Search

CN-122027927-A - Optimizing system of intelligent sound equipment

CN122027927ACN 122027927 ACN122027927 ACN 122027927ACN-122027927-A

Abstract

The invention discloses an intelligent sound optimizing system, which relates to the technical field of intelligent audio, and comprises distributed intelligent sound nodes and intelligent decision modules on a main control node, wherein each intelligent sound node forms an environment sensing and cooperative execution layer, each node is provided with a wireless sensing, audio processing and network communication module, the wireless sensing module transmits and receives wireless sensing signals and collects environment sensing data of link signals between user equipment and nodes, after receiving the data, a user state sensing unit of the intelligent decision module calculates the real-time position of a user and predicts a short-term movement track, a joint decision unit determines a target sound node set and sound field rendering parameter according to the data, and calculates an optimal data distribution path by combining a dynamic network quality map, and the network communication module and the audio processing module are respectively adjusted according to related parameters. The invention realizes cross-layer joint optimization of acoustics and networks, and ensures audio experience of users in mobile and wireless environment time-varying scenes.

Inventors

  • CHEN ZELIN
  • CHEN CUIFENG
  • DING HUA

Assignees

  • 深圳市狂热者数码科技有限公司

Dates

Publication Date
20260512
Application Date
20260214

Claims (10)

  1. 1. The intelligent sound optimizing system is characterized by comprising a plurality of intelligent sound nodes distributed and an intelligent decision module arranged on a main control node; The intelligent sound nodes form an environment sensing and cooperative execution layer, and each intelligent sound node is provided with a wireless sensing module, an audio processing module and a network communication module; The wireless sensing module is used for transmitting and receiving wireless sensing signals so as to acquire environment sensing data comprising user equipment signals and link signals between nodes; the intelligent decision module is used for receiving the environment sensing data and executing the acoustic and network joint optimization decision taking the user as a center; The intelligent decision module comprises a user state sensing unit and a joint decision unit; The user state sensing unit is used for calculating the real-time space position of the user and predicting the future short-term movement track of the user based on the user equipment signals; the joint decision unit is used for executing the following operations: Determining one or more target sound node sets serving target sound field rendering based on a real-time spatial position of a user and a predicted moving track, and calculating sound field rendering parameters required by each target sound node, wherein the sound field rendering parameters comprise beam forming parameters, equalizer parameters, relative delay parameters and gain parameters; based on the inter-node link signals, constructing a dynamic network quality map reflecting the connection quality and the interference state between the intelligent sound nodes; Calculating an optimal data distribution network path for the target acoustic node set based on the dynamic network quality map and the target acoustic node set; The network communication module is used for dynamically adjusting network connection topology and data transmission strategies according to the optimal data distribution network path; the audio processing module is used for driving the loudspeaker unit to conduct audio rendering according to the sound field rendering parameters.
  2. 2. The optimizing system of intelligent sound according to claim 1, wherein the user state sensing unit calculates a real-time spatial position of a user and predicts a future short-term movement track thereof, and specifically comprises: Acquiring channel state information or signal arrival time difference measurement values in the wireless sensing signals; Processing measured values from a plurality of intelligent sound nodes through a fusion filtering algorithm to obtain a three-dimensional coordinate estimated value of a user; Inputting the historical coordinate sequence of the user into a prediction model, and outputting a predicted track coordinate sequence of the user in a future preset time window; The prediction model is a time sequence prediction model based on a long-term and short-term memory network.
  3. 3. The optimizing system of intelligent sound according to claim 1, wherein the joint decision unit calculates an optimal data distribution network path based on the dynamic network quality map and the target sound node set, in particular by minimizing a joint cost function; the joint cost function is defined as: Wherein, the Representing the total cost of the device and, Representing the acoustic cost function of the device, Representing the cost function of the network, Representing the current and predicted set of locations of the user, A set of target sound nodes representing candidates and a combination of sound field rendering parameters thereof, Represents the quality map of the dynamic network, And The method is used for balancing the acoustic performance and the network transmission cost for a preset weight coefficient; The acoustic cost function For evaluating position The place adopts a sound field scheme When the sound pressure level is equal to the ideal target, the sound image positioning accuracy is equal to the ideal target; The network cost function For evaluating quality maps in dynamic networks Next, a sound field scheme The target sound node set in the network establishes the estimated total transmission delay, jitter and packet loss risk required by the data transmission path.
  4. 4. The optimizing system of an intelligent sound according to claim 1, wherein the wireless sensing signal is an ultra wideband pulse signal or a Wi-Fi signal with channel state information detection capability; The wireless sensing module performs double utilization on the same physical signal: The first reuse is to use the same physical signal as a user equipment signal for positioning and tracking the user; the second reuse is to use the same physical signal as the inter-node link signal for evaluating the wireless link quality between intelligent audio nodes.
  5. 5. The optimizing system of intelligent sound according to claim 1, wherein the joint decision unit further comprises a network topology optimizing subunit; The network topology optimization subunit generates and transmits a network configuration instruction set to related intelligent sound nodes after determining an optimal data distribution network path; the network configuration instruction set includes at least one of: a father node change instruction of a designated node, a transmission priority marking instruction of a specific audio data stream, a node transmitting power adjusting instruction and a node working frequency band switching instruction.
  6. 6. The optimizing system of claim 1, wherein the network communication module adjusts the working state of the network protocol stack of the network communication module according to the instruction after receiving the network configuration instruction set, and the adjusting includes: Updating the routing table according to the father node change instruction; setting a service quality label for the corresponding audio data packet according to the transmission priority marking instruction; controlling the gain of the radio frequency front end according to the transmitting power adjusting instruction; and controlling the radio frequency switch to the designated frequency channel according to the working frequency channel switching instruction.
  7. 7. The optimizing system of claim 1, further comprising a closed loop feedback module deployed at the master control node or at each intelligent audio node; The closed loop feedback module is used for monitoring the deviation of the actual position and the predicted position of the user and the end-to-end play delay and the packet loss rate of each active audio data stream in real time; When the deviation or the end-to-end play delay exceeds a preset threshold, the closed loop feedback module sends a re-decision trigger signal to the intelligent decision module; And after receiving the trigger signal, the intelligent decision module restarts the calculation process of the joint decision unit so as to dynamically fine-tune the rendering parameters of the sound field or the path of the data distribution network.
  8. 8. The optimizing system of intelligent sound according to claim 1, wherein the joint decision unit uses the user prediction track coordinate sequence output by the user state sensing unit as prior information when calculating the optimal data distribution network path; the joint decision unit performs joint modeling on the household space layout, the real-time network quality map and the user prediction track by using a graph neural network model; the output of the graphic neural network model is the prediction of a link with transmission performance which does not meet the preset requirement on the moving path of the user, and the joint decision unit pre-adjusts the data distribution network path according to the prediction result before the user reaches the area with the potential transmission performance which does not meet the preset requirement.
  9. 9. The optimizing system of intelligent sound according to claim 1, wherein the audio processing module uses a clock synchronization mechanism based on reference broadcast when performing audio rendering according to sound field rendering parameters; The master control node periodically broadcasts synchronous signaling containing a global time reference; And after receiving the synchronous signaling, all the target sound nodes correct the local audio clocks and synchronously execute audio playing according to the corrected clocks and the relative delay parameters.
  10. 10. A smart sound optimization system as claimed in claim 3, wherein said joint decision unit solves said joint cost function Adopting an optimization algorithm or a heuristic search algorithm based on gradient descent; In each optimization iteration, the algorithm searches for acoustic costs simultaneously And network cost The weighted sum is the smallest solution space point, and the solution space point corresponds to a group of determined sound field rendering parameter sets and network path configuration sets; the sound field rendering parameter set and the network path configuration set are the output of the joint decision unit.

Description

Optimizing system of intelligent sound equipment Technical Field The invention relates to the technical field of intelligent audio, in particular to an optimization system of intelligent sound equipment. Background With the progress of wireless communication and digital signal processing technologies, distributed intelligent sound systems composed of a plurality of sound box nodes are becoming increasingly popular, and are aimed at providing users with immersive surround sound field experiences. In an ideal static environment, such systems have achieved good hearing through fixed-position acoustic tuning. However, in a real home scenario, the user is in a mobile state, and there are changeable wireless propagation conditions such as wall shielding, interference of home appliances, etc. in the environment, which presents serious challenges for maintaining the continuity and stability of high-quality audio playing. The prior art generally attempts to solve part of the problems from two relatively independent aspects of acoustic rendering optimization or network transmission optimization, but has not been able to systematically address the synergistic contradiction of acoustic requirements and network foundations in dynamic scenarios. At the acoustic optimization level, the prior art has focused on creating an optimal sound field for a fixed listening position or multiple fixed areas. For example, patent document CN119996899a discloses an intelligent processing system for the power amplifier effect of an automobile DSP. The system collects acoustic data through a microphone array, builds a three-dimensional acoustic geometric model, evaluates the spectrum difference of each seat to generate an adjustment coefficient, and further guides a directional beam forming filter to optimize so as to improve the balance and consistency of a sound field in a vehicle. The scheme effectively optimizes the hearing of the fixed seat area through accurate acoustic modeling and beam forming technology, but the optimization target is a static or preset acoustic area, and does not relate to how to perform real-time and self-adaptive sound field tracking and reconstruction on a listener continuously moving in space. Once the listener leaves the preset optimization area, his experience will drop significantly. At the network transmission optimization level, the prior art focuses on predicting and alleviating the problems of delay and instability of audio data transmission. For example, a method and a system for intelligent regulation of a bluetooth speaker and a bluetooth speaker are disclosed, and the disclosure number of the method and the system is CN120640194A. According to the method, the probability and the duration of the play state in the next period which are not in accordance with the requirements are predicted by analyzing the association training probability model of the historical transmission parameters and the delay faults, and the cache depth or the audio gain is adjusted in advance based on the prediction result so as to increase the running stability of the Bluetooth sound box. The scheme is used for resisting the uncertainty of network transmission through predictive regulation, but the optimization object is a single audio data stream and a playing buffer mechanism thereof, and the data stream is not considered to be finally rendered in a collaborative sound field consisting of a plurality of physical nodes. Its network optimization decisions are decoupled from the upper level specific, dynamically changing acoustic rendering requirements (e.g., which speakers are currently required to work in concert, what data they each need). Therefore, one problem faced by the current technology is that in a home scenario where a user moves and the wireless environment is time-varying, there is a significant gap between the dynamic requirement that the acoustic rendering layer provide a coherent sound field for a moving object and the guaranteed capability that the network transmission layer provides a stable, synchronized data stream for a plurality of dynamically varying speaker nodes. If the acoustic algorithm is only optimized, the audio interruption or synchronization misalignment caused by the degradation of the bottom layer network cannot be overcome, and if the network transmission is only optimized, the specific acoustic intention of the upper layer is difficult to perceive, and the most targeted resource scheduling cannot be performed. The prior art lacks a cross-layer joint optimization mechanism capable of cooperatively scheduling sound field rendering parameters and network transmission paths by taking the acoustic experience of a mobile user as a unified target, which causes that the system is difficult to ensure continuous, stable and high-fidelity immersive audio experience in a real and complex environment. It is difficult to provide users with a continuous, stable, high-fidelity immersive audio experience i