detector array (4), and corrects the arrival angle; the ultrasonic receiver (2) receives the ultrasonic signal, determines the distance R between the target point and the ultrasonic receiver (2) by measuring the time difference of signal propagation; and determines the position of the target point in three-dimensional space based on the corrected arrival angle and distance R; Based on the position of the target point in three-dimensional space, an adaptive dynamic weighted hybrid filtering algorithm is run to obtain the spatial coordinates of the target point, thereby realizing the indoor positioning of the target point.

2. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The LED array (3) adopts an intrinsic polarization excitation mechanism. Specifically, under the action of a forward pulse of 0-5 mA, carriers will preferentially recombine along the normal direction of the quantum well, generating linearly polarized light with a polarization axis parallel to the lattice; under the action of a reverse pulse of -2-0 mA, the symmetry of the carrier distribution is destroyed, thereby generating elliptically polarized light with a long axis at 45 degrees to the lattice direction.

3. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The LED array (3) adopts a honeycomb layout mode, wherein every seven LED nodes forming a regular hexagon form a group, and within the group, the central LED node and the six adjacent nodes on the periphery form a retractable regular hexagonal unit, at which time, the distance between the central LED node and any node on the periphery is the reference spacing; the area where the regular hexagonal unit formed by the seven nodes in the middle of the array is located is the central area, a circle of regular hexagonal units located at the edge of the array forms the edge area, and the rest is the middle area; the distance between the central LED node of the regular hexagonal unit in the central area and any node on the periphery is extended to 1.2 times of the reference spacing, while the distance between the central LED node of the regular hexagonal unit in the edge area and any node on the periphery is compressed to 80% of the reference spacing; the height of the LED array is controlled in layers by a retractable bracket, the LED nodes in the central area are raised, the height of the LED nodes in the middle area remains unchanged, and the height of the LED nodes in the edge area is lowered, and the ratio of the LED node raising, the height remaining unchanged, and the height lowering is 6:5:

4.

4. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The ultrasonic transmitter (1) and the ultrasonic receiver (2) constitute an ultrasonic distance measurement module, and the ultrasonic distance measurement module adopts an intelligent micro-vibration cleaning mechanism: When the ultrasonic receiver (2) detects an increase in broadband noise caused by particle scattering, the microcontroller (5) determines that it is dust interference, and the ultrasonic transmitter (1) adopts a high-frequency pulse blasting mode, initially exciting micron-sized cavitation bubbles with a high-frequency vibration of 150kHz, and using the shock wave of the bubble rupture to peel off the particles; the ultrasonic transmitter (1) increases the effective frequency band by 5kHz every 10ms, and at the same time, the ultrasonic receiver (2) dynamically monitors the reflected wave signal-to-noise ratio. If the reflected wave signal-to-noise ratio is lower than 15dB, the microcontroller (5) triggers the ultrasonic transmitter (1) to increase the energy density with a short pulse of 5ms, and the ultrasonic transmitter (1) gradually reduces the frequency to 120kHz; When the attenuation of the high-frequency reflected wave caused by the surface tension is detected, the microcontroller (5) determines that there is interference from water droplets on the surface and adopts the standing wave shear control mode. The microcontroller (5) tracks the impedance phase angle change rate Δθ' / Δt of the driving circuit of the ultrasonic transmitter (1) in real time, so that Δθ' suddenly increases by 30% of the resonant frequency. At this frequency, the microcontroller (5) controls the ultrasonic transmitter (1) to apply 3.0μm standing wave vibration, forming a periodic shear stress field, and causing the water film to break into droplets and roll down due to the imbalance of the liquid surface tension; When the microcontroller (5) detects an increase in the low-frequency resonance peak caused by viscosity, it is determined to be an oil-related interference. A low-frequency mechanical shock and heat-assisted mode is adopted. The ultrasonic transmitter (1) initially vibrates at a low frequency of 40kHz and a large amplitude to generate an impact acceleration exceeding 500g. At the same time, the piezoelectric inverse effect is used to generate Joule heat to reduce the viscosity of the oil. The microcontroller (5) controls the ultrasonic transmitter (1) to decrease the frequency in steps of 1kHz until the residual oil is controlled below 2%.

5. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The acousto-optic combined positioning system is placed in a protective shell (6), which adopts a double-layer composite structure and includes a gradient functional layer (7) and a fractal resonance layer (8), wherein the fractal resonance layer (8) is located inside the gradient functional layer (7); the outer surface of the gradient functional layer (7) is engraved with a Koch snowflake fractal groove, the fractal dimension of which is 1.26-1.58, the number of layers is 5, and the groove depth is 0.2-0.5 mm; the overall depth of the groove is gradient distributed, decreasing from the center to the edge at a gradient of 0.05 mm / mm, so that the groove extension direction and the sound wave incident direction form an angle of 10°-20°; a micropore with a diameter of 20 μm is provided at the bottom of the groove, and the micropore depth is 1 / 5 of the groove depth.

6. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The ultrasonic receiver (2) is provided with a filter cover (9) on the outside, and a leading edge sawtooth array (11) made of ceramic is provided on the outer surface of the filter cover (9). The leading edge sawtooth array (11) is designed by integrating bionics and quantum biotechnology: the leading tooth inclination angle of the sawtooth in the leading edge sawtooth array (11) is 45°, the rear tooth inclination angle is 30°, the root fillet radius of the sawtooth is ≤0.05mm, and a quantum biofilm is coated on the surface of the sawtooth, wherein the biofilm is composed of a spider silk protein network synthesized by gene editing and a cadmium sulfide quantum dot array.

7. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The LED array (3) performs frequency modulation on the light signal, specifically: Design a nonlinear energy efficiency ratio evaluation function: Among them, the intermediate quantity s(t)=[SNR(t),e dec (t)], SNR(t) is the signal-to-noise ratio of the average power of the sound wave signal received by the ultrasonic receiver (2) at time t to the noise power, e dec (t) is the decoding error rate at time t, ΔSNR(t) = SNR(t) - SNR ref , SNR ref is the reference signal-to-noise ratio, K1 is the smoothing coefficient, λ is a constant, (1-f / f max ) l is a concave function, f max is the upper limit of the sending frequency, l is the adjustment attenuation rate; When SNR<20dB and e dec When >0.05, the main constraint of the evaluation function is the error term e -λedec(t) ; When SNR>30dB and e dec <0.01, the frequency term (1-f / f max ) l Become the main constraint; Energy consumption hierarchical control is achieved through the parameter l: when l = 2, it is superlinear penalty, when l = 0.5, it is sublinear attenuation, and when l = 1, it is a linear energy consumption model.

8. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The ultrasonic transmitter (1) performs pulse modulation on the ultrasonic signal so that the ultrasonic emission power satisfies: Among them, P max and P min are the maximum power and minimum power of ultrasonic emission respectively, E is the environmental index, and: Wherein, L is the ambient light intensity measured during the calculation of the LED array transmission frequency, SNR is the signal-to-noise ratio measured during the calculation of the LED array transmission frequency, T is the ambient temperature measured by the temperature sensor, and H is the humidity measured by the humidity sensor.

9. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The position of the target point in three-dimensional space is: x=x i +R·cos(θ correct,y )·sin(θ correct,x ) y=y i +R·cos(θ correct,y )·cos(θ correct,x ) z=z i +R sin(θ correct,y ) Among them, x i 、y i 、z i is the spatial three-dimensional coordinate of the i-th acousto-optic positioning system, θ correct，x and θ correct，y is the corrected arrival angle.

10. The indoor positioning method based on optical AOA and ultrasonic assisted ranging according to claim 1, characterized in that: The adaptive dynamic weighted hybrid filtering algorithm includes: (1) Dual-modal noise perception mechanism: real-time monitoring of the angular noise σ of the optical AOA θ Distance noise from ultrasonic waves (σ R ), establish the dynamic noise covariance matrix: Q k =diag([σ θ 2 ,s R 2 ]) in: s θ =α·(1+e -β·SNR ) Where α, β, and δ are calibration coefficients, SNR is the signal-to-noise ratio, ΔT is the temperature change, ΔH is the humidity change, ΔT0 and ΔH0 are the normalized reference values of temperature and humidity, and ΔT and ΔH are the actual changes in temperature and humidity. (2) Nonlinear geometric constraint fusion: For each acousto-optic positioning system, calculate the theoretical measurement value: Among them, θ ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) is the theoretical angle between the position of the positioning system and the target point, R ideal is the coordinate of the i-th acousto-optic positioning system (x i ,y i ,z i ) the theoretical distance between the position of the positioning system and the target point calculated; A geometric position verification layer is introduced to construct a cost function using the physical constraint relationship between AOA angle and ultrasonic distance: J(X)=w optical ·(θ corrected -θ ideal (X)) 2 +w ultrasonic ·(R-||X-X sensor ||) 2 +ε·||X-X prior || 2 Among them, ||XX sensor || is the current position estimate X and the ultrasonic receiver position X sensor The Euclidean distance between ||XX prior || 2 Estimated current position X and previous position X prior The squared distance, weight w optical 、w ultrasonic Dynamic changes are made based on the data measured by the AOA angle measurement device and the ultrasonic ranging module, respectively. ε is the regularization term. (3) Sliding window adaptive weighting: Use time sliding window to analyze historical data consistency, weight w optical and w ultrasonic They are: w ultrasonic =correlation(R window ,R motion_model ) Among them, θ window 、R window are the AOA angle measurement value sequence and ultrasonic ranging value sequence of the latest N frames, R motion_model is the "expected distance" sequence estimated based on the kinematic model and the state at the previous moment. entropy() is used to calculate the discrete entropy of N groups of angle values, and correlation() is used to obtain the Pearson correlation coefficient of the two sequences. N represents the window length. When entropy() increases, the weight w optical When <0.5, increase the weight w ultrasonic The value of w optical +w ultrasonic =1; The adaptive dynamic weighted hybrid filtering algorithm updates the state and covariance in the following way: x k|k =x k|k-1 +K global -1 ·(∑K i ·z residual ) P k|k =(I-K global -1 ·∑K i )·P k|k-1 Among them, x k|k-1 is the predicted state at the current moment k, P k|k-1 is the prior estimation error covariance matrix of k at the current moment, the improved Kalman gain H i is the Jacobian matrix, R θ is the covariance matrix of the noise values measured in the horizontal direction of the ultrasonic receiver installation angle, R R is the covariance matrix of the noise values measured in the straight line between the ultrasonic receiver and the target point, and the global Kalman gain K global =(∑K i )+κ·I, κ is the regularization coefficient, I is the unit matrix, z residual is the residual.

Images