Positioning method, apparatus and electronic device

Abstract

The specification provides a positioning method, device and electronic equipment, the method comprising: collecting a sound wave positioning pulse signal emitted by a target reference device; obtaining characteristic information representing a time difference between different array elements in the acoustic array when the sound wave positioning pulse signal arrives; based on the characteristic information, the relative spatial arrival angle of the target reference device relative to the device to be positioned is calculated; obtaining the relative distance between the device to be positioned and the target reference device, and the device posture parameter of the device to be positioned; based on the device posture parameter, the relative spatial arrival angle, the relative distance and the known position of the target reference device, the spatial position of the device to be positioned is determined. The technical problem that in the prior art, the positioning system needs multiple positioning reference devices, which is difficult to arrange due to environmental restrictions, and the positioning accuracy is greatly affected by the environment.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of communication technology, and in particular to a positioning method, apparatus, and electronic device. Background Technology

[0002] With the rapid development of mobile robots, unmanned aerial vehicles, and other technologies, the demand for real-time, high-precision positioning of targets in three-dimensional space is increasing.

[0003] Among existing positioning technologies, acoustic positioning technology primarily uses multi-station spatial intersection to calculate location, which imposes significant environmental limitations and makes deployment difficult. Other methods, such as wireless ranging and radio frequency fingerprinting, currently lack the conditions for widespread adoption in terms of positioning accuracy, system cost, environmental tolerance, and ease of application.

[0004] Among existing positioning technologies, fusion positioning, including Bluetooth, AOA, UWB, LoRa, SLAM, etc., can effectively obtain relatively accurate locations in enclosed spaces. However, there are trade-offs in terms of system complexity, cost, positioning accuracy, and reliability, and there is a lack of universal positioning solutions for underground spaces. Summary of the Invention

[0005] This specification provides a positioning method, device, and electronic device to solve the technical problems in existing solutions where positioning systems are difficult to deploy due to environmental limitations and their positioning accuracy is greatly affected by the environment.

[0006] A first aspect of this specification provides a positioning method applied to a device to be positioned, the device including an acoustic array, the method comprising: Acquire the acoustic positioning pulse signal emitted by the target reference device; Acquire characteristic information representing the time difference between arrival of the acoustic positioning pulse signal to different array elements in the acoustic array; Based on the feature information, the relative spatial angle of arrival of the target reference device relative to the device to be located is calculated; The relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located, are obtained. The spatial position of the device to be located is determined based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device.

[0007] Further, acquiring the characteristic information representing the arrival time difference between different array elements in the acoustic array of the acoustic positioning pulse signal includes: The acoustic positioning pulse signals received by each element in the acoustic array are subjected to baseband demodulation and matched filtering to obtain the matched filtered output signal of each element. For each array element, the matching peak position is determined from the matched filter output signal, and the signal is truncated based on the matching peak position to obtain the direct wave signal segment corresponding to each array element. The feature information is obtained based on the direct wave signal segment corresponding to each array element.

[0008] Further, the acoustic array includes a first acoustic subarray, and the step of calculating the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information includes: The array elements in the first acoustic subarray are combined to construct multiple sets of first array element connecting arm combinations, each first array element connecting arm combination containing two non-parallel first array element connecting arms. For each combination of first array element connecting arms, based on the arrival time difference of each first array element connecting arm under the combination, the candidate arrival angle corresponding to each combination of first array element connecting arms is calculated. For each combination of first array element connecting arms, the arrival time difference of each first array element connecting arm under the combination and the spatial angle between different first array element connecting arms are used to assign confidence weights to the first array element connecting arm combination. Using the candidate arrival angle as the direction and the confidence weight as the modulus, construct a three-dimensional vector corresponding to each combination of the first array element connecting arms; The three-dimensional vectors corresponding to the connecting arms of each of the first array elements are superimposed, and the relative spatial arrival angle is determined based on the direction of the superimposed resultant vector.

[0009] Furthermore, the confidence weights are calculated based on the following formula:

[0010] in, Indicates the arm connected by the first element. Second array element connecting arm The first element of the combination is the connecting arm combination. The corresponding confidence weights; Indicates the first element connecting the arms The corresponding arrival time difference; Indicates the second array element connecting arm The corresponding arrival time difference; Indicates the first element connecting the arms Second Array Link Arm The angle between them; Indicates the first prevention and elimination of zero items; This indicates the second prevention and control item.

[0011] Furthermore, the acoustic array is a multi-scale acoustic array, comprising a first acoustic subarray and at least one second acoustic subarray; wherein the element spacing of the second acoustic subarray is greater than the element spacing of the first acoustic subarray; the method further includes: The relative spatial angle of arrival calculated based on the first acoustic subarray is used as the initial spatial angle of arrival. Based on all the elements of the first acoustic subarray and the second acoustic subarray, construct multiple second element connecting arms with different baseline lengths; Directional filtering is performed based on the initial spatial arrival angle to select candidate connecting arms from the plurality of second array element connecting arms whose spatial angle with the direction indicated by the initial spatial arrival angle is within a preset angle range; the preset angle range includes 90 degrees. Select a first preset number of first long baseline arms from the candidate connecting arms in descending order of baseline length. From the candidate connecting arms, a second preset number of second long baseline arms are selected in descending order of baseline length; the spatial angle between the second long baseline arm and the first long baseline arm is within the preset angle range; A long baseline arm combination is constructed based on the first long baseline arm and the second long baseline arm; each long baseline arm combination includes one first long baseline arm and one second long baseline arm. Based on the time difference of arrival corresponding to the long baseline arm combination, multiple candidate angles with phase ambiguity are calculated; an angle calculation window with a boundary span is set with the initial spatial angle of arrival as the center, and the target candidate angle falling within the angle calculation window is selected from the multiple candidate angles with phase ambiguity as the relative spatial angle of arrival without phase ambiguity; wherein, the boundary span of the angle calculation window is smaller than the phase ambiguity period generated by the long baseline arm combination.

[0012] Furthermore, the target reference device is one of a plurality of reference devices, and the reference device is configured to simultaneously transmit wireless broadcast signals and acoustic positioning pulse signals; The step of performing baseband demodulation and matched filtering on the acoustic positioning pulse signals received by each element in the acoustic array to obtain the matched filtered output signal of each element specifically includes: Receive the wireless broadcast signal transmitted by the target reference device, and parse the wireless broadcast signal to obtain the transmission frequency band information of the target reference device; Based on the transmission band information, a corresponding template is selected as a local reference signal from a plurality of pre-configured matching signal templates; Using the local reference signal, the acoustic positioning pulse signals received by each array element in the acoustic array are subjected to baseband demodulation and matched filtering.

[0013] Further, obtaining the relative distance between the device to be located and the target reference device includes: Receive the wireless broadcast signal transmitted by the target reference device, and use the time of receiving the wireless broadcast signal as the reference time; Based on the preset physical ranging boundary, determine the effective observation time window for the target reference device; Within the effective observation time window, feature matching is performed on the acoustic signals acquired by the acoustic array; If the time at which the matching peak position is determined to occur is within the effective observation time window, then the time at which the matching peak position is determined to occur is taken as the actual arrival time, and the relative distance is calculated based on the time difference between the actual arrival time and the reference time, combined with the ambient sound speed.

[0014] Further, the direct wave signal segment is a time-domain complex baseband signal sequence; the acquisition of the feature information based on the direct wave signal segment corresponding to each array element includes: From the direct wave signal segments corresponding to each array element, extract multiple time-domain signal sampling points containing the matching peak positions to form a two-dimensional complex matrix, which serves as the feature information; The step of calculating the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information includes: The two-dimensional complex matrix is ​​input into a pre-trained relative angle calculation network, and the relative angle calculation network is used to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located.

[0015] Further, the relative angle calculation network is a complex convolutional neural network, which includes cascaded complex feature extraction network layers and real fully connected mapping layers. The step of using the relative angle calculation network to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located includes: In the complex feature extraction network layer, the two-dimensional complex matrix is ​​subjected to multi-layer complex two-dimensional convolution operations in sequence, combined with nonlinear activation processing of complex batch normalization, to extract complex feature maps; wherein, the nonlinear activation processing is configured to perform activation calculation on the magnitude of the complex number while preserving the original phase features of the complex number; The final output of the complex feature extraction network layer is moduloed to generate real parameters, which are then compressed into a real feature vector through global average pooling. The real-valued feature vector is input into the real-valued fully connected mapping layer. After fully connected mapping and dimensionality reduction, the relative spatial angle of arrival of the target reference device relative to the device to be located is obtained.

[0016] A second aspect of this application provides a positioning device for use with a device to be positioned, the device including an acoustic array, the device comprising: The acquisition module is used to acquire the acoustic positioning pulse signal emitted by the target reference device; The first acquisition module is used to acquire feature information characterizing the arrival time difference between different array elements in the acoustic array of the acoustic positioning pulse signal; The calculation module is used to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information; The second acquisition module is used to acquire the relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located. The determination module is used to determine the spatial position of the device to be located based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device.

[0017] A third aspect of this application provides a positioning device, including the positioning apparatus described in the second aspect above.

[0018] A fourth aspect of this application also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the method described above.

[0019] A fifth aspect of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0020] As can be seen, the positioning method, device, and electronic device provided in the embodiments of this specification can achieve the following beneficial effects: Traditional acoustic positioning schemes (such as polygonal positioning) typically require at least three or four reference devices with known coordinates to calculate the three-dimensional coordinates of the device to be located. This scheme, however, uses an acoustic array to obtain the relative spatial angle of arrival and combines it with relative distance to geometrically construct a complete spatial vector. This allows the device to be located to determine its relative position by establishing a connection with only a single target reference device, significantly reducing the system's requirements for reference device deployment density and hardware deployment costs. Furthermore, in terms of positioning system infrastructure construction, it can be equivalent to optical illumination, with the target reference device acting as the light source and the illuminated area being the locatable region. This greatly reduces the difficulty of designing and constructing positioning system infrastructure in complex spaces. On the other hand, the reference devices do not require strict time alignment or complex coordination mechanisms; they only need to sense surrounding reference broadcast signals (low power consumption, effective range within approximately 20 meters) and select idle time slices (a reasonable length is 100ms) to perform wireless positioning broadcasting and ultrasonic transmission operations.

[0021] Mobile devices often undergo random rotation and tilting during movement. Since the angle of arrival output by the acoustic array is based on the device's own coordinate system, if attitude is not considered, the device's own rotation can cause a significant shift in the positioning coordinates. This solution dynamically compensates for the impact of device attitude changes on angle observation by acquiring device attitude parameters in real time and performing coordinate transformation mapping. This ensures that a stable global spatial position can be output under any device attitude, improving the system's robustness in dynamic environments.

[0022] In addition to the technical problems solved by at least one embodiment of this application, the technical features constituting the technical solution, and the beneficial effects brought about by the technical features of these technical solutions as described above, other technical problems that can be solved by at least one embodiment of this application, other technical features included in the technical solution, and the beneficial effects brought about by these technical features will be further described in detail in the specific embodiments. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A schematic flowchart of a positioning method provided in an embodiment of this specification is shown; Figure 2 This specification illustrates a flowchart of a method for calculating the time difference of arrival provided in an embodiment. Figure 3This specification illustrates a flowchart of a method for calculating the relative spatial angle of arrival, as provided in an embodiment. Figure 4 A schematic diagram of an acoustic array provided in an embodiment of this specification is shown; Figure 5 This document illustrates another flowchart for calculating the relative spatial angle of arrival, as provided in an embodiment of this specification. Figure 6 This specification illustrates a flowchart of an embodiment for obtaining a matched filter output signal. Figure 7 This invention provides a schematic diagram of a process for obtaining relative distance according to an embodiment of the present invention. Figure 8 A schematic diagram of a positioning system provided in an embodiment of this specification is shown; Figure 9 A schematic diagram of a positioning device provided in an embodiment of this specification is shown; Figure 10 A schematic diagram of a computer device provided in an embodiment of this specification is shown.

[0024] [Explanation of Figure Markers]: 901. Data Acquisition Module; 902. First Acquisition Module; 903. Solving Module; 904. Second Acquisition Module; 905. Determine the module; 1002. Computer equipment; 1004. Processing equipment; 1006. Storage resources; 1008. Drive mechanism; 1010. Input / Output Module; 1012. Input devices; 1014. Output devices; 1016. Presentation device; 1018. Graphical User Interface; 1020. Network interface; 1022. Communication link; 1024. Communication bus. Detailed Implementation

[0025] The present specification will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the scope of this specification, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of this specification.

[0026] It should be understood that in the description of the specific embodiments in this specification, terms such as "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature.

[0027] In the specific embodiments described in this specification, unless otherwise explicitly stated and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a movable connection, a detachable connection, or an integral part; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this specification according to the specific circumstances.

[0028] In the specific embodiments described herein, unless otherwise expressly specified and limited, the term "above" or "below" the second feature may include direct contact between the first and second features, or contact between the first and second features not in direct contact but through another feature between them.

[0029] In the specific implementation of this specification, unless otherwise expressly stated and limited, the term "more" means two or more.

[0030] This specification provides the operational steps of the methods described in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operational steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiments or drawings can be executed sequentially or in parallel.

[0031] Among related technologies, acoustic positioning technology primarily employs a multi-station spatial intersection method to calculate location, which imposes significant environmental limitations and makes deployment difficult. Furthermore, acoustic positioning technology also faces the challenge of environmental complexity, including multiple reflections and scattering paths, as well as interference from environmental noise, all of which severely impact positioning accuracy.

[0032] To address the aforementioned technical problems, this specification provides a positioning method, apparatus, and electronic device.

[0033] To facilitate understanding, we will first provide an overall description of the system. The system's physical architecture mainly consists of two parts: the target reference device and the device to be located.

[0034] Target reference devices, typically serving as reference nodes with prior known coordinates, are deployed in the physical environment of the positioning area. These devices are configured to periodically or on-demand synchronously transmit wireless broadcast signals (such as radio frequency signals) and acoustic positioning pulse signals (such as ultrasonic signals modulated by spread spectrum sequences).

[0035] The positioning device is typically mounted on a moving object (such as a person, drone, or mobile robot) that requires three-dimensional coordinates. This device integrates a radio receiver module, an acoustic array, an inertial measurement unit (IMU), and a processor.

[0036] To facilitate understanding, typical application scenarios of this solution are introduced. Specifically, the positioning method proposed in this specification is particularly suitable for complex spaces where satellite positioning signals are lacking and severe multipath / reverberation interference exists.

[0037] Taking the typical scenario of an underground mine environment as an example, the space in mine tunnels is relatively narrow, and the rock walls and metal supports will strongly reflect and refract sound waves and electromagnetic waves. In this environment, traditional single-channel ranging or narrowband acoustic ranging is very prone to mistakenly locking onto reflected waves instead of direct waves, resulting in large deviations in distance and angle calculations. At the same time, the posture of underground workers or unmanned logistics vehicles changes frequently when they move, and pure ranging schemes without attitude calibration cannot provide accurate three-dimensional positioning information.

[0038] The solution provided in this manual can also achieve high-precision positioning in the above typical scenarios, as detailed below.

[0039] The technical solutions in this specification will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] Figure 1 This is a flowchart illustrating a positioning method provided in an embodiment of this specification. The method in this embodiment can be executed by the device to be positioned, specifically by the processor of the device to be positioned. For ease of description, the device to be positioned will be referred to as the execution subject in the following description. The device to be positioned includes an acoustic array. For example... Figure 1 As shown, the method specifically includes the following steps: S101. Acquire the acoustic positioning pulse signal emitted by the target reference device.

[0041] The device to be located controls the acoustic array to maintain signal listening. The target reference device transmits acoustic positioning pulse signals into the space medium. The device to be located uses the acoustic array to receive these acoustic positioning pulse signals. The acoustic array contains multiple spatially discrete array elements (e.g., acoustic microphones). Each element in the acoustic array independently receives the signal. The device to be located amplifies and filters the analog acoustic signals received by each element. After filtering, the device to be located uses an analog-to-digital converter to discretely sample the analog acoustic signals. Through discrete sampling, the device to be located obtains a digitized acoustic positioning pulse signal.

[0042] S102. Obtain feature information characterizing the arrival time difference between different array elements in the acoustic array when the acoustic positioning pulse signal arrives.

[0043] Because different array elements in an acoustic array are distributed at different physical spatial locations, the physical path lengths of the acoustic positioning pulse signals arriving at different array elements vary. This difference in physical path length results in a sequential arrival time of the acoustic positioning pulse signals at different array elements. The positioning device performs synchronization alignment processing on the digitized acoustic positioning pulse signals corresponding to each array element. Specifically, the positioning device acquires characteristic information representing the arrival time difference between different array elements. The arrival time difference is the time difference between the arrival times of the acoustic positioning pulse signals at any two different array elements.

[0044] It should be noted that the feature information can encompass not only the specific physical values ​​that directly characterize the time difference, but also complex sequence feature matrices that implicitly contain the arrival time difference and relative phase relationship. In some implementations, the device to be located directly performs correlation peak detection on the sampled waveform data to extract the time difference values ​​at the nanosecond or microsecond level as feature information.

[0045] In other embodiments, since the physical differences of sound waves under different time delays are directly reflected as phase differences on the complex plane, the feature information can also be a two-dimensional complex matrix that implicitly contains the physical characteristics of arrival time difference / phase difference without explicit numerical calculation, and can be directly used as the input source of a neural network. Both of the above fall within the scope of feature information protected in this application.

[0046] S103. Calculate the relative spatial angle of arrival of the target reference device relative to the device to be located based on the arrival time difference.

[0047] The device to be located reads the internal geometric topology parameters of the acoustic array. These parameters include the three-dimensional coordinates of each array element within the device's body coordinate system. The device then combines these internal geometric topology parameters with feature information to calculate the spatial angle. Through this spatial angle calculation, the device obtains the relative spatial angle of arrival (AOA) of the target reference device relative to itself. The AOA characterizes the pointing vector direction of the target reference device within the device's body coordinate system. For example, the AOA includes a horizontal azimuth component and a vertical pitch component.

[0048] In some embodiments, when the feature information is the arrival time difference with a specific value, the device to be located calculates the path difference by combining the arrival time difference with the current ambient sound speed, and uses the path difference and the array element spacing to substitute into the parameters of the geometric constraint equation set to solve the geometric constraint equation set to obtain the relative spatial angle of arrival.

[0049] In some embodiments, the sensor array mounted on the device to be positioned includes at least two non-collinear physical baselines in geometric arrangement, forming a first connecting arm and a second connecting arm (e.g., arranged in an "L" shape or a "+" shape).

[0050] Based on the above physical structure, the process of calculating the relative spatial angle of arrival of the target reference device using two connecting arms relies on a standard spatial geometric analytical mathematical model, as follows: After extracting the direct wave signal segments corresponding to each array element, the arrival timestamp of the signal is first established by cross-correlation algorithm or peak detection, and the time difference between the arrival of the direct wave signal and the different array elements on the same connecting arm is calculated.

[0051] Subsequently, for the first connecting arm, the incident angle (e.g., azimuth projection angle) of the target signal relative to the axis of the first connecting arm is calculated using the time difference between the signals received by the array elements at both ends of the connecting arm, the known length of the physical baseline, and the speed of sound propagation, based on spatial geometric trigonometric function relationships (e.g., inverse cosine operation).

[0052] Similarly, using the corresponding matrix data on the second connecting arm, the incident angle (e.g., elevation projection angle) of the target signal relative to the axis of the second connecting arm is calculated using the same geometric direction finding principle.

[0053] Finally, by combining the spatial angle relationship between the first and second connecting arms in the local coordinate system of the sensor itself, the incident angles of the above two dimensions are orthogonally or obliquely projected and synthesized in three dimensions, which can eliminate spatial conical ambiguity and calculate the relative spatial arrival angle of the target reference device relative to the device to be positioned (i.e., the spatial vector angle composed of the horizontal azimuth angle and the vertical pitch angle).

[0054] S104. Obtain the relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located.

[0055] To obtain the relative distance, the flight time of the sound wave signal from the target reference device to the receiving device can be obtained through methods such as acoustic-optical synchronization or one-sided two-way ranging, and the physical distance between the two can be calculated accordingly.

[0056] To acquire the device's attitude parameters, the inertial measurement unit (IMU) built into the device can collect its three-dimensional attitude data in the geographic coordinate system in real time. The attitude parameters include at least the yaw, pitch, and roll angles, which describe the rotational state of the device's coordinate system relative to an external absolute reference frame.

[0057] S105. Based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device, determine the spatial position of the device to be located.

[0058] The known location of the target reference device is represented by its latitude and longitude or absolute spatial coordinates in an absolute physical three-dimensional coordinate system. This known location can be pre-stored in the device to be located. Alternatively, for example, the known location of the target reference device can be obtained by looking up a table using the ID number given by the wireless broadcast signal, or it can be directly given in the message of the broadcast signal.

[0059] For example, the device to be positioned uses the direction indicated by the relative spatial arrival angle as the vector direction and the relative distance as the vector magnitude. Based on this vector direction and magnitude, the device establishes a polar coordinate spatial vector. It then establishes a three-dimensional orthogonal coordinate system and transforms the polar coordinate spatial vector into a relative spatial position vector in the body coordinate system. Next, the device constructs a three-dimensional rotation space matrix using the heading, pitch, and roll angle data from its attitude parameters. The device then multiplies the relative spatial position vector on the left using this three-dimensional rotation space matrix. Through this matrix multiplication operation, the device aligns the relative spatial position vector to a coordinate system that conforms to the known position of the target reference device. This matrix mapping eliminates and compensates for reference system deviations caused by the device's fuselage attitude rotation. Finally, the device performs a vector addition and translation operation on the known position coordinates and the aligned and transformed relative spatial position vector. The output value of this vector addition and translation operation constitutes the absolute spatial position of the device.

[0060] Traditional acoustic positioning methods (such as polygonal positioning) typically require at least three or four reference devices with known coordinates to calculate the three-dimensional coordinates of the device to be located. This method, however, uses an acoustic array to obtain the relative spatial angle of arrival and combines it with relative distance to geometrically construct a complete spatial vector. This allows the device to be located to determine its relative position by establishing a connection with only a single target reference device, significantly reducing the system's requirements for reference device deployment density and hardware deployment costs.

[0061] Mobile devices often undergo random rotation and tilting during movement. Since the angle of arrival output by the acoustic array is based on the device's own coordinate system, if attitude is not considered, the device's own rotation can cause a significant shift in the positioning coordinates. This solution dynamically compensates for the impact of device attitude changes on angle observation by acquiring device attitude parameters in real time and performing coordinate transformation mapping. This ensures that a stable global spatial position can be output under any device attitude, improving the system's robustness in dynamic environments.

[0062] In some embodiments of this specification, in order to accurately obtain the time difference between array elements in complex multipath environments (such as indoor environments with many reflective surfaces), this invention employs a direct wave extraction strategy based on matched filtering, which can effectively remove environmental multipath interference while retaining accurate time delay information. See also Figure 2 The method for obtaining characteristic information representing the time difference between arrival of the acoustic positioning pulse signal to different elements in the acoustic array includes: S201: The acoustic positioning pulse signals received by each element in the acoustic array are subjected to baseband demodulation and matched filtering to obtain the matched filtered output signal of each element.

[0063] For example, the device to be located discretizes the multiple raw analog waveforms acquired by the acoustic array. For each sampled array element signal, baseband demodulation is first performed to downconvert it from the carrier frequency to the baseband, and then low-pass filtering is performed to remove high-frequency environmental noise.

[0064] Using a local reference signal pre-stored on the device to be located (this reference signal has consistency or a definite correlation with the original pulse waveform transmitted by the target reference device), matched filtering is performed on each baseband signal. The matched filtering process is essentially calculating the cross-correlation between the received signal and the local reference signal. Because the local reference signal perfectly matches the transmitted signal, after filtering, the signal energy is concentrated within an extremely short time window, thus achieving pulse compression. This step significantly improves the signal-to-noise ratio and produces a matched-filtered output signal with sharp peaks.

[0065] S202: For each array element, determine the matching peak position from the matched filter output signal, and extract the signal based on the matching peak position to obtain the direct wave signal segment corresponding to each array element.

[0066] In the output signal of a matched filter, multiple peaks are typically generated due to reflections from walls or obstacles in the indoor environment. The earliest peak on the time axis with an intensity exceeding a preset threshold corresponds to a direct wave propagating directly through a straight spatial path; while subsequent peaks correspond to multipath reflected waves.

[0067] In this step, the device to be located executes a peak detection algorithm to find the first maximum point where the amplitude exceeds a specific background noise threshold. The first maximum point corresponds to the direct wave energy in the air medium without physical obstruction. The device to be located extracts the matching peak position corresponding to this first maximum point. Subsequently, using the time scale of this matching peak position as the center, the signal is truncated forward and backward according to a preset time length (e.g., 1.5 to 2 times the pulse width), retaining the waveform near the center point and removing reflected wave components and other reverberation noise outside the window. Through this operation, the long-term continuous signal, which was originally mixed with reflected waves, is processed into a pure direct wave signal segment containing only single-pulse characteristics.

[0068] S203: Based on the direct wave signal segment corresponding to each array element, obtain the feature information.

[0069] The direct wave signal segment is a clean signal stripped of environmental multipath reflection interference. When the positioning device extracts feature information based on the processed direct wave signal segment, if the feature information is represented as a specific numerical value of the arrival time difference between different array elements, the positioning device extracts the timestamp corresponding to the matching peak position in the direct wave signal segment of different array elements. For example, the calculation of this timestamp can be performed using the cross-spectral estimation method (GCC) in broadband pulse delay estimation. During the calculation, a specific array element in the acoustic array is used as a reference, and the physical time difference between the timestamps of other array elements and the timestamp of the reference is calculated. Through the above algebraic subtraction operation, the time delay when the acoustic positioning pulse signal arrives at array elements at different spatial positions is restored, thereby obtaining specific nanosecond or microsecond-level time delay data. This time delay data, i.e., the arrival time difference, can be directly used as feature information for subsequent analytical calculation of the direction angle by substituting it into the spatial geometric constraint equations.

[0070] In some embodiments, the feature information may be a two-dimensional complex matrix containing physical laws such as time difference of arrival and relative phase. For information on obtaining this feature information, please refer to the following text.

[0071] By adopting the above-mentioned processing flow based on matched filtering and signal segment truncation, multipath artifacts in the indoor environment can be effectively filtered out, ensuring that the time difference data input into the subsequent positioning solution model is accurate and direct path data that is not affected by reverberation.

[0072] In some embodiments, the acoustic array includes a first acoustic subarray. Under this solution logic, the feature information acquired by the device to be located is specifically the numerical value of the time difference of arrival between different array elements. See also Figure 3 The relative spatial angle of arrival of the target reference device relative to the device to be located is calculated based on feature information, specifically including the following steps: S301: Combine the array elements in the first acoustic subarray to construct multiple sets of first array element connecting arm combinations, each first array element connecting arm combination containing two non-parallel first array element connecting arms.

[0073] In a three-dimensional spatial angle measurement scenario, within the first acoustic subarray, any two array elements can establish a line segment with a fixed baseline length and orientation; this line segment serves as the connecting arm of the first array elements. Multiple connecting arms of the first array elements are combined to construct multiple combinations. Each combination of connecting arms contains two non-parallel connecting arms of the first array elements.

[0074] See Figure 4 , Figure 4 A schematic diagram of an acoustic array provided in an embodiment of this specification is shown. Figure 4 The diagram illustrates an octagonal planar array comprising eight array elements, which is divided into two square quaternary arrays of different sizes. The four array elements 401 in the middle region constitute the first acoustic subarray, while the four array elements 402 on the periphery constitute the second acoustic subarray. For example, the spacing between adjacent array elements in the first acoustic subarray and the spacing between adjacent array elements in the second acoustic subarray are 15 mm and 30 mm, respectively.

[0075] S302: For each combination of first array element connecting arms, calculate the candidate angle of arrival for each combination of first array element connecting arms based on the arrival time difference of each first array element connecting arm under the combination.

[0076] For each first element connecting arm combination, the device to be located extracts the time difference between the two end elements of each first element connecting arm constituting the combination, thereby obtaining the arrival time difference corresponding to each first element connecting arm. Combined with the constant sound wave propagation speed in the current environment, the above arrival time difference is converted into the objective path difference traversed by the target reference device to the two end elements of the corresponding connecting arm. The device to be located establishes the phase deflection constraint equation based on the path difference and the geometric baseline length of the first element connecting arm. By simultaneously solving the phase deflection constraint equations corresponding to the two first element connecting arms belonging to the same combination, the candidate angle of arrival of each first element connecting arm combination pointing to the target reference device in the local solution plane is calculated. Since the local multipath interference of different elements exists independently, each candidate angle of arrival represents the preliminary direction finding result calculated based on the local element field of view.

[0077] S303: For each combination of first array element connecting arms, the arrival time difference of each first array element connecting arm under the combination and the spatial angle between different first array element connecting arms are used to assign confidence weights to the combination of first array element connecting arms.

[0078] In array signal detection theory, when the measurement baseline faces the same incident sound wave, it is limited by the geometric projection relationship, and the resulting measurement error is strongly correlated with the incident angle of the sound wave.

[0079] Specifically, in theory, two non-parallel arms can provide a set of spatial orientation estimates. However, in reality, measurements are subject to noise, and the accuracy is relatively poor if the sound source is in the direction of the line connecting the two array elements. Therefore, different combinations of non-parallel arms have different confidence weights. By accumulating the orientation results of different arm combinations according to a certain weight relationship (vector superposition), the final orientation estimate can be obtained.

[0080] In some embodiments, the confidence weight is calculated based on the following formula:

[0081] in, Indicates the arm connected by the first element. Second array element connecting arm The first element of the combination is the connecting arm combination. The corresponding confidence weights; Indicates the first element connecting the arms The corresponding arrival time difference; Indicates the second array element connecting arm The corresponding arrival time difference; Indicates the first element connecting the arms Second Array Link Arm The angle between them; Indicates the first prevention and elimination of zero items; This indicates the second prevention and control item.

[0082] For example, in the case where the first acoustic subarray contains 4 array elements, the time delay difference corresponding to the connecting arm of the 6 array elements can be calculated.

[0083] Referring to the formula above, the closer the spatial angle between the two element connecting arms in the first array element connecting arm combination is to 90 degrees, the better the orthogonal independent observation of the two measurement baselines, and the corresponding absolute value of the sine... The larger the parameter, the higher the confidence weight of the first element connecting arm combination.

[0084] Meanwhile, in array geometry direction finding, the smaller the absolute value of the time difference of arrival, the closer the physical incident direction of the acoustic wave is to the perpendicular bisector normal of the array element connecting arms. Within this range, the angular deviation generated when the underlying measurement delay error is mapped to spatial angle is minimal, resulting in the most reliable direction finding results. Therefore, the formula places the absolute value of the time difference of arrival in the denominator to construct a reciprocal structure, aiming to adaptively assign greater confidence weights to array element connecting arms that possess extremely high solution accuracy due to their proximity to the normal incident direction, thereby effectively improving the noise and distortion resistance of the global resultant vector.

[0085] In addition, the first and second zero-prevention terms are introduced into the denominator structure as pre-configured minimal positive real number penalty factors to avoid computer zero-prevention overflow interruption caused by the time difference being completely zero when the target reference device is exactly on the normal line, thus ensuring the normal operation of the algorithm under extreme conditions.

[0086] S304: Using the candidate arrival angle as the direction and the confidence weight as the modulus, construct a three-dimensional vector corresponding to each combination of the first array element connecting arms.

[0087] In the specific implementation, the first array element is used as the connecting arm. Second array element connecting arm For example, the angle between the two is denoted as . When the two arms are parallel, the acoustic observation baseline completely overlaps in space or degenerates into a linear projection of the same dimension, losing orthogonality and independence. Therefore, geometrically, it is impossible to obtain the complete three-dimensional spatial orientation of the acoustic pulse solely through the parallel array element connecting arms.

[0088] Based on the non-parallel structure, the devices to be positioned integrate their independent numerical angle transformations into a geometric Cartesian coordinate system. Specifically, for any combination of connecting arms of the first array element... Let the candidate arrival angle initially calculated for this combination in spherical coordinates be . .in These represent the azimuth and elevation angles of the local solution, respectively. The confidence weights corresponding to the first array element connecting arm combination calculated in the previous step are also taken.

[0089] By mapping trigonometric functions from spherical coordinates to Cartesian coordinates, the device to be positioned constructs the three-dimensional orientation vector corresponding to the combination of the connecting arms of the array elements. The calculation formula is as follows:

[0090] S305: Superimpose the three-dimensional vectors corresponding to the connecting arms of each of the first array elements, and determine the relative spatial arrival angle based on the direction of the superimposed resultant vector.

[0091] After completing the numerical evaluation of all confidence weights, the device under test transforms and integrates the originally independent numerical angle information into the geometric vector space. The device under test uses the spatial direction indicated by the calculated candidate angle of arrival as the three-dimensional vector direction, and directly uses the magnitude of the confidence weight output by the formula as the corresponding spatial modulus parameter to construct the three-dimensional vector associated with the combination of the connecting arms of each independent first element.

[0092] Next, the device under test superimposes the three-dimensional vectors independently generated by combining all the first array element connecting arms. By executing the spatial polygon vector addition rule of multiple three-dimensional vectors, the jump error points with high variance are decoupled and dissipated, and finally the stable aggregation resultant vector direction is smoothly calculated. The synthesis process of the resultant vector fully amplifies the decision-making proportion of high-quality orthogonal arms and wide-angle high-sensitivity arms. The device under test ultimately determines the global relative spatial angle of arrival directly from the superimposed resultant vector direction for subsequent positioning distance mapping.

[0093] In the above implementation, by constructing a combination of non-parallel array element connecting arms and calculating candidate angles of arrival based on each combination, the geometric redundancy information of the array can be fully utilized for multi-view observation. When fusing multiple observation results, confidence weights based on the orthogonality of the measurement baseline and the magnitude of the time delay are introduced. Since the observation signal in the normal region of the array element connecting arm has a higher tolerance to the noise of the underlying time delay measurement, the reciprocal weighting strategy can adaptively filter and amplify the proportion of high-precision channels. Finally, all the calculation results are transformed into a three-dimensional vector space for superposition. The natural smoothing mechanism of the combined vector effectively dilutes the local direction finding distortions caused by multipath effects or sudden environmental noise interference. This calculation method, which relies on physical laws to allocate weights and perform noise reduction fusion, does not require the introduction of complex nonlinear iterative search calculations. While ensuring low computational overhead, it significantly improves the accuracy of target angle of arrival estimation and the robustness of the algorithm to complex acoustic environments.

[0094] In some embodiments, the acoustic array is a multi-scale acoustic array, which includes a first acoustic subarray and at least one second acoustic subarray; wherein the element spacing of the second acoustic subarray is greater than the element spacing of the first acoustic subarray.

[0095] To balance the global uniqueness and high accuracy of spatial angle measurement, the positioning device employs a combined solution method of coarse measurement followed by fine measurement. Specifically, the first acoustic subarray, due to its small spacing between adjacent elements (e.g., less than or equal to half the wavelength of the corresponding acoustic carrier frequency band), ensures, according to the spatial data sampling theorem, that no phase aliasing or spatial ambiguity occurs from any angle of physical incidence within the entire spatial detection field of view, thus enabling the calculation of a unique and precise direction. However, limited by the small spatial aperture, its absolute angle measurement accuracy has a physical upper limit. The second acoustic subarray has a larger element spacing, several times the wavelength of the acoustic wave. Increasing the observation baseline length significantly improves the sensitivity and angle measurement accuracy, but due to insufficient spatial sampling rate, periodic phase flips occur when solving inverse trigonometric functions over a wide-angle range, leading to multiple equally probable ambiguous solutions within the spatial solution domain.

[0096] Based on this physical property, see Figure 5 In some embodiments, the step of performing high-precision, unambiguous angle measurement on the device to be located based on a multi-scale acoustic array specifically includes: S501: The relative spatial angle of arrival calculated based on the first acoustic subarray is used as the initial spatial angle of arrival.

[0097] The unambiguous observation results obtained by the device to be positioned in the preliminary step, which rely solely on the output of the first acoustic subarray, are marked as the initial spatial angle of arrival and used as the prior constraint benchmark for subsequent large-aperture fine solution.

[0098] S502: Based on all the array elements of the first acoustic subarray and the second acoustic subarray, construct multiple second array element connecting arms with different baseline lengths.

[0099] To maximize the utilization of the hardware array, the entire array element set is combined and paired. By using physical array element connections spanning the first and second acoustic subarrays, a set of second array element connection arms covering both short and long baselines is generated. By adding combinations of longer arms, more accurate direction finding results can be obtained.

[0100] S503: Based on the initial spatial arrival angle, perform directional filtering to select candidate connecting arms from the plurality of second array element connecting arms whose spatial angle with the direction indicated by the initial spatial arrival angle is within a preset angle range; the preset angle range includes 90 degrees.

[0101] The closer the baseline direction is to the incident direction of the sound source, the smaller the angular calculation distortion caused by the time delay error mapping. Therefore, the positioning device uses the direction indicated by the initial spatial angle of arrival as the approximate reference axis for the incident sound wave, and calculates the relative spatial angle between each second element connecting arm and this reference axis. Subsequently, axial connecting arms that are prone to divergence due to severe polarization are eliminated, and those spatial connecting arms that fall within a preset angle range (e.g., 80 to 100 degrees) containing the 90-degree normal around them are retained as candidate connecting arms.

[0102] S504: Select a first preset number of first long baseline arms from the candidate connecting arms in order of baseline length from longest to shortest.

[0103] Under the same conditions of high sensitivity observation in the normal direction, the longer the physical length of the baseline, the higher the spatial phase resolution it provides. The positioning device measures the straight-line distance between the first and last array elements of each candidate baseline arm, sorts them in descending order, and preferentially selects the baseline with the longest physical span from the first preset number (e.g., 3) at the top of the sequence, identifying it as the first long baseline arm. For example, taking 3 as the first preset number, the determined first long baseline arm is denoted as... .

[0104] S505: From the candidate connecting arms, select a second preset number of second long baseline arms in descending order of baseline length; the spatial angle between the second long baseline arm and the first long baseline arm is within the preset angle range.

[0105] To establish complete cross-positioning constraints, another set of observation baselines needs to be determined. The device to be positioned continues to search for a second, predetermined number of long baseline arms in descending order of length from the remaining candidate connecting arms. During the selection process, in addition to considering the baseline length, an orthogonality fitness test is also required: the extracted second long baseline arm must maintain a preset angle range of approximately 90 degrees with the determined first long baseline arm in space to ensure that the combination of the two sets of measurement baselines can maximize measurement accuracy.

[0106] For example, taking a second preset number of 3 as an example, the determined second long baseline arm is denoted as... .

[0107] S506: Construct a long baseline arm combination based on the first long baseline arm and the second long baseline arm; each long baseline arm combination includes a first long baseline arm and a second long baseline arm.

[0108] The device to be positioned will combine the first long baseline arm and the second long baseline arm, which have been selected through dual constraints of length and orthogonality, into pairs to generate a long baseline arm combination with the largest effective observation aperture and the best noise-resistant projection surface, providing a foundation for the high-precision solution in the next stage.

[0109] For example, the first long baseline arm includes The second long baseline arm includes Therefore, the two can be combined to obtain 9 long baseline arm combinations.

[0110] S507: Based on the time difference of arrival corresponding to the long baseline arm combination, calculate multiple candidate angles with phase ambiguity; set an angle calculation window with a boundary span centered on the initial spatial angle of arrival, and select the target candidate angle that falls within the angle calculation window from the multiple candidate angles with phase ambiguity as the relative spatial angle of arrival without phase ambiguity; wherein, the boundary span of the angle calculation window is smaller than the phase ambiguity period generated by the long baseline arm combination.

[0111] The positioning device extracts the time delay data of the endpoint array elements of the long baseline arm combination and substitutes them into the nonlinear spatial constraint equation. Since the equivalent spacing of the long baseline combination is much larger than half a physical wavelength, the equation will inevitably output a series of equally spaced candidate solutions when analyzing the entire spatial domain.

[0112] Subsequently, the device to be positioned calls upon the previously acquired low-precision initial spatial angle of arrival. Using the pointing degree of this initial spatial angle of arrival in the coordinate system as the center position, a certain tolerance limit is extended to both sides to construct a spatial angle calculation window. The boundary span of this window is strictly limited by the system configuration parameters to ensure that its physical width is strictly less than the natural phase ambiguity period (i.e., the angle difference between the directions of two adjacent grating lobes) generated by the long baseline arm combination at the corresponding frequency.

[0113] The positioning device performs a geometric inclusion comparison between each candidate angle with phase ambiguity output from the previous equations and the angle calculation window. Since the window width is smaller than the period span, among multiple candidate angles with ghosting, only one will necessarily fall within the window's intercept interval due to its close proximity to the true physical direction. The positioning device captures this unique target candidate angle and eliminates all other spurious angles caused by spatial aliasing, thereby obtaining the final relative spatial angle of arrival that inherits the ultra-high observation accuracy of long baselines while eliminating the hidden dangers of periodic spatial ambiguity.

[0114] As can be seen, in the embodiments of this specification, the first acoustic subarray with small element spacing provides an initial spatial angle of arrival without phase aliasing as a priori directional guide. Subsequently, the original subarray separations are broken down to perform global long baseline combination, and a selection process is conducted based on the principle of physical geometric optimality. Long baseline arms that are nearly perpendicular to the incident direction of the target acoustic wave, have the longest possible physical span, and are nearly orthogonal to each other are merged into the optimal long baseline combination. This selection mechanism significantly reduces the mapping distortion when converting the underlying time delay error to spatial angle from the measurement mechanism, providing high-quality input for high-precision calculation. Finally, a small-span calculation window is set using the unambiguous initial angle from the previous stage, and the unique true candidate angle is determined from the periodic spatial multiple solutions (phase ambiguity and ghosting) inevitably generated by the long baseline combination. This fully integrates and utilizes the maximum effective spatial aperture resource composed of all array elements distributed across multiple scales on the measured object, outputting a relative spatial angle of arrival with both global uniqueness and high accuracy without introducing highly complex global spatial spectrum search calculations and with relatively low system computing power overhead.

[0115] In some embodiments, the position calculation of the device to be located needs to be unified to a preset global physical coordinate system. The following section will further introduce the step of "determining the spatial position of the device to be located based on the device attitude parameters, relative spatial angle of arrival, relative distance, and the known position of the target reference device" in conjunction with the coordinate system transformation.

[0116] For example, assuming the absolute reference system upon which the algorithm is based is the geodetic coordinate system, in which the positive X-axis corresponds to the due east direction, the positive Y-axis corresponds to the due north direction, and the positive Z-axis corresponds to the vertical upward direction.

[0117] First, acquire and parse the device attitude parameters of the device to be located at the current instant of motion.

[0118] The device to be positioned outputs the Euler angle information of its sensor array in space in real time through its integrated inertial navigation sensor module. Specifically, the device's attitude parameters mainly include three rotational parameters: yaw angle (denoted as...). ), representing the angle of rotation about the Z-axis; pitch angle (denoted as Z-axis) ), representing the angle of rotation about the Y-axis; roll angle (Yaw, denoted as ), represents the angle of rotation about the Y-axis. ), representing the angle of rotation about the X-axis.

[0119] Secondly, the observation results in the local spherical coordinate system are transformed into an azimuth vector in the carrier space rectangular coordinate system. Using the relative distance R and relative spatial angle of arrival (including the determined unambiguous azimuth and elevation angles of arrival) obtained in the previous steps, the relative coordinate estimates of the target reference device are constructed in the spherical coordinate system of the device to be located. By performing a three-dimensional rectangular projection of the above spherical coordinate parameters, the sound source azimuth vector based on the sensor's own carrier coordinate system can be obtained, denoted as... .

[0120] Subsequently, a vector mapping transformation from the carrier coordinate system to the geodetic coordinate system is completed using a coordinate system rotation matrix. To align the local observation azimuth vector with the global absolute space, the device to be positioned is based on the aforementioned device attitude parameters. Construct a coordinate system rotation matrix, perform a coordinate system rotation operation, and obtain the position vector in the geodetic rectangular coordinate system. The transformation process satisfies the following space matrix multiplication formula:

[0121] in, The rotation matrix representing the transformation from the carrier coordinate system to the preset geodetic coordinate system, taking into account the rotation components of each Euler angle, can be expressed in its matrix expansion form as follows:

[0122] In the matrix operator above, to simplify the expression, the triangular operator is defined. And operators .

[0123] Finally, the position vector is reversed to determine the final absolute spatial position of the device to be located.

[0124] After the above rotation transformation, the position vector In a physical sense, it represents the global direction and its three-dimensional distance difference from the device to be positioned to the target reference device. According to the vector law, to determine the unknown position of the device to be positioned, the known position coordinates of the target reference device, pre-calibrated in the global geodetic coordinate system, are used. Using the absolute spatial reference origin, the three-dimensional spatial position of the device to be positioned at this moment is calculated by applying the spatial vector difference rule. The calculation formula is as follows:

[0125] Through the rigorous three-dimensional matrix rotation and vector back-calculation described above, the high-precision acoustic ranging and multi-subarray joint angle measurement results are decoupled and converted with the instantaneous attitude of the equipment itself, eliminating the positioning offset caused by random rotation and tilt of the equipment, and thus accurately giving the precise spatial coordinates of the equipment to be positioned in the geodetic coordinate system.

[0126] In practical indoor or localized positioning scenarios, multiple reference devices are typically deployed within a space to achieve large-area coverage and multi-point positioning convergence. The following details how the device to be positioned can accurately extract the acoustic signal from a specific target reference device in such an environment with multiple reference devices coexisting.

[0127] Specifically, in some embodiments, the target reference device is one of a plurality of reference devices configured to simultaneously transmit wireless broadcast signals and acoustic positioning pulse signals.

[0128] For example, wireless broadcast signals may include radio frequency signals such as Bluetooth, Wi-Fi, and UWB. Wireless broadcast signals propagate at the speed of light and are used to transmit data; while specific acoustic positioning pulse signals propagate at the speed of sound and are used for physical ranging and angle measurement.

[0129] Accordingly, see Figure 6 The acoustic positioning pulse signals received by each element in the acoustic array are subjected to baseband demodulation and matched filtering to obtain the matched filtered output signal of each element, specifically including: S601: Receive the wireless broadcast signal transmitted by the target reference device, and parse the wireless broadcast signal to obtain the transmission frequency band information of the target reference device.

[0130] Because multiple reference devices exist in space, each reference device can be assigned a different acoustic frequency band to prevent interference between acoustic signals. When the target reference device is operational, it can simultaneously emit wireless broadcast signals and acoustic signals. Due to the speed of light, the device to be located first captures the wireless broadcast signal. Subsequently, the processor decodes the wireless broadcast signal, extracting the identification identifier and corresponding transmission frequency band information associated with the reference device currently transmitting the radio frequency signal. This transmission frequency band information indicates in which sound frequency range (e.g., 18kHz-20kHz, or 20kHz-22kHz) the reference device will transmit acoustic pulses.

[0131] S602: Based on the transmission band information, select the corresponding template as the local reference signal from a plurality of pre-configured matching signal templates.

[0132] The local memory of the device to be located contains a set of matching signal templates. These templates physically correspond to standard reference waveforms with different frequency bands and different sweep characteristics. Based on the transmission frequency band information parsed in the previous step, the processor directly retrieves a unique template from the local library that is completely consistent with the transmitted waveform characteristics of the target reference device through table lookup or mapping, and uses it as the local reference signal for the next step.

[0133] S603: Using the local reference signal, perform baseband demodulation and matched filtering on the acoustic positioning pulse signals received by each array element in the acoustic array.

[0134] After the sound wave positioning pulse signal, which travels at the speed of sound, reaches the acoustic array, the processor uses the pre-retrieved local reference signal to perform baseband demodulation operations such as mixing and down-conversion on the multiple mixed ambient sound waves collected by the microphone array; then, matched filtering is performed, which can be achieved through cross-correlation operations.

[0135] By selecting a highly matched local reference signal, the filtering process can accurately amplify and extract signal peaks belonging to a specific frequency band, while suppressing or filtering out sound waves emitted by reference devices in other unmatched frequency bands in the space as out-of-band noise.

[0136] It is evident that in environments with multiple reference devices, the acoustic waves emitted by each device are prone to severe acoustic crosstalk. This solution utilizes the frequency band configuration of radio frequency signal transmission, enabling the device to be located to clearly identify the frequency band of the acoustic positioning pulse signal, thereby selecting a suitable local reference signal. Through the orthogonality of matched filtering, the acoustic waves of specific target devices can be accurately isolated, effectively shielding crosstalk interference from other devices and ensuring the reliable operation of the multi-node positioning network.

[0137] In some embodiments of this specification, see Figure 7 Obtaining the relative distance between the device to be located and the target reference device includes: S701: Receive the wireless broadcast signal transmitted by the target reference device, and use the time of receiving the wireless broadcast signal as the reference time.

[0138] At the physical level, the propagation speed of wireless broadcast signals is much greater than that of acoustic positioning pulse signals. Therefore, in specific positioning environments, the propagation time of wireless signals is extremely short and can be ignored. Thus, when the radio frequency antenna of the device to be located receives the wireless broadcast signal from the target reference device, the processor immediately adds a high-precision timestamp to its local timeline, using it as the reference moment. This reference moment can be considered the moment when the acoustic positioning pulse signal was just emitted from the reference device.

[0139] S702: Determine the effective observation time window for the target reference device based on the preset physical ranging boundary.

[0140] In practical applications, the effective coverage of a reference device is physically limited, for example, covering a radius of 50m. The processor determines the physical ranging boundary (e.g., 50m) based on factory configuration or the current application scenario, and converts it into a time boundary using the speed of sound. The processor then adds the time boundary to the reference time, thus planning an effective observation time window on the timeline.

[0141] S703: Within the effective observation time window, feature matching is performed on the acoustic wave signals acquired by the acoustic array.

[0142] For example, the device to be located does not perform position calculations at all times, but rather performs cross-correlation calculations on the acoustic signals received by the acoustic array within the effective observation time window, i.e., feature matching.

[0143] S704: If it is determined that the time of generation of the matching peak position is within the effective observation time window, then the time of generation of the matching peak position is taken as the actual arrival time, and the relative distance is calculated based on the time difference between the actual arrival time and the reference time, combined with the ambient sound speed.

[0144] During the matching operation within the time window, if the result of the cross-correlation operation exceeds the preset signal-to-noise ratio detection threshold, the processor determines that the matching peak has been successfully captured.

[0145] The method for the processor to capture direct acoustic signals is described above and will not be repeated here.

[0146] The device to be located records the exact time corresponding to the matching peak as the actual arrival time of the sound wave. Then, this arrival time is subtracted from the reference time obtained in the first step to obtain the actual time difference in the sound wave's flight through the air. Finally, this time difference is multiplied by the real-time ambient sound speed to obtain the relative distance.

[0147] As can be seen, in the embodiments of this specification, the significant difference in propagation speed between wireless radio frequency signals and sound wave pulse signals is utilized, with the arrival time of the delay-free wireless signal directly used as the timing starting point for ranging calculation. The device to be located only needs to perform unidirectional signal listening locally, without the need to establish a complex bidirectional clock coordination protocol, to accurately reconstruct the true flight time of the sound wave in space, reducing the deployment difficulty and hardware cost of the positioning system.

[0148] Furthermore, in highly reflective environments such as indoors or underground, sound waves are easily reflected by walls or obstacles, resulting in multipath fading and causing the receiver to capture delayed echoes that do not propagate in a straight line. This scheme calculates the effective observation time window through a preset physical sensing boundary, establishing a reasonable physical filtering boundary. It can directly eliminate long-path reflected signals that arrive too late on the time axis. This is equivalent to building an anti-interference barrier in the time domain, reducing the distance misjudgment rate caused by incorrect matching of reflected waves, and enhancing positioning stability.

[0149] To aid understanding, further explanation is provided below with reference to the accompanying diagram. See also... Figure 8 The scenario in this embodiment mainly consists of a target reference device 801 located at a fixed position in space and a moving device to be positioned 802.

[0150] The target reference device 801 is fixed in an absolute coordinate system (e.g., a geodetic coordinate system), which is defined by the azimuth reference. (As shown in the picture, due east, E) (as shown in the figure, due north N) and vertical reference (For example, vertically upward) forming a three-dimensional space.

[0151] The device 802 to be located has multiple receiving array elements distributed on the same plane (as shown by the black dots on the label surface in the figure, which together constitute the aforementioned multi-scale planar acoustic array) for receiving acoustic positioning pulses emitted by the reference device. Since the device to be located is usually worn by personnel or mounted on mobile devices, its spatial attitude changes randomly. Therefore, a carrier coordinate system bound to the device itself (based on...) is established. During actual positioning calculations, there is a spatial rotational misalignment between the carrier coordinate system and the absolute coordinate system, consisting of yaw angle, pitch angle, and roll angle. This rotational misalignment is monitored and output in real time by the inertial measurement unit installed inside the device to be positioned.

[0152] on the other hand, Figure 8 The lower right corner of the image shows the waveform envelope of a pulse signal from a certain array channel in the device to be located after matched filtering. The horizontal axis represents the sampling point sequence (or time axis), and the vertical axis represents the amplitude of the detector output.

[0153] In complex environments such as underground spaces or indoors, the acoustic pulses emitted by the reference equipment not only reach the receiver over a straight distance but also undergo strong multipath reflections from walls and the ground. As shown in the waveform diagram, multiple energy peaks will appear sequentially in a real acoustic environment. The first peak that arrives and whose amplitude exceeds the threshold is the direct wave, while the subsequent secondary peaks are multipath interference echoes generated by environmental reflections.

[0154] This solution, during front-end signal processing, uses a specific time-domain window to precisely capture the main pulse waveform data that first crosses the threshold (i.e., the direct-reach matching peak, such as...). Figure 8 The 803 shown in the figure serves as the reference input for subsequent solutions. This truncation operation filters out interference from multipath reflections.

[0155] In summary, the positioning logic of this embodiment includes: firstly utilizing... Figure 8 High-precision time delay estimation is performed on the direct wave matching peak signal shown in 803 to calculate the spatial angle of arrival (and relative distance) of the reference device relative to the device to be positioned in the carrier coordinate system. Subsequently, the rotation transformation matrix is ​​calculated based on the spatial attitude information output by the internal inertial measurement unit to transform the relative spatial position vector from the carrier coordinate system and map it to the absolute coordinate system. Finally, the absolute three-dimensional coordinates of the device to be positioned are calculated in reverse by combining the known absolute coordinates of the reference device.

[0156] In some embodiments, in environments with complex spatial structures, materials such as glass curtain walls act as strong reflective surfaces for acoustic signals such as ultrasonic waves, which can cause strong multipath reflection interference.

[0157] To overcome the aforementioned environmental complexity, provide good error tolerance for local adverse conditions such as individual array element failures, severe multipath interference, and unfavorable sensor attitude, and make full use of the redundant array element information of multi-scale arrays, this application provides a scheme for localization calculation based on a pre-trained deep learning model.

[0158] It should be noted that the method of solving the problem using the model itself does not generate any new physical information. The rationality of the result is still based on the physical feasibility of the acoustic baseline array arrangement. Therefore, the data input of the deep network in this embodiment still uses the direct wave signal characteristics generated by the hardware array elements mentioned above.

[0159] Specifically, for the step of "obtaining feature information based on the direct wave signal segment corresponding to each array element," in some embodiments, the direct wave signal segment is a time-domain complex baseband signal sequence. The device to be located first extracts multiple time-domain signal sampling points containing the matching peak (i.e., the position of the first leading matching peak in the aforementioned embodiment) from the direct wave signal segment corresponding to each array element. These sampling points are used to construct a two-dimensional complex matrix, which is then used as the input feature information. For example, the two-dimensional complex matrix is ​​defined as CSig(nch, msmpl), where nch represents the number of channels and msmpl represents the number of sampling points.

[0160] Taking a specific application scenario as an example, assuming that the multi-scale acoustic array contains a total of 8 effective acoustic array elements (i.e., the number of channels nch=8), and 16 time-domain signal sampling points (i.e., the number of sampling points msmpl=16) are extracted before and after the matching peak of each array element (including the peak point), a two-dimensional complex matrix with a dimension of 8×16 is generated. Subsequently, it is input into a pre-trained relative angle calculation network for angle calculation.

[0161] In some embodiments, the relative angle calculation network is specifically configured as a complex convolutional neural network (ComplexCNN), whose overall network topology includes cascaded complex feature extraction network layers and real fully connected mapping layers. The operational mechanism of this network is described in detail below.

[0162] The complex feature extraction network layer is mainly used to extract deep features that preserve phase. Specifically, considering that the core of acoustic localization lies in the arrival time difference of the signals received by each array element, and that this tiny time difference is strictly mapped to a complex "phase difference" in the complex baseband signal, its phase information must be preserved when performing convolutional activation.

[0163] In this network layer, the input two-dimensional complex matrix is ​​subjected to multiple layers of complex two-dimensional convolution operations in sequence, and nonlinear activation processing is performed in conjunction with complex batch normalization to extract complex feature maps. Among them, the nonlinear activation processing is specifically configured to perform activation calculation on the magnitude of the complex number while preserving the original phase features of the complex number (for example, by using the modulo nonlinear activation function modReLU).

[0164] For example, assuming the batch dimension of the input data is batch (in training mode, the batch value is typically 128 or 256), this complex feature extraction network layer can consist of two cascaded complex 2D convolutional layers forming the network backbone: The first layer of complex 2D convolution has 1 input channel and 32 output channels (i.e., 1→32). The kernel size is set to 3×3, followed by complex batch normalization and modReLU activation function. The second layer of complex 2D convolution has 32 input channels and 64 output channels (i.e., 32→64). The kernel size remains 3×3, followed by complex batch normalization and the modReLU activation function.

[0165] Furthermore, it should be noted that the complex two-dimensional convolution operation achieves complex frequency domain feature fusion through the cross-multiplication of the real and imaginary parts. Taking a complex feature x and a complex weight W as an example, the strict mathematical relationship of its complex convolution operation is expressed as follows:

[0166] Where the subscripts re and im represent the real and imaginary parts of the complex parameter, respectively, and the operator This represents a two-dimensional discrete convolution operator.

[0167] After completing all complex convolutions, a modulus operation is performed on the final output of the complex feature extraction network layer to degenerate and reduce the dimensionality of the extracted complex feature map containing deep spatial characteristics, generating real-valued parameters. Next, this real-valued parameter matrix is ​​input into a global average pooling layer for spatial compression, thereby stretching and compressing the three-dimensional feature map into a one-dimensional real-valued feature vector. For example, the feature data format is now converted to a (batch, 64) real-valued vector form.

[0168] Finally, the generated real-valued feature vectors are input into a real-valued fully connected mapping layer for nonlinear regression. For example, a fully connected hidden layer with the number of neurons reduced from 64 to 16 (i.e., 64→16) can be set up, and a LeakyReLU(0.1) linear unit with a slope of 0.1 can be added as an activation function. At the same time, to prevent overfitting, a random deactivation strategy with a dropout rate of 0.1 can be configured.

[0169] Following this fully connected hidden layer (16 nodes), three regression output heads are derived from the network end, generating the azimuth, elevation, and confidence probability of the target reference device relative to the device to be located. The azimuth and elevation angles obtained above can be combined to form the relative spatial angle of arrival.

[0170] The training process of the model will be described below.

[0171] In order to enable the relative angle calculation network to not only output the predicted value of the relative spatial angle of arrival, but also to adaptively evaluate the prediction reliability in the current complex multipath environment, this embodiment adopts a loss function based on heteroscedastic Gaussian uncertainty when training the complex convolutional neural network.

[0172] For example, the prediction errors of the model for the relative spatial angles of arrival in the two dimensions mentioned above are both modeled as Gaussian distributions. Specifically, for the azimuth and elevation angles, let their corresponding true label values ​​be respectively... and The actual forward output generated by the network includes the predicted mean values ​​for the two angles mentioned above (denoted as ). and ) and the log-variance characterizing the uncertainty of prediction (denoted as ) and ).

[0173] After establishing the above distribution model, the overall loss function of the network ( The negative log-likelihood of the predicted values ​​from two independent angles is constructed as the sum of their negative log-likelihoods, and its basic expression is as follows:

[0174] in, ,

[0175] To ensure absolute stability of numerical computations during actual code training processes such as gradient backpropagation, and to avoid anomalies such as division by zero or exponential overflow, the above loss function is rewritten while maintaining mathematical equivalence as follows:

[0176] The physical meaning of this loss function is: when encountering extremely poor data samples, such as strong reflections from glass curtain walls, which cause network prediction errors (e.g., When the Gaussian variance is large, the training mechanism tends to automatically increase it and actively decrease the confidence level of the prediction, thereby reducing the overall calculated loss penalty and protecting the network from the degradation caused by gradients from outliers. Conversely, if the environment is favorable and the error is small, the network will actively decrease the Gaussian variance to increase the confidence level. Through this training mechanism, the network can automatically learn the reliability of each observed sample in the current acoustic environment.

[0177] Finally, during the inference and deployment phase, the aforementioned output confidence probability... That is, the log-variance parameter learned by the network itself is directly obtained through joint exponential mapping. For example, by comprehensively integrating the evaluation values ​​from the two angle channels, the final location reliability probability can be calculated as follows:

[0178] The confidence probability value is constrained to the range of (0,1). The closer it is to 1, the less multipath contamination the given relative spatial angle of arrival data is, and the more reliable the solution is.

[0179] By combining the aforementioned complete complex network architecture with an uncertainty loss function, the trained parameter set (e.g., approximately 38,000 parameters in total) is ultimately embedded directly onto the processor at the edge of the device to be located and executed. High-risk errors can be alerted or weighted and discarded based on the output confidence probability, further improving the system's robust positioning performance in complex and irregularly reflective building environments.

[0180] As can be seen, in this embodiment, a deep network based on complex feature extraction is used to calculate the spatial angle of arrival. By directly using a temporal complex matrix as input, the network can automatically capture the phase difference features inherent between multiple redundant array elements. This not only effectively addresses multipath reflection crosstalk but also maintains a certain degree of error tolerance and calculation capability even when individual acoustic array elements malfunction or the orientation of the device to be located is unfavorable, ensuring the continuity of basic positioning services. The used complex small convolutional network is a lightweight network that can be directly deployed on the miniature NPU or processor of the device to be located, meeting the terminal's requirement for real-time and rapid output of the angle of arrival.

[0181] Furthermore, during the model training phase, by having the network output both the mean and variance simultaneously, even if extremely damaged sample data (such as extreme environmental noise or strong reflection occlusion) is fed in, the network will reduce the penalty by increasing the variance, thus avoiding the exaggerated gradients generated by abnormal samples from destroying the learned normal feature weights and making the model performance more stable.

[0182] Based on the same inventive concept, embodiments of this specification also provide a positioning device applied to a device to be positioned, the device to be positioned including an acoustic array, see [link to documentation]. Figure 9 The device includes: Acquisition module 901 is used to acquire the acoustic positioning pulse signal emitted by the target reference device; The first acquisition module 902 is used to acquire feature information characterizing the arrival time difference between different array elements in the acoustic array when the acoustic positioning pulse signal arrives at the acoustic array. The calculation module 903 is used to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information; The second acquisition module 904 is used to acquire the relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located. The determination module 905 is used to determine the spatial position of the device to be located based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device.

[0183] The beneficial effects obtained by the above-described device are the same as those obtained by the above-described method, and will not be described in detail in the embodiments of this specification.

[0184] like Figure 10 The diagram shown is a structural schematic of a computer device according to an embodiment of this specification. The positioning method of this invention can be applied to the computer device in this embodiment. The computer device 1002 may include one or more processing devices 1004, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads.

[0185] Computer device 1002 may also include any storage resource 1006 for storing any kind of information such as code, settings, data, etc.

[0186] Non-limiting, for example, storage resource 1006 may include any one or more of the following: any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc.

[0187] More generally, any storage resource can use any technology to store information.

[0188] Furthermore, any storage resource can provide volatile or non-volatile retention of information.

[0189] Furthermore, any storage resource can represent a fixed or removable component of the computer device 1002.

[0190] In one scenario, when processing device 1004 executes associated instructions stored in any storage resource or combination of storage resources, computer device 1002 can perform any operation of the associated instructions. Computer device 1002 also includes one or more drive mechanisms 1008 for interacting with any storage resource, such as hard disk drive mechanisms, optical disk drive mechanisms, etc.

[0191] Computer device 1002 may further include an input / output module 1010 (I / O) for receiving various inputs (via input device 1012) and providing various outputs (via output device 1014). A specific output mechanism may include a presentation device 1016 and an associated graphical user interface (GUI) 1018. In other embodiments, the input / output module 1010 (I / O), input device 1012, and output device 1014 may be omitted, and the device may function solely as a computer device within a network. Computer device 1002 may also include one or more network interfaces 1020 for exchanging data with other devices via one or more communication links 1022. One or more communication buses 1024 couple the components described above together.

[0192] The communication link 1022 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. The communication link 1022 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0193] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0194] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the above-described method.

[0195] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0196] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the embodiments of this specification, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0197] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the embodiments in this specification.

[0198] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0199] In the embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0200] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments described in this specification, depending on actual needs.

[0201] Furthermore, the functional units in the various embodiments of this specification can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0202] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this specification, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, 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 specification. 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.

[0203] This specification describes the principles and implementation methods of the embodiments using specific examples. The above descriptions of the embodiments are only for the purpose of helping to understand the methods and core ideas of the embodiments in this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the embodiments in this specification. Therefore, the content of this specification should not be construed as a limitation on the embodiments in this specification.

Claims

1. A positioning method, characterized by, Applied to a device to be located, the device to be located includes an acoustic array, the method includes: Acquire the acoustic positioning pulse signal emitted by the target reference device; Acquire characteristic information representing the time difference between arrival of the acoustic positioning pulse signal to different array elements in the acoustic array; Based on the feature information, the relative spatial angle of arrival of the target reference device relative to the device to be located is calculated; The relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located, are obtained. The spatial position of the device to be located is determined based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device.

2. The method according to claim 1, characterized in that, The acquisition of feature information characterizing the arrival time difference between different array elements in the acoustic array of the acoustic positioning pulse signal includes: The acoustic positioning pulse signals received by each element in the acoustic array are subjected to baseband demodulation and matched filtering to obtain the matched filtered output signal of each element. For each array element, the matching peak position is determined from the matched filter output signal, and the signal is truncated based on the matching peak position to obtain the direct wave signal segment corresponding to each array element. The feature information is obtained based on the direct wave signal segment corresponding to each array element.

3. The method according to claim 1, characterized in that, The acoustic array includes a first acoustic subarray, and the step of calculating the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information includes: The array elements in the first acoustic subarray are combined to construct multiple sets of first array element connecting arm combinations, each first array element connecting arm combination containing two non-parallel first array element connecting arms. For each combination of first array element connecting arms, based on the arrival time difference of each first array element connecting arm under the combination, the candidate arrival angle corresponding to each combination of first array element connecting arms is calculated. For each combination of first array element connecting arms, the arrival time difference of each first array element connecting arm under the combination and the spatial angle between different first array element connecting arms are used to assign confidence weights to the first array element connecting arm combination. Using the candidate arrival angle as the direction and the confidence weight as the modulus, construct a three-dimensional vector corresponding to each combination of the first array element connecting arms; The three-dimensional vectors corresponding to the connecting arms of each of the first array elements are superimposed, and the relative spatial arrival angle is determined based on the direction of the superimposed resultant vector.

4. The method according to claim 3, characterized in that, The confidence weights are calculated based on the following formula: in, Indicates the arm connected by the first element. Second array element connecting arm The first element of the combination is the connecting arm combination. The corresponding confidence weights; Indicates the first element connecting the arms The corresponding arrival time difference; Indicates the second array element connecting arm The corresponding arrival time difference; Indicates the first element connecting the arms Second Array Link Arm The angle between them; Indicates the first prevention and elimination of zero items; This indicates the second prevention and control item.

5. The method according to claim 3 or 4, characterized in that, The acoustic array is a multi-scale acoustic array, comprising a first acoustic subarray and at least one second acoustic subarray; wherein the element spacing of the second acoustic subarray is greater than the element spacing of the first acoustic subarray; the method further includes: The relative spatial angle of arrival calculated based on the first acoustic subarray is used as the initial spatial angle of arrival. Based on all the elements of the first acoustic subarray and the second acoustic subarray, construct multiple second element connecting arms with different baseline lengths; Directional filtering is performed based on the initial spatial arrival angle to select candidate connecting arms from the plurality of second array element connecting arms whose spatial angle with the direction indicated by the initial spatial arrival angle is within a preset angle range; the preset angle range includes 90 degrees. Select a first preset number of first long baseline arms from the candidate connecting arms in descending order of baseline length. From the candidate connecting arms, a second preset number of second long baseline arms are selected in descending order of baseline length; the spatial angle between the second long baseline arm and the first long baseline arm is within the preset angle range; A long baseline arm combination is constructed based on the first long baseline arm and the second long baseline arm; each long baseline arm combination includes one first long baseline arm and one second long baseline arm. Based on the time difference of arrival corresponding to the long baseline arm combination, multiple candidate angles with phase ambiguity are calculated; an angle calculation window with a boundary span is set with the initial spatial angle of arrival as the center, and the target candidate angle falling within the angle calculation window is selected from the multiple candidate angles with phase ambiguity as the relative spatial angle of arrival without phase ambiguity; wherein, the boundary span of the angle calculation window is smaller than the phase ambiguity period generated by the long baseline arm combination.

6. The method according to claim 2, characterized in that, The target reference device is one of a plurality of reference devices, and the reference device is configured to simultaneously transmit wireless broadcast signals and acoustic positioning pulse signals; The step of performing baseband demodulation and matched filtering on the acoustic positioning pulse signals received by each element in the acoustic array to obtain the matched filtered output signal of each element specifically includes: Receive the wireless broadcast signal transmitted by the target reference device, and parse the wireless broadcast signal to obtain the transmission frequency band information of the target reference device; Based on the transmission band information, a corresponding template is selected as a local reference signal from a plurality of pre-configured matching signal templates; Using the local reference signal, the acoustic positioning pulse signals received by each array element in the acoustic array are subjected to baseband demodulation and matched filtering.

7. The method according to claim 6, characterized in that, The step of obtaining the relative distance between the device to be located and the target reference device includes: Receive the wireless broadcast signal transmitted by the target reference device, and use the time of receiving the wireless broadcast signal as the reference time; Based on the preset physical ranging boundary, determine the effective observation time window for the target reference device; Within the effective observation time window, feature matching is performed on the acoustic signals acquired by the acoustic array; If the time at which the matching peak position is determined to occur is within the effective observation time window, then the time at which the matching peak position is determined to occur is taken as the actual arrival time, and the relative distance is calculated based on the time difference between the actual arrival time and the reference time, combined with the ambient sound speed.

8. The method according to claim 2, characterized in that, The direct wave signal segment is a time-domain complex baseband signal sequence; the acquisition of the feature information based on the direct wave signal segment corresponding to each array element includes: From the direct wave signal segments corresponding to each array element, extract multiple time-domain signal sampling points containing the matching peak positions to form a two-dimensional complex matrix, which serves as the feature information; The step of calculating the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information includes: The two-dimensional complex matrix is ​​input into a pre-trained relative angle calculation network, and the relative angle calculation network is used to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located.

9. The method according to claim 8, characterized in that, The relative angle calculation network is a complex convolutional neural network, which includes cascaded complex feature extraction network layers and real fully connected mapping layers. The step of using the relative angle calculation network to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located includes: In the complex feature extraction network layer, the two-dimensional complex matrix is ​​subjected to multi-layer complex two-dimensional convolution operations in sequence, combined with nonlinear activation processing of complex batch normalization, to extract complex feature maps; wherein, the nonlinear activation processing is configured to perform activation calculation on the magnitude of the complex number while preserving the original phase features of the complex number; The final output of the complex feature extraction network layer is moduloed to generate real parameters, which are then compressed into a real feature vector through global average pooling. The real-valued feature vector is input into the real-valued fully connected mapping layer. After fully connected mapping and dimensionality reduction, the relative spatial angle of arrival of the target reference device relative to the device to be located is obtained.

10. A positioning device, characterized in that, Applied to a device to be located, the device including an acoustic array, the device comprising: The acquisition module is used to acquire the acoustic positioning pulse signal emitted by the target reference device; The first acquisition module is used to acquire feature information characterizing the arrival time difference between different array elements in the acoustic array of the acoustic positioning pulse signal; The calculation module is used to calculate the relative spatial angle of arrival of the target reference device relative to the device to be located based on the feature information; The second acquisition module is used to acquire the relative distance between the device to be located and the target reference device, as well as the device attitude parameters of the device to be located. The determination module is used to determine the spatial position of the device to be located based on the device attitude parameters, the relative spatial angle of arrival, the relative distance, and the known position of the target reference device.

11. A positioning device, characterized in that, Includes the apparatus as described in claim 10.

12. An electronic device, characterized in that, It includes at least one processor and a memory; the memory stores a computer program or instruction set; at least one of the processors executes the computer program or instruction set stored in the memory, such that at least one of the processors performs the method as described in any one of claims 1 to 9.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program or instruction set, which, when executed by a processor, implements the method as described in any one of claims 1 to 9.

Images