Acoustic motion capture method and system based on receiver synchronization mechanism

Abstract

The present application relates to the technical field of action capture and spatial positioning, in particular to an acoustic action capture method and system based on a receiving end synchronization mechanism. The core innovation is to build a system architecture with the receiving end as the unified time reference, completely eliminating the asynchronous error of the traditional transmitting end synchronization architecture. To solve the problems of difficult determination of multiple transmitting node signals at the same time and signal aliasing of multiple sources, a unified time calibration mechanism of receiving end synchronization is adopted, combined with a hybrid signal scheduling scheme of grouping time division multiplexing and intra-group code division multiplexing. The system includes a coordinator, an ultrasonic transmitting terminal, an acoustic receiving array and other components. The coordinator provides a global clock through the IEEE 1588v2 protocol and PPS signal, the transmitting terminal transmits ultrasonic signals with a unique pseudo-random noise code, the receiving array separates the signals through parallel matched filtering and records accurate time stamps, and finally realizes high-precision positioning and attitude restoration through the time difference of arrival algorithm, environmental parameter correction and inverse kinematics model.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of motion capture and spatial positioning technology, in particular to an acoustic motion capture method and system based on a receiving end synchronization mechanism. BACKGROUND

[0002] Human motion capture technology is the core support in the fields of virtual reality, motion analysis and medical rehabilitation, and the key is to realize the real-time and accurate reconstruction of the spatial pose of the motion target. The current mainstream technical solutions mainly fall into two categories: one is a motion capture system based on visual recognition, which can achieve high accuracy under ideal conditions, but is extremely sensitive to environmental occlusion, light changes and device calibration, and the system deployment is complex and costly; the other is a solution based on wearable sensors, which is flexible in deployment and not limited by the field of view, but has inherent integral drift error and cannot provide absolute spatial coordinates, and needs to rely on external reference for periodic calibration.

[0003] Acoustic positioning technology has become a potential technical route in the field of motion capture due to its unique advantages of anti-visual occlusion, low cost and ability to provide absolute coordinates. Traditional acoustic motion capture systems usually adopt a transmitting end synchronization architecture, which requires strict time synchronization among multiple acoustic beacons, and the receiving array performs positioning calculation based on the time difference of signal arrival, but this architecture has significant limitations: first, the wireless synchronization between mobile beacons is easily affected by multipath effect and clock drift, making it difficult to maintain microsecond-level synchronization accuracy; second, the signals transmitted by different beacons in a multi-target scenario are easily mixed, and existing coding strategies cannot effectively distinguish them, resulting in serious code interference; third, the speed of sound propagation is significantly affected by changes in environmental temperature and humidity, and multipath interference and background noise can introduce arrival time judgment errors.

[0004] Existing improvement solutions mainly focus on local optimization and fail to solve the coupled problems of insufficient transmitting end synchronization reliability and multi-source signal correlation ambiguity from the system architecture level, resulting in the difficulty of existing acoustic motion capture systems to meet the needs of high precision, low delay and multi-target expansion in complex application scenarios. Currently, there is no acoustic positioning solution that uses receiving end time synchronization as the system time reference and combines grouping time division and code division multiplexing mechanisms to achieve multi-source concurrent positioning. SUMMARY

[0005] The present application aims to provide an acoustic motion capture method and system based on a receiving end synchronization mechanism, which solves the problems of chaotic management of multiple transmitting sources, inconsistent time reference leading to low positioning accuracy, and difficulty in distinguishing signals in the prior art.

[0006] The technical solution adopted by the present application to solve its technical problems is: an acoustic motion capture method based on a receiving end synchronization mechanism, comprising the following steps:

[0007] S1. System synchronization step: the coordinator acts as the time reference source of the system, and establishes a global unified high-precision time reference for the acoustic receiving array by using the precision time protocol; the microsecond-level time consistency between the nodes in the receiving array is realized by periodically issuing time synchronization messages and cooperating with the pulse per second signal;

[0008] S2. Synchronization state judgment step: monitor the system synchronization state, and enable the degraded synchronization mode to maintain the basic synchronization function of the system when the synchronization failure is detected;

[0009] S3. Instruction configuration step: the coordinator allocates a unique transaction identifier for each acoustic emission event according to the task scheduling requirement, and generates a trigger instruction containing the target emission time, signal type, identity code and power control parameter; the trigger instruction is sent to the corresponding acoustic emission terminal through the wireless link to realize the unified time scheduling and identification of multiple emission sources;

[0010] S4. Signal emission step: after receiving the trigger instruction, the acoustic emission terminal aligns the local clock according to the target emission time, and emits the acoustic signal carrying the identity code at the predetermined time; the acoustic signal uses an ultrasonic carrier wave, and can be modulated and encoded by a linear frequency modulation or a pseudo-random sequence; wherein the acoustic emission terminal is divided into multiple groups, and the groups sequentially emit in a time division multiplexing manner, and the multiple emission terminals in the group concurrently emit in a code division multiplexing manner, and each emission terminal is allocated a unique pseudo-random noise code as the identity code;

[0011] S5. Signal receiving step: each receiving node of the acoustic receiving array operates synchronously under the unified time reference, monitors the spatial acoustic signal in real time, and records the absolute time of arrival of each acoustic pulse through a hardware-level time stamping circuit; the receiving node collects the received mixed acoustic signal, and processes it by using a parallel matched filtering technology, performs cross-correlation operation on the mixed signal by using all possible identity code templates stored in advance, and separates and identifies the signals of different emission terminals by detecting the significant peaks in the cross-correlation output;

[0012] S6. Data association step: the processing device associates the signals from different receiving nodes at the event level based on the transaction identifier and the absolute time stamp; when a signal group with the same identity code and within a preset fault tolerance time window is detected, it is classified as the same emission event;

[0013] S7. Positioning solution step: for the associated same emission event, a spatial positioning algorithm based on time difference of arrival is used to calculate the coordinates of the sound source in three-dimensional space; the algorithm combines the sound wave propagation speed and the spatial geometric distribution of the receiving nodes to perform least squares fitting, and introduces an environmental temperature correction model in real time;

[0014] S8. Pose restoration step: input the spatial coordinate sequence of multiple sound sources into the inverse kinematics model, obtain the skeletal pose parameters of the human body or object through constraint solving, including joint rotation angle, quaternion pose and spatial pose change, realize continuous motion capture and three-dimensional motion reconstruction.

[0015] Specifically, the coordinator in S1 acts as a precision time protocol master clock, communicates with the receiving array through Ethernet or optical fiber network, and periodically sends time synchronization messages, while providing physical pulse per second signals for time reference calibration to the receiving array, so that the clock drift error is maintained within the microsecond level for a long time.

[0016] Specifically, the degraded synchronization mode enabled in S2 is a bidirectional ranging mode between receiving nodes, which maintains system synchronization by measuring relative time deviation.

[0017] Specifically, the coordinator in S3 issues trigger instructions to the transmitting terminal through ultra-wideband or Bluetooth low power communication mode; in addition to containing transaction identifier and target transmission time, the instructions can also contain transmission signal type, acoustic encoding parameters and node identity information.

[0018] Specifically, the transmitting terminal in S4 includes a high-precision timer, a digital signal modulator, and an ultrasonic transducer driving unit, and the transmission signal frequency range is 35 kHz to 45 kHz; the consistency and repeatability of sound wave transmission are realized through pre-emphasis filtering and power compensation.

[0019] Specifically, the acoustic receiving node in S5 contains a hardware timestamp recording circuit, an analog-to-digital conversion module, and a local cache module; the timestamp recording is realized based on envelope peak detection of the received signal, and is timed with a global synchronous clock.

[0020] Specifically, the data association in S6 further includes: analyzing the identity encoding information in the acoustic signal to distinguish signals belonging to different transmitting terminals; when a signal set with the same identity encoding and an absolute arrival time difference less than a fault tolerance time window is detected, it is determined as the same event signal group.

[0021] Specifically, the fault tolerance time window is dynamically adjusted according to the system synchronization accuracy and environmental noise, and the range can be configured to ±1 microsecond to ±50 microseconds; when the synchronization link drifts, the window threshold is corrected by an adaptive algorithm.

[0022] Specifically, the multi-node arrival time difference positioning algorithm is used in S7, and the calculation process includes: extracting the sound wave arrival time difference of different receiving nodes, constructing the positioning equation set based on the spherical wave propagation model, and solving the three-dimensional coordinates of the sound source using least squares or Kalman filtering algorithm; real-time temperature, humidity and air pressure parameters are introduced to correct the sound speed model.

[0023] Specifically, the sound speed model is based on an empirical formula where T is the ambient temperature, H is the air humidity, the temperature and humidity parameters are collected by real-time sensors and the calculation results are corrected.

[0024] Specifically, the pose restoration step in S8 is based on a human joint hierarchical structure model, and the joint rotation angle and pose vector are solved by an inverse kinematics algorithm; the algorithm can use a constrained least squares method or a pose fusion filtering method.

[0025] Specifically, the pseudo-random noise code length used in the code division multiplexing in S4 is 31 to 127 bits, the chip rate is set to 2 to 10 kHz, and the signal duration is in the range of 3 to 20 milliseconds.

[0026] Specifically, the parallel matched filter processing in S5 is implemented on an FPGA or a dedicated digital signal processor, and the mixed signal of each receiving channel is simultaneously correlated with all identity code templates, and a constant false alarm rate detection algorithm is used to adaptively set the peak detection threshold.

[0027] An acoustic motion capture system for implementing the acoustic motion capture method based on the receiving end synchronization mechanism, comprising:

[0028] A coordinator module for establishing a global unified time reference and generating a transmission task instruction;

[0029] An acoustic transmission terminal for receiving the instruction and transmitting an acoustic signal carrying an identity code at a specified time;

[0030] An acoustic receiving array composed of multiple receiving nodes with hardware timestamp function, for receiving acoustic signals and recording their absolute arrival times;

[0031] A signal conditioning circuit for amplifying and filtering the received signals;

[0032] A signal acquisition circuit for synchronously acquiring multiple acoustic signals; a time synchronization module for providing a unified time reference for the system;

[0033] A positioning solution module for calculating the coordinates of the sound source based on the time difference of arrival algorithm;

[0034] A pose restoration module for pose reconstruction processing of the positioning results;

[0035] An environmental sensing unit for providing real-time temperature and humidity parameters to correct the sound speed model.

[0036] Specifically, the coordinator module comprises a precision time protocol master clock unit and a wireless communication unit; the master clock unit is used to issue precision time protocol messages through a wired network to establish network time synchronization; and the wireless communication unit is used to issue trigger instructions to acoustic emission terminals.

[0037] Specifically, the acoustic emission terminal comprises a high-precision timing control unit, a signal encoding circuit and an ultrasonic transducer; the signal encoding circuit stores a pseudo-random noise code sequence, which is used to modulate an ultrasonic carrier to generate an acoustic signal carrying a unique identity code.

[0038] Specifically, the receiving nodes in the acoustic receiving array comprise a hardware timestamp unit and a matched filter processing unit; the hardware timestamp unit is used to record the absolute time of arrival of acoustic waves with a global time reference; and the matched filter processing unit is used to separate and identify signals emitted by different emission terminals through parallel cross-correlation operations.

[0039] Specifically, the positioning calculation module adopts a time difference of arrival algorithm and can receive environmental sensor data to compensate for the sound speed model in real time; and the attitude restoration module converts the acoustic source coordinate sequence into skeletal attitude data through an inverse kinematics algorithm based on a human skeletal model.

[0040] Specifically, the signal acquisition circuit realizes multi-channel synchronous sampling through FPGA control of multiple ADC chips, and the sampling rate is not less than 192 kHz; and the signal conditioning circuit comprises a programmable gain amplifier and an anti-aliasing filter.

[0041] A non-volatile storage medium storing a computer program, which, when executed by a processor, performs all steps of the acoustic motion capture method based on the receiver synchronization mechanism.

[0042] The present application has the following advantages:

[0043] The global time synchronization mechanism based on PTP is adopted, the coordinator realizes microsecond-level time consistency among receiver nodes through precision time protocol message and PPS signal calibration, and the time difference of arrival measurement precision and positioning precision of acoustic sources are greatly improved.

[0044] Through the mechanism combining group time division multiplexing and intra-group code division multiplexing, a unique pseudo-random noise code identity code is added to each emission signal, so that multiple acoustic emission sources can work concurrently in the group and the signals do not interfere with each other, effectively solving the multi-source signal aliasing problem.

[0045] The event correlation method combining hardware timestamp recording and adaptive fault-tolerant time window can accurately match the same emission event from different nodes in the signal arrival timing, and improve the reliability of data association.

[0046] The real-time introduction of environmental temperature, humidity and other parameters in the positioning calculation corrects the sound velocity model, effectively reduces the influence of environmental changes on the positioning result, and further improves the three-dimensional positioning accuracy.

[0047] The inverse kinematics posture restoration algorithm based on the human body skeletal hierarchy can output stable and accurate skeletal joint angles and posture changes, realizes continuous three-dimensional motion capture, and enhances the real-time performance and robustness of the system.

[0048] The receiving end uses parallel matched filtering technology to process mixed acoustic signals, effectively separates and identifies signals of different transmitting terminals through cross-correlation operation, and significantly improves the robustness and multi-target resolution capability of the system in a complex sound field environment. BRIEF DESCRIPTION OF DRAWINGS

[0049] The application will be further described below in combination with the drawings and embodiments.

[0050] Figure 1 : The system overall architecture and data flow diagram of the embodiment one of the application;

[0051] Figure 2 : The synchronization mechanism flow chart of the embodiment one of the application;

[0052] Figure 3 : The signal transmission, propagation and reception sequence diagram of the embodiment one of the application;

[0053] Figure 4 : The data association logic judgment diagram of the embodiment one of the application;

[0054] Figure 5 : The sound source positioning effect diagram based on TDOA of the embodiment one of the application;

[0055] Figure 6 : The TDOA three-dimensional positioning principle schematic diagram based on hyperboloid intersection of the embodiment one of the application;

[0056] Figure 7 : The positioning solution algorithm flow chart of the embodiment one of the application;

[0057] Figure 8 : The posture restoration architecture diagram of the embodiment one of the application;

[0058] Figure 9 : The acoustic motion capture system module and working process schematic diagram of the embodiment two of the application. DETAILED DESCRIPTION

[0059] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the application will be further described below in combination with specific embodiments.

[0060] The application discloses an acoustic motion capture method based on a receiving end synchronization mechanism.

[0061] S1: system synchronization step: the coordinator serves as a time reference source of the system, and a precision time protocol (PTP) is used to establish a globally unified high-precision time reference for the acoustic receiving array; the coordinator communicates with the receiving array through an Ethernet or optical fiber network, periodically issues a time synchronization message, and outputs a physical pulse per second (PPS) signal for time reference calibration, so that the clock long-term drift error between nodes in the receiving array is maintained within a microsecond range.

[0062] S2: synchronization state judgment step: when it is monitored that synchronization fails or the master-slave clock is abnormal, a degraded synchronization mode is enabled; the relative time deviation of each receiving node is measured through a bidirectional ranging mode, so that the basic synchronization function of the system is maintained, and the whole synchronization failure caused by single-point clock failure is prevented.

[0063] S3: instruction configuration step: the coordinator allocates a unique transaction identifier for each acoustic emission event according to task scheduling requirements, and generates a trigger instruction containing target emission time, emission signal type, identity coding and power control parameters; the instruction is sent to the corresponding acoustic emission terminal through an ultra-wideband (UWB) or Bluetooth low power consumption (BLE) wireless communication mode, so that unified time scheduling and node identification management for multiple emission source parallel emission are realized.

[0064] S4: signal emission step: after receiving the trigger instruction, the acoustic emission terminal aligns the local clock according to the target emission time, and emits an acoustic signal carrying the identity coding at the predetermined moment; the acoustic signal takes an ultrasonic wave with a frequency of 35 kHz to 45 kHz as a carrier, multiple acoustic emission terminals are divided into several groups, time division multiplexing is used between the groups to sequentially emit, and code division multiplexing is used in the terminals in the group to concurrently emit; each emission terminal is allocated a unique pseudo-random noise code (preferably a Gold code) as the identity coding, the code length ranges from 31 bits to 127 bits, and the ultrasonic carrier is modulated through binary phase shift keying by using the code.

[0065] S5: signal receiving step: each receiving node in the acoustic receiving array is synchronously operated under the global time reference, and the spatial acoustic signal is monitored in real time; when the acoustic pulse is detected, the envelope peak arrival time of the pulse is marked through a hardware time stamp recording circuit, and the timing accuracy is better than 10 microseconds; after the receiving node collects the mixed acoustic signal, parallel matched filtering technology is used, all identity coding templates are stored in advance and are correlated with the mixed signal, and the signals of different emission terminals are separated and identified by detecting significant peaks.

[0066] S6: Data association step: the processing device associates the signals from different receiving nodes at the event level based on the transaction identifier and the absolute timestamp; first, the signals of different transmitting terminals are distinguished by identity coding information, and then the signals with the same identity coding and the time difference within the preset fault tolerance time window (±1 microsecond to ±50 microseconds) are determined as the same transmission event; the fault tolerance time window can be dynamically adjusted according to the system synchronization accuracy and environmental noise.

[0067] S7: Positioning solution step: for the associated same transmission event, the three-dimensional coordinates of the sound source are calculated based on the time difference of arrival recorded by each receiving node using the multi-node time difference of arrival (TDOA) positioning algorithm; the positioning equation set is constructed according to the sound wave propagation speed and the spatial geometric distribution of the receiving nodes, and the least square method or Kalman filtering algorithm is used for solving; the sound speed model is corrected by using the empirical formula c=331.3+0.606T+0.0124H (T is the environmental temperature, and H is the air humidity) to eliminate the positioning error caused by the change of environmental conditions.

[0068] S8: Attitude restoration step: the three-dimensional spatial coordinate sequence of multiple sound sources is input into the inverse kinematics model based on the hierarchical structure of human skeleton, and the skeletal attitude parameters of the human body or object are calculated by constraint optimization to solve the rotation angle and quaternion attitude vector of each joint, so that the continuous three-dimensional motion attitude is obtained; the constraint least square method or attitude fusion filtering algorithm can be used to realize stable fusion of multi-source coordinate data.

[0069] The acoustic action capture system for realizing the above-mentioned acoustic action capture method based on the receiving end synchronization mechanism comprises a coordinator module, an acoustic transmitting terminal, an acoustic receiving array, a signal conditioning circuit, a signal acquisition circuit, a time synchronization module, a positioning solution module and an attitude restoration module, and specifically as follows:

[0070] The coordinator module: as the system control core, it contains a precision time protocol master clock unit and a wireless communication unit; the master clock unit issues synchronization messages through an Ethernet or optical fiber network, and the wireless communication unit communicates with the transmitting terminal in the UWB or Bluetooth low power consumption mode, which is used to establish a global unified time reference and assign a unique transaction identifier to each acoustic transmission event.

[0071] The acoustic transmitting terminal: it contains a high-precision timer, a signal coding circuit and an ultrasonic transducer array, and stores a Gold code sequence (code length 31-127 bits), which generates coded ultrasonic signals with a center frequency of 40 kHz±5 kHz through BPSK modulation; it is used to receive the trigger instruction issued by the coordinator and transmit ultrasonic signals carrying a unique identity code at a specified time.

[0072] Acoustic receiving array: composed of multiple spatially distributed receiving nodes, each node containing an analog front-end, an ADC conversion module, a hardware timestamp unit and a matched filter processing unit; the matched filter unit separates different transmitting source components in the mixed signal through parallel correlation operation, and the correlation peak detection threshold is determined by a CFAR adaptive algorithm; each node is equipped with a hardware timestamp function for synchronously receiving acoustic signals and recording their absolute arrival times with a precision better than 10 microseconds.

[0073] Signal conditioning circuit: connected after the receiving array, containing a programmable gain amplifier and an anti-aliasing filter, for gain adjustment and anti-aliasing filtering of the received analog signals.

[0074] Signal acquisition circuit: through FPGA control multiple ADCs to realize synchronous sampling with a sampling rate not less than 192 kHz, for synchronous sampling and digitization of the conditioned multi-channel signals.

[0075] Time synchronization module: integrated in each component of the system, for maintaining the microsecond-level time consistency of the entire system.

[0076] Positioning solution module: based on the time difference of arrival data of each receiving node, using the TDOA algorithm to calculate the three-dimensional spatial coordinates of the sound source; capable of receiving environmental sensor data for real-time compensation of sound velocity.

[0077] Attitude restoration module: based on the human body skeletal hierarchical model, converting the coordinate sequence of multiple sound sources into continuous human body skeletal attitude data through inverse kinematics algorithm, and realizing stable fusion of multi-source coordinates using constrained least squares method.

[0078] Embodiment one: acoustic motion capture method based on receiving end synchronization

[0079] The implementation process of the method will be described in detail below with reference to the accompanying Figures 1 to 7

[0080] S1: System synchronization and global time reference establishment step After the coordinator is powered on, run the Precision Time Protocol (PTP) and start as the master clock. Periodically send time synchronization messages (Sync, Follow Up messages) to all nodes in the acoustic receiving array through a wired network (such as Ethernet), and generate and output a physical pulse per second (PPS) signal to the receiving array. The rising edge of the PPS signal is strictly aligned with the second clock boundary of the coordinator. The receiving nodes process the PTP messages and combine the PPS signal for hardware interrupt calibration, synchronize the local clock with the coordinator master clock, ensure that the clock deviation of the receiving array is stable within microseconds in the long term, and establish a global unified high-precision time reference. The synchronization deviation is checked by comparing the rising edge time of the PPS signal, and the long-term error is not more than ±2 microseconds. ​

[0081] S2: Synchronization state judgment and fault handling step The system continuously monitors the synchronization state between the coordinator and the receiving nodes, and the judgment module checks whether the PTP message reception is continuous and the PPS signal is normal in real time, and evaluates whether the clock deviation exceeds the threshold. If the synchronization signal is normal, the normal PTP master-slave synchronization mode is maintained; if the synchronization signal is abnormal, the degraded synchronization mode is automatically triggered, and bidirectional ranging is performed between each receiving node through a wireless link (such as UWB) to measure the round-trip time (Round-Trip Time) to calculate the clock deviation and compensate for the transmission delay, and an auxiliary time reference is established to enable the system to maintain basic synchronization functions even if the main clock reference is lost.

[0082] S3: Instruction configuration and issuing step The coordinator assigns a globally unique transaction identifier (such as TxID=T001) to each acoustic signal transmission event according to the motion capture task plan; divides all acoustic transmission terminals (PUCK) into G groups (such as G=2 or 3), and assigns each PUCK a unique pseudo-random noise code (preferably Gold code) as an identity code; generates a trigger instruction frame containing parameters such as target absolute transmission time, transaction identifier, group membership, and identity code information, and issues it to the designated acoustic transmission terminal through a ultra-wideband (UWB) or Bluetooth Low Energy (BLE) wireless communication link.

[0083] S4: Signal transmission and spatial propagation step The acoustic transmission terminal receives the trigger instruction and performs analysis, the system uses a hybrid scheme of group time division multiplexing (TDM) combined with intra-group code division multiplexing (CDM), PUCKs of different groups transmit in different time slots in sequence, and multiple PUCKs in the same group transmit concurrently in the same time slot; the high-precision timer inside the terminal performs local clock calibration and countdown according to the target absolute transmission time, starts the signal transmission circuit at the specified time, uses the assigned unique pseudo-random code to perform binary phase shift keying (BPSK) modulation on the ultrasonic carrier (preferably 40 kHz), and the modulated coded ultrasonic signal is transmitted through the ultrasonic transducer, propagates in space at the speed of sound, and linearly superimposes to form a mixed signal.

[0084] S5: Signal receiving, decoding and time stamp recording step Each node in the acoustic receiving array continuously monitors acoustic signals, receives and collects mixed signals formed by the superposition of sound waves emitted by multiple PUCKs; the collected mixed signals are subjected to parallel matched filtering processing (implemented on an FPGA, with a processing delay of less than 1 millisecond), all identity encoding templates are stored in the node, and each encoding template is used in parallel to perform cross-correlation operation with the mixed signals; the peak value of the cross-correlation output of each encoding template is detected to determine the detected PUCK signal and identify its identity; once a valid peak value is detected, a hardware time stamp circuit is triggered immediately to record the absolute time (with a precision better than 10 microseconds) at which the correlation peak occurs with reference to a global synchronous clock; the receiving node encapsulates and uploads data such as transaction identifier, PUCK identity code, absolute arrival time stamp, etc. to the data processing center.

[0085] S6: Data association and event clustering step After the data processing device receives the data from each receiving node, the signals are grouped according to the identity code based on the identity code; for a group of signals with the same identity code, the absolute arrival time stamp is extracted and the time difference is calculated to determine whether the time difference is within a preset fault tolerance time window (such as ±10 microseconds); if it is within the window, these signals are determined to be the same transmission event and are clustered to form a complete data set which is sent to the TDOA calculation module.

[0086] S7: Positioning solution step For the same transmission event after association, the time difference of arrival data (TDOA) is obtained, a positioning equation set is constructed, environmental parameters (temperature T, humidity H) are collected in real time, and the sound speed model (v = 331.3 + 0.606T + 0.0124H) is used to correct the sound speed value; the least squares method or Kalman filter algorithm is used to solve the three-dimensional coordinates of the sound source, ensuring the uniqueness of the solution under the condition of symmetrical arrangement of the receiving nodes, avoiding positioning ambiguity, and performing optimized filtering.

[0087] S8: Pose restoration step The three-dimensional coordinate sequence of multiple sound sources is input into an inverse kinematics model, which is based on an N-joint simplified skeletal structure, and the joint constraints are obtained through human body calibration. After steps such as skeletal mapping, constraint application and inverse solution, continuous human motion data is finally output.

[0088] Embodiment two: Acoustic motion capture system

[0089] The acoustic motion capture system provided by the embodiment adopts a modular design, and each module has clear functions and works cooperatively. The specific work flow is as follows:

[0090] The coordinator module first establishes and maintains the time synchronization of the whole system, groups the acoustic emission terminals, and assigns a unique pseudo-random identity code to each terminal; according to the motion capture task, the trigger instruction is issued to the designated acoustic emission terminal; each acoustic emission terminal concurrently emits the coded acoustic wave modulated by the unique pseudo-random code at the precise time specified by the instruction; the acoustic receiving array synchronously receives the mixed signal formed by the superposition of these acoustic waves under the unified time reference, separates and identifies the signal components of each emission terminal through the matched filter processing unit, and records the accurate arrival time of each signal component; the signal conditioning circuit and the signal acquisition circuit pre-process and digitize the signal; the data with time stamp are sent to the data processing device, the positioning calculation module calculates the coordinates of each sound source, and finally the attitude restoration module generates the continuous skeletal attitude action sequence.

[0091] The realizability of the system has been verified by experiments, the receiving node is realized based on an STM32 or FPGA platform, a 192kHz sampling rate and a 40kHz ultrasonic carrier are used, and actual tests show that the positioning error is less than 5mm, proving the practicability and high precision of the system.

[0092] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. An acoustic motion capture method based on a receiver synchronization mechanism, characterized in that, Includes the following steps: S1. System synchronization steps: The coordinator, as the system's time reference source, uses a precision time protocol to establish a globally unified high-precision time reference for the acoustic receiving array; by periodically sending time synchronization messages and cooperating with a pulse per second signal, microsecond-level time consistency among the nodes in the receiving array is achieved. S2. Synchronization status judgment steps: Monitor the system synchronization status. When a synchronization failure is detected, enable the degraded synchronization mode to maintain the basic synchronization function of the system. S3. Command Configuration Steps: Based on task scheduling requirements, the coordinator assigns a unique transaction identifier to each acoustic transmission event and generates a trigger command containing the target transmission time, signal type, identification code, and power control parameters; the trigger command is sent to the corresponding acoustic transmitting terminal via a wireless link to achieve unified time scheduling and identification of multiple transmission sources. S4. Signal Transmission Steps: After receiving the trigger command, the acoustic transmitting terminal aligns its local clock according to the target transmission time and transmits an acoustic signal carrying an identification code at a predetermined time. The acoustic signal uses an ultrasonic carrier and can be encoded using linear frequency modulation or pseudo-random sequence modulation. The acoustic transmitting terminal is divided into multiple groups, and the groups transmit sequentially using time division multiplexing. Multiple transmitting terminals within a group transmit concurrently using code division multiplexing. Each transmitting terminal is assigned a unique pseudo-random noise code as its identification code. S5. Signal Reception Steps: Each receiving node of the acoustic receiving array operates synchronously under a unified time reference, monitors spatial acoustic signals in real time, and records the absolute arrival time of each acoustic pulse through a hardware-level timestamp circuit; the receiving nodes acquire the received mixed acoustic signals and process them using parallel matched filtering technology, and perform cross-correlation calculations with the mixed signals using all pre-stored possible identity coding templates, and separate and identify the signals of different transmitting terminals by detecting significant peaks in the cross-correlation output; S6. Data association step: The processing device performs event-level association on signals from different receiving nodes based on transaction identifiers and absolute timestamps; when a group of signals with the same identity code and a time difference within a preset fault-tolerant time window is detected, they are classified as the same transmission event; S7. Positioning calculation steps: For the same transmission event that is associated, the spatial positioning algorithm based on the time difference of arrival is used to calculate the coordinates of the sound source in three-dimensional space; the algorithm combines the sound wave propagation speed and the spatial geometric distribution of the receiving node to perform least squares fitting, and introduces the ambient temperature correction model in real time; S8. Posture Reconstruction Step: Input the spatial coordinate sequence of multiple sound sources into the inverse kinematics model, and obtain the skeletal posture parameters of the human body or object through constraint solution, including joint rotation angles, quaternion postures and spatial pose changes, to achieve continuous motion capture and three-dimensional motion reconstruction.

2. The acoustic motion capture method based on receiver synchronization mechanism according to claim 1, characterized in that: In S1, the coordinator acts as the master clock for the precision time protocol. It communicates with the receiving array via Ethernet or fiber optic network and periodically sends time synchronization messages. At the same time, it provides the receiving array with a physical pulse signal per second for time reference calibration. The clock drift error is maintained within the microsecond range for a long time. The downgraded synchronization mode enabled in S2 is a two-way ranging mode between receiving nodes, which maintains system synchronization by measuring the relative time deviation. In S3, the coordinator sends a trigger command to the transmitting terminal via ultra-wideband or Bluetooth low-power communication. In addition to the transaction identifier and target transmission time, the command may also include the transmission signal type, acoustic coding parameters, and node identity information. The transmitting terminal in S4 includes a high-precision timer, a digital signal modulator, and an ultrasonic transducer drive unit, with a transmission signal frequency range of 35kHz to 45kHz; the consistency and repeatability of the sound wave transmission are achieved through pre-emphasis filtering and power compensation. The acoustic receiving node in S5 includes a hardware timestamp recording circuit, an analog-to-digital conversion module, and a local cache module. The timestamp recording is achieved based on the envelope peak detection of the received signal and is timed using a globally synchronized clock; The data association in S6 further includes: parsing the identity coding information in the acoustic signal to distinguish signals belonging to different transmitting terminals; when a set of signals with the same identity coding and an absolute arrival time difference less than the fault tolerance time window is detected, they are determined to be the same event signal group. The S7 employs a multi-node time-of-arrival (TOA) positioning algorithm. The calculation process includes: extracting the TOA of sound waves from different receiving nodes, constructing a set of positioning equations based on the spherical wave propagation model, and using the least squares or Kalman filter algorithm to solve for the three-dimensional coordinates of the sound source; and introducing real-time temperature, humidity, and air pressure parameters to correct the sound speed model. The posture restoration step in S8 is based on the human joint hierarchical structure model and solves the joint rotation angle and pose vector through inverse kinematics algorithm; the algorithm can adopt constrained least squares method or posture fusion filtering method.

3. The acoustic motion capture method based on the receiver synchronization mechanism according to claim 2, characterized in that: The fault-tolerant time window is dynamically adjusted according to the system synchronization accuracy and environmental noise, and can be configured to range from ±1 microsecond to ±50 microsecond; when the synchronization link drifts, the window threshold is corrected by an adaptive algorithm.

4. The acoustic motion capture method based on the receiver synchronization mechanism according to claim 2, characterized in that: The sound speed model is based on empirical formulas. Where T is the ambient temperature and H is the air humidity, the temperature and humidity parameters are collected by real-time sensors and the calculation results are corrected.

5. The acoustic motion capture method based on receiver synchronization mechanism according to claim 1, characterized in that: The pseudo-random noise code length used in the intra-group code division multiplexing in S4 is 31 to 127 bits, the chip rate is set to 2 kHz to 10 kHz, and the signal duration is 3 to 20 milliseconds.

6. The acoustic motion capture method based on receiver synchronization mechanism according to claim 1, characterized in that: The parallel matched filtering process in S5 is implemented on an FPGA or a dedicated digital signal processor. The mixed signal of each receiving channel is cross-correlated with all identity coding templates simultaneously, and a constant false alarm rate detection algorithm is used to adaptively set the peak detection threshold.

7. An acoustic motion capture system for implementing the acoustic motion capture method based on a receiver synchronization mechanism as described in any one of claims 1-6, characterized in that, include: The coordinator module is used to establish a globally unified time base and generate launch mission commands; An acoustic transmitting terminal is used to receive the instruction and transmit an acoustic signal carrying an identity code at a specified time; An acoustic receiver array consists of multiple receiver nodes with hardware timestamp functionality, used to receive acoustic signals and record their absolute arrival times; Signal conditioning circuitry is used to amplify and filter the received signal; The signal acquisition circuit is used to synchronously acquire multiple acoustic signals; the time synchronization module is used to provide a unified time reference for the system. The localization calculation module is used to calculate the coordinates of the sound source based on the time difference of arrival algorithm; The attitude reconstruction module is used to reconstruct the attitude of the localization results. An environmental sensing unit is used to provide a sound velocity model that corrects for real-time temperature and humidity parameters.

8. The acoustic motion capture system according to claim 7, characterized in that: The coordinator module includes a precision time protocol master clock unit and a wireless communication unit; the master clock unit is used to send precision time protocol messages through a wired network to establish network time synchronization; the wireless communication unit is used to send trigger commands to the acoustic transmitting terminal. The acoustic transmitting terminal includes a high-precision timing control unit, a signal encoding circuit, and an ultrasonic transducer; the signal encoding circuit stores a pseudo-random noise code sequence, which is used to modulate the ultrasonic carrier to generate an acoustic signal carrying a unique identification code. The receiving node in the acoustic receiving array includes a hardware timestamp unit and a matched filter processing unit; the hardware timestamp unit is used to record the absolute time of arrival of the sound wave with a global time reference; the matched filter processing unit is used to separate and identify signals transmitted by different transmitting terminals through parallel cross-correlation operations. The positioning and calculation module uses the time difference of arrival algorithm and can receive environmental sensor data to compensate the sound speed model in real time; the attitude restoration module is based on the human skeleton model and converts the sound source coordinate sequence into skeleton attitude data through inverse kinematics algorithm.

9. The acoustic motion capture system according to claim 7, characterized in that: The signal acquisition circuit uses an FPGA to control multiple ADC chips to achieve multi-channel synchronous sampling with a sampling rate of not less than 192kHz; the signal conditioning circuit includes a programmable gain amplifier and an anti-aliasing filter.

10. A non-volatile storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it performs all the steps of the acoustic motion capture method based on the receiver synchronization mechanism as described in any one of claims 1-6.

Images