Seat control method, device and system and electronic equipment

Abstract

The invention provides a seat control method, device and system and electronic equipment, and the method comprises the steps: obtaining a to-be-played audio signal, carrying out the extraction of low-frequency coefficients of different scales and high-frequency coefficients of different scales, carrying out the multi-scale entanglement reforming simulation processing of the audio signal, all low-frequency coefficients and all high-frequency coefficients, and carrying out the multi-scale entanglement reforming simulation processing of the low-frequency coefficients and all high-frequency coefficients; the method comprises the steps that a reformed low-frequency signal and a reformed high-frequency signal are obtained, a seat control instruction is determined, and when an audio signal is played, a seat is controlled to execute the seat control instruction corresponding to the audio signal, and when the method is applied to a vehicle, actions such as vibration of the seat can be linked based on the audio signal while the vehicle-mounted entertainment system executes functions, and the safety of the vehicle is improved. The requirement of the user for immersive entertainment is further met, and the user experience is improved.

Description

Technical Field

[0001] This application relates to the field of seat control technology, and in particular to a seat control method, device, system and electronic device. Background Technology

[0002] With the rapid development of automotive intelligent technology, in-vehicle entertainment systems (such as games, movies, and music) are gradually becoming an important part of the user's driving experience.

[0003] In related technologies, in-vehicle entertainment systems mainly provide entertainment content to users through vision (display screen) and hearing (speakers). However, this single sensory experience often fails to meet users' needs for immersive entertainment. Summary of the Invention

[0004] This application provides a seat control method, device, system, and electronic device to address the technical problem that the single sensory experience of in-vehicle entertainment systems often fails to meet users' needs for immersive entertainment in related technologies.

[0005] This application provides a seat control method, the method comprising: acquiring an audio signal to be played; extracting low-frequency coefficients and high-frequency coefficients at different scales of the audio signal; performing multi-scale entanglement renormalization simulation processing on the audio signal, all the low-frequency coefficients and high-frequency coefficients to obtain a renormalized low-frequency signal and a renormalized high-frequency signal; determining a seat control command based on the renormalized low-frequency signal and the renormalized high-frequency signal; and controlling the seat to execute the seat control command corresponding to the audio signal when the audio signal is played.

[0006] In one embodiment of this application, extracting low-frequency coefficients and high-frequency coefficients at different scales of the audio signal includes: performing wavelet decomposition on the audio signal to obtain low-frequency coefficients and high-frequency coefficients at different scales.

[0007] In one embodiment of this application, extracting low-frequency coefficients and high-frequency coefficients at different scales of the audio signal includes: obtaining a continuous time variable, a translation parameter, a scaling function, and a wavelet basis function; determining low-frequency coefficients at different scales based on the continuous time variable, the audio signal, the scaling function, and the translation parameter; and determining high-frequency coefficients at different scales based on the continuous time variable, the audio signal, the wavelet basis function, and the translation parameter.

[0008] In one embodiment of this application, performing multi-scale entanglement renormalization on the audio signal, all the low-frequency coefficients, and the high-frequency coefficients to obtain the renormalized low-frequency signal and the renormalized high-frequency signal includes: obtaining the scale of the audio signal decomposition; dividing the audio signal, all the low-frequency coefficients, and the high-frequency coefficients into a number of scale levels to obtain a coefficient tree, wherein the audio signal is located at the bottom level of the coefficient tree, and each level of the coefficient tree represents the low-frequency coefficients and high-frequency coefficients of the audio signal at the corresponding scale; and performing unitary transformation and isometric transformation on the data at each level of the coefficient tree alternately according to the level to realize the multi-scale entanglement renormalization process and obtain the renormalized low-frequency signal and the renormalized high-frequency signal, wherein the bottom level data undergoes unitary transformation.

[0009] In one embodiment of this application, the data at each level of the coefficient tree are subjected to alternating unitary and isometric transformations according to the level to achieve the multi-scale entanglement renormalization simulation processing, obtaining renormalized low-frequency signals and renormalized high-frequency signals. This includes: inputting the coefficient tree into a pre-trained multi-scale entanglement renormalization simulation network, and implementing the multi-scale entanglement renormalization simulation processing through the multi-scale entanglement renormalization simulation network to obtain renormalized low-frequency signals and renormalized high-frequency signals; wherein, the multi-scale entanglement renormalization simulation network includes multiple unentanglers and isometric transformations. The transformation module includes a non-bottom-level unentanglement unit used to perform unitary transformation on the layer data corresponding to the unentanglement unit and the iso-distance transformed data, a bottom-level unentanglement unit used to perform unitary transformation on the layer data corresponding to the unentanglement unit, and an iso-distance transformation module used to perform iso-distance transformation on the unitarily transformed data and the layer data of the layer where the iso-distance transformation module is located. The network layer of the multi-scale entanglement renormalization simulation network is equal to the scale, and each grid point of the multi-scale entanglement renormalization simulation network is used to store high-frequency coefficients or low-frequency coefficients at the corresponding position and scale.

[0010] In one embodiment of this application, the training method of the multi-scale entanglement renormalization simulation network includes: acquiring sample data, the sample data including sample renormalized low-frequency signals, sample renormalized high-frequency signals, sample low-frequency signals, and sample high-frequency signals; processing the sample low-frequency signals and sample high-frequency signals through an initial multi-scale entanglement renormalization simulation network to obtain processed renormalized low-frequency signals and processed renormalized high-frequency signals; determining a loss value based on the sample renormalized low-frequency signals, sample renormalized high-frequency signals, renormalized low-frequency signals, and processed renormalized high-frequency signals; and updating the parameters of the initial multi-scale entanglement renormalization simulation network based on the loss value to obtain the trained multi-scale entanglement renormalization simulation network.

[0011] In one embodiment of this application, after the seat executes the seat control command corresponding to the audio signal, the method further includes: acquiring user adjustment data, wherein the user adjustment data includes at least one of seat vibration intensity feedback data, seat vibration mode feedback data, and audio source feedback data; adjusting the matching relationship between the renormalized low-frequency signal and the renormalized high-frequency signal and the seat control command based on the user adjustment data; and determining a new seat control command based on the newly determined renormalized low-frequency signal, the renormalized high-frequency signal, and the adjusted matching relationship to control the seat.

[0012] This application embodiment also provides a seat control device, the seat control device comprising: an audio signal acquisition module for acquiring an audio signal to be played; a coefficient extraction module for extracting low-frequency coefficients and high-frequency coefficients at different scales of the audio signal; a multi-scale entanglement renormalization simulation processing module for performing multi-scale entanglement renormalization simulation processing on the audio signal, all the low-frequency coefficients and high-frequency coefficients to obtain a renormalized low-frequency signal and a renormalized high-frequency signal; a seat control command generation module for determining a seat control command based on the renormalized low-frequency signal and the renormalized high-frequency signal; and a control module for controlling the seat to execute the seat control command corresponding to the audio signal when the audio signal is played.

[0013] This application embodiment also provides a seat control system, which includes a vehicle, a wavelet decomposition module, and a multi-scale entanglement renormalization simulation processing network. The vehicle includes a seat, a seat controller, and an audio playback device, wherein: the wavelet decomposition module is used to extract low-frequency coefficients and high-frequency coefficients of different scales of the audio signal to be played by the audio playback device; the multi-scale entanglement renormalization simulation processing network is used to perform multi-scale entanglement renormalization simulation processing on the audio signal, all the low-frequency coefficients, and the high-frequency coefficients to obtain renormalized low-frequency signals and renormalized high-frequency signals; the seat controller is used to determine seat control commands based on the renormalized low-frequency signals and renormalized high-frequency signals; the audio playback device is used to play the audio signal; and the seat controller is also used to control the seat to execute the seat control commands corresponding to the audio signal when the audio signal is played.

[0014] This application also provides an electronic device, including: a memory storing a computer program thereon; and a processor for executing the computer program in the memory to implement the steps of the method described in any of the above embodiments.

[0015] This invention also provides a computer-readable storage medium having a computer program stored thereon, the computer program being used to cause a computer to perform the method provided in any of the above embodiments.

[0016] The beneficial effects of this application are as follows: The seat control method, device, system, and electronic device proposed in this application acquire an audio signal to be played, extract low-frequency coefficients and high-frequency coefficients at different scales, perform multi-scale entanglement renormalization on the audio signal, all low-frequency coefficients, and high-frequency coefficients to obtain renormalized low-frequency signals and renormalized high-frequency signals, and determine seat control commands. When the audio signal is played, the seat is controlled to execute the seat control commands corresponding to the audio signal. When this method is applied to a vehicle, it can realize the vibration and other actions of the seat based on the audio signal while the in-vehicle entertainment system performs its functions, further meeting the user's needs for immersive entertainment and improving the user experience. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0018] In the attached diagram: Figure 1 This is a schematic diagram illustrating an application scenario of a seat control method provided in an embodiment of this application; Figure 2 A schematic flowchart of a seat control method provided in one embodiment of this application; Figure 3 A schematic diagram of a multi-scale entanglement renormalization proposed network provided in an embodiment of this application; Figure 4 A schematic diagram of a de-entanglement device provided in one embodiment of this application; Figure 5 A schematic diagram of an isometric transformation module provided in one embodiment of this application; Figure 6 A schematic flowchart illustrating a seat control method provided in one embodiment of this application; Figure 7 A schematic diagram of the structure of a seat control device provided in an embodiment of this application; Figure 8 A schematic diagram of a seat control system provided in one embodiment of this application; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the shape, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the present application. However, it will be apparent to those skilled in the art that embodiments of the present application may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the present application.

[0022] With the rapid development of automotive intelligent technology, in-vehicle entertainment systems (such as games, movies, and music) are gradually becoming an important part of the user's driving experience. Traditional in-vehicle entertainment systems mainly provide entertainment content to users through vision (display screens) and hearing (speakers). However, this single sensory experience often cannot meet users' needs for immersive entertainment. In recent years, haptic feedback technology has been gradually introduced into the automotive field, but existing technologies are mainly focused on safety reminder functions (such as lane departure warnings and collision warnings) and have not yet been fully applied to entertainment scenarios.

[0023] In-vehicle entertainment systems typically include a central control screen, audio system, and multimedia playback functions, providing a high-quality audio-visual experience. Some high-end models also feature virtual reality (VR) or augmented reality (AR) technology to further enhance the entertainment experience. Seat vibration functions are primarily used for safety alerts, such as reminding the driver when the vehicle deviates from its lane or a collision risk is detected. Some models also introduce massage functions, providing a simple massage effect through vibration motors built into the seats. Haptic feedback technology is widely used in consumer electronics (such as mobile phones and game controllers), simulating different tactile effects through vibration. In the automotive field, haptic feedback technology is mainly used for force feedback in components such as the steering wheel and pedals to enhance the driving feel.

[0024] However, in related technologies, seat vibration functions are primarily used for safety alerts and lack integration with in-vehicle entertainment systems, failing to provide users with an immersive entertainment experience. In-vehicle entertainment systems only offer experiences through sight and sound, lacking tactile feedback, resulting in insufficient user immersion. Furthermore, they do not adequately consider personalized user needs, failing to dynamically adjust vibration effects based on user preferences or entertainment content. Seat vibration functions and in-vehicle entertainment systems are typically designed independently, lacking the ability to work collaboratively, leading to resource waste, a fragmented user experience, and low technological integration.

[0025] To address the aforementioned technical problems, this application provides a seat control method. By acquiring the audio signal to be played, low-frequency coefficients and high-frequency coefficients at different scales are extracted. The audio signal, all low-frequency coefficients, and high-frequency coefficients are then subjected to multi-scale entanglement renormalization to obtain renormalized low-frequency and high-frequency signals. Seat control commands are then determined. When the audio signal is played, the seat is controlled to execute the corresponding seat control command. When this method is applied to a vehicle, it can simultaneously enable the in-vehicle entertainment system to perform functions while the seat vibrates based on the audio signal, further satisfying the user's need for immersive entertainment and enhancing the user experience.

[0026] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating an application scenario of a seat control method provided in an embodiment of this application. For example... Figure 1 As shown, taking the application of this seat control method to a vehicle as an example, when the vehicle performs entertainment functions such as playing music and movies, the audio signal to be played can be acquired, and wavelet decomposition can be performed on it to obtain high-frequency and low-frequency coefficients corresponding to different scales (resolutions). Then, multi-scale entanglement renormalization simulation processing is performed to obtain the corresponding renormalized high-frequency and low-frequency signals. Based on this, the corresponding seat control command is determined. When the audio signal is played, the corresponding seat control command is executed synchronously to control the seat to move or vibrate, thereby enhancing the user's immersive experience.

[0027] It should be noted that the above scenario is only an example of an application scenario provided by the embodiments of this application. The method can also be applied to scenarios such as massage chairs and cinemas according to the user's needs. The embodiments of this application do not limit the actual form of various devices, components, etc. included in the scenario. In the specific application of the solution, it can be set according to actual needs. Please see Figure 2 , Figure 2 A schematic flowchart of a seat control method provided in one embodiment of this application is shown below. Figure 2 As shown, the method includes the following steps: Step S210: Obtain the audio signal to be played.

[0028] As an example, if the audio file to be played is predetermined, such as a specific movie, the audio signal corresponding to that audio file can be obtained in advance and subsequent steps can be executed, which can improve the synchronization of subsequent seat control.

[0029] As another example, if the currently playing audio file is music, since the user's music playback is somewhat random and may switch songs at any time, an audio signal of a preset duration can be pre-set as the audio signal to be played, such as an audio signal within 1 minute that is about to be played. This signal can then be processed and corresponding seat control commands can be generated. This can ensure the timeliness of subsequent seat control and reduce the waste of computing resources caused by the user switching audio.

[0030] Step S220: Extract low-frequency coefficients and high-frequency coefficients at different scales of the audio signal.

[0031] In one embodiment, extracting low-frequency coefficients and high-frequency coefficients at different scales of an audio signal includes: performing wavelet decomposition on the audio signal to obtain low-frequency coefficients and high-frequency coefficients at different scales. The wavelet decomposition method described above is just one example; those skilled in the art can also use other known methods to extract low-frequency and high-frequency coefficients.

[0032] In one embodiment, extracting low-frequency coefficients and high-frequency coefficients at different scales of an audio signal includes: acquiring a continuous-time variable, a translation parameter, a scaling function, and a wavelet basis function; determining low-frequency coefficients at different scales based on the continuous-time variable, the audio signal, the scaling function, and the translation parameter; and determining high-frequency coefficients at different scales based on the continuous-time variable, the audio signal, the wavelet basis function, and the translation parameter.

[0033] Following the above embodiments, taking the Discrete Wavelet Transform (DWT) as an example, the decomposition formula is as follows: cA=k∑f(t) (t k) Formula (1), cD=k∑f(t) ψ(t k) Formula (2), Where cA represents the low-frequency coefficient, cD represents the high-frequency coefficient, t represents the continuous time variable, and k represents the translation parameter. Analyzing the characteristics of the signal at different time points, by continuously changing k, allows... (t k) and ψ(t) k) This "analysis window" slides across the entire signal to obtain the low-frequency and high-frequency information of the signal at various time points. f(t) is the original continuous-time audio signal. (t k) is the scaling function, used to extract low-frequency components (the "parent" function for capturing low-frequency components of the signal), ψ(t) k) is a wavelet basis function used to extract high-frequency components (the “mother” function that captures the high-frequency components and details of a signal).

[0034] Wavelet decomposition is a signal analysis method based on wavelet transform, which separates signals according to their frequency components through multi-scale decomposition. Discrete wavelet transform (DWT) employs iterative decomposition using filter banks, and its mathematical expression relies on integer shifts and scaling of the wavelet basis function ψ. Its core processes include selecting the number of decomposition levels and determining the wavelet basis functions. It outputs arrays of low-frequency and high-frequency coefficients. Wavelet decomposition can capture local information in audio signals.

[0035] Wavelet transform is used to decompose audio signals at multiple scales, extracting low-frequency and high-frequency components at different scales. Wavelet transform can capture the local features of a signal and is particularly suitable for processing non-stationary signals, such as audio in music and movies.

[0036] Step S230: Perform multi-scale entanglement renormalization simulation processing on the audio signal, all low-frequency coefficients and high-frequency coefficients to obtain the renormalized low-frequency signal and the renormalized high-frequency signal.

[0037] The Multiscale Entanglement Renormalization Proposal (MERA) is a tensor network state used to describe quantum many-body systems, effectively handling quantum entanglement at different scales. This method progressively reduces the system's degrees of freedom while preserving its low-energy state properties by dividing the system into different scales and applying entanglement renormalization operations at each scale. In practice, the MREA method describes quantum states by constructing a network of tensors.

[0038] By renormalizing, MERA compresses low-frequency signals into a low-dimensional representation through the recursive application of quantum entanglement operations.

[0039] In one embodiment, performing multi-scale entanglement renormalization on the audio signal, all low-frequency coefficients, and high-frequency coefficients to obtain renormalized low-frequency and high-frequency signals includes: obtaining the scale of audio signal decomposition; dividing the audio signal, all low-frequency coefficients, and high-frequency coefficients into levels of scale to obtain a coefficient tree, with the audio signal located at the bottom level of the coefficient tree, and each level of the coefficient tree representing the low-frequency and high-frequency coefficients of the audio signal at the corresponding scale; and performing unitary transformation and isometric transformation on the data at each level of the coefficient tree alternately according to the level to achieve multi-scale entanglement renormalization to obtain renormalized low-frequency and high-frequency signals, wherein the bottom level data undergoes unitary transformation.

[0040] Scale can also be understood as resolution. For a given scale, that is, the audio signal, for each scale of the gas, there are corresponding low-frequency coefficients and high-frequency coefficients.

[0041] The coefficient tree is a pyramid-like structure that matches the architecture of MERA. By processing the audio signal, all low-frequency coefficients, and high-frequency coefficients through unitary or isometric transformations at corresponding levels, the internal information of the audio signal can be correlated. Furthermore, compared to methods such as machine learning algorithms, the method used in this embodiment supports backtracking, which can preserve the correlation between data, rather than being a "black box" that only knows the final output, thus facilitating backtracking.

[0042] Following the above embodiments, the data at each level of the coefficient tree are subjected to alternating unitary and isometric transformations according to the level to achieve multi-scale entanglement renormalization simulation processing, obtaining renormalized low-frequency and high-frequency signals. This includes: inputting the coefficient tree into a pre-trained multi-scale entanglement renormalization simulation network, and performing multi-scale entanglement renormalization simulation processing through the multi-scale entanglement renormalization simulation network to obtain renormalized low-frequency and high-frequency signals; wherein, the multi-scale entanglement renormalization simulation network includes multiple unentanglers and... The isometric transformation module is used to perform unitary transformation on the layer data corresponding to the unentangled unit and the data after isometric transformation. The unentangled unit at the bottom layer is used to perform unitary transformation on the layer data corresponding to the unentangled unit. The isometric transformation module is used to perform isometric transformation on the data after unitary transformation and the layer data of the layer where the isometric transformation module is located. The network layers and scales of the multi-scale entanglement renormalization simulation network are equal. Each grid point of the multi-scale entanglement renormalization simulation network is used to store the high-frequency coefficients or low-frequency coefficients at the corresponding position and scale.

[0043] The Multiscale Entangled Renormalization (MERA) network is a hierarchical network. Wavelet coefficients need to be arranged in a pyramid-like structure to match the MERA architecture.

[0044] One exemplary approach is to construct a complete wavelet coefficient tree (coefficient tree): The bottom layer (layer 0): the original signal f[n] itself. This is the finest scale of the "physical lattice".

[0045] Layer 1: Composed of coefficients from the first-level wavelet decomposition. Typically, the approximation coefficients cA_1 (low-frequency coefficients) and detail coefficients cD_1 (high-frequency coefficients) are interleaved to form a new one-dimensional sequence. The length of this sequence is the same as the original signal.

[0046] Level 2: Perform a second-level decomposition on cA_1 to obtain cA_2 and cD_2. Similarly, arrange them alternately and use them as the "parent layer" of the previous layer (the layer where cA_1 is located). ... Layer J (topmost layer): consists of the coarsest scale cA_J and cD_J.

[0048] Now, we have a hierarchical structure, where each layer represents the information of the signal at a specific resolution (scale). This hierarchical structure is the basic framework of the MERA network, namely the coefficient tree (wavelet coefficient pyramid) mentioned above. When this coefficient tree is input into the MERA network, the network's levels directly correspond to the wavelet decomposition scales J, J-1, ..., 1, 0. Each node (grid point) stores the wavelet coefficient value at the corresponding location and scale.

[0049] Please see Figure 3 , Figure 4 and Figure 5 , Figure 3 This is a schematic diagram of a multi-scale entanglement renormalization proposed network provided in one embodiment of this application. Figure 4 This is a schematic diagram of a de-entanglement device provided in one embodiment of this application. Figure 5 This is a schematic diagram of the structure of an isometric transformation module provided in an embodiment of this application, as shown below. Figure 3 As shown, the MERA network consists of two basic operations (two types of tensors) that are alternately applied to this pyramid structure. The entangler, for example... Figure 4 As shown, at the same scale, it is a series of unitary transformations that act on adjacent grid points (i.e., wavelet coefficients) with the aim of decoupling short-range entanglements / correlations.

[0050] Isometry transformation module, such as Figure 5As shown, the transformation from a finer scale to a coarser scale is a dimensionality reduction transformation. It compresses and merges the information from several lattice points (such as two) into a new lattice point representing a coarser degree of freedom. For standard binary MERA, both tensors need to satisfy specific constraints, and the unentangler must be unitary: Formula (3), in, U This represents the unit operator itself. U † yes U Hermitian conjugate, I It is a unit operator.

[0051] Equidity needs to satisfy: Formula (4), in, V It is an operator that maps vectors from a low-dimensional space to a high-dimensional space. V † yes V Hermitian conjugate, I It is a unit operator.

[0052] The unentangler preserves the entanglement between two data points input to the same unentangler and de-entangles data input to different unentanglers; then, the output of the unentanglement layer is coarsened by the equidistant layer. The characteristics of the unentangler and the equidistant layer enable the MERA network to capture all entanglements between data at the same scale in the same layer.

[0053] Since the isometric mapping in the MERA network is fixed (determined by the wavelet basis), it is necessary to learn the entangler operation of each level through optimization algorithms.

[0054] In one embodiment, the training method of the multi-scale entanglement renormalization simulated network includes: acquiring sample data, which includes sample renormalized low-frequency signals, sample renormalized high-frequency signals, sample low-frequency signals, and sample high-frequency signals; processing the sample low-frequency signals and sample high-frequency signals through an initial multi-scale entanglement renormalization simulated network to obtain processed renormalized low-frequency signals and processed renormalized high-frequency signals; determining a loss value based on the sample renormalized low-frequency signals, sample renormalized high-frequency signals, renormalized low-frequency signals, and processed renormalized high-frequency signals; and updating the parameters of the initial multi-scale entanglement renormalization simulated network based on the loss value to obtain the trained multi-scale entanglement renormalization simulated network.

[0055] The training objective of the multi-scale entanglement renormalization proposed network is to enable the entire MERA network to most effectively represent (or compress) the information contained in the input wavelet coefficient pyramid.

[0056] As an example, taking the principles of backpropagation and stochastic gradient descent for model training, the training method of the multi-scale entanglement renormalization hypothetical network is as follows: Input: Training data Learning rate Batch size m.

[0057] Where N is the sample size. Let i be the i-th input feature (sample low-frequency signal and sample high-frequency signal). For the i-th true label (low-frequency signal after sample renormalization, high-frequency signal after sample renormalization), the step size of parameter updates is controlled by the learning rate, which determines the "step size" of each gradient descent. The number of samples selected in each iteration is determined by the batch size, which is used to balance training efficiency and stability.

[0058] process: 1. Initialize parameters W, b: Where W is the weight matrix of the l-th layer, used to perform a weighted transformation on the output of the previous layer. b is the bias term of the l-th layer, used to adjust the baseline value of the weighted input.

[0059] 2. Repeat until convergence: (1) Randomly select m samples ; This allows for the extraction of a small batch of samples from the training set to update the parameters for this iteration.

[0060] (2) Forward propagation: Formula (5), Among them, z (1) W is the weighted input for layer 1. (1) Let b be the weight matrix of the l-th layer. (1) Here, x is the bias term of the l-th layer, and x is the input feature vector. a (1) The output is the activation value of layer 1. It is the activation function, z (2) W is the weighted input for the second layer. (2) Let b be the weight matrix of the second layer. (2) For the bias term of the second layer, a (2) The activation output of layer 2, z (L) W is the weighted input for the Lth layer. (L) Let b be the weight matrix of the Lth layer. (L) For the bias term of the Lth layer, a (L) The activation output of layer L. a (L-1)The activation output of layer L-1, y ^ represents the network's predicted output, which is the activation output of the Lth layer.

[0061] (3) Calculate the loss: Formula (6), Where J is the batch loss, is the single-sample loss of the i-th sample (such as cross-entropy, mean squared error, etc.), and m is the number of samples.

[0062] (4) Backpropagation: Calculate the output layer error: Formula (7), in, It is the error term of the output layer. It is the gradient of the loss with respect to the predicted output. Is the activation function in The derivative at point, This indicates element-wise multiplication.

[0063] Backpropagation layer by layer: Formula (8), in, It is the error term of the l-th layer. It is the first The transpose of the layer weight matrix is ​​used to propagate errors from higher layers to lower layers. It is the error term of the (l+1)th layer. Is the activation function in The derivative at point, This indicates element-wise multiplication.

[0064] Calculate the gradient: Formula (9), Formula (10), in, Let the gradient of the weight matrix (weights) of the l-th layer be denoted as . It is the transpose of the activation output of layer (l-1). It is the error term of the l-th layer. The gradient of the bias term (bias term) at the l-th layer.

[0065] (5) Parameter update: Formula (11), Formula (12), Formula (11) updates the weights of the l-th layer using gradient descent. It is the learning rate, which controls the update magnitude. Formula (12) updates the bias of the l-th layer using gradient descent. For the weights of the l-th layer, Let the gradient of the weight matrix (weights) of the l-th layer be denoted as . Let be the gradient of the bias term at layer l. This is the bias for the l-th layer.

[0066] In one embodiment, taking low-frequency coefficients as an example, the renormalization process can be expressed as follows: cAE=R(cA) formula (13). Where R represents the renormalization operation, cAE is the renormalized low-frequency signal, and cA is the low-frequency coefficient.

[0067] In low-frequency signal processing, MERA can further optimize the low-frequency signals extracted by wavelet transform. Its core idea is to gradually extract the key features of the signal through a hierarchical structure while reducing redundant information.

[0068] Inputting the wavelet transform result into MERA is a process of "quantizing" classical signals. It uses wavelet transform as the front end of multi-scale analysis to decompose the signal into different time-frequency domains. Then, it uses the MERA network as the back end for multi-scale correlation modeling to deeply reveal and quantify the complex cross-scale correlation structures hidden among these coefficients. Unlike traditional statistical methods, the method provided in this embodiment examines familiar classical signals through a tool describing the most complex correlations in the quantum world, potentially uncovering unprecedented deep patterns and features.

[0069] Step S240: Determine seat control commands based on the renormalized low-frequency signal and the renormalized high-frequency signal.

[0070] The method of determining seat control commands based on renormalized low-frequency and high-frequency signals can be achieved by using rules pre-defined by those skilled in the art regarding different seat control methods and amplitudes corresponding to different signals. Alternatively, the method of determining seat control commands based on signals can be implemented using other methods known to those skilled in the art.

[0071] As an example, the control methods of seat control commands include, but are not limited to, the forward and backward movement of the seat, the activation of seat massage, and one or more of the following control methods: massage mode, massage intensity, seat vibration, vibration amplitude, vibration frequency, etc.

[0072] The above method can be used to associate audio signals with seat controls, so that the seat controls can be activated when audio is played.

[0073] Step S250: When the audio signal is played, control the seat to execute the seat control command corresponding to the audio signal.

[0074] As an example, seat control commands can be tagged according to the time domain of their corresponding audio signals. When the audio is played, seat control is performed simultaneously, so that the user can get good auditory and visual feedback.

[0075] In one embodiment, after controlling the seat to execute the seat control command corresponding to the audio signal, the method further includes: acquiring user adjustment data, the user adjustment data including at least one of seat vibration intensity feedback data, seat vibration mode feedback data, and audio source feedback data; adjusting the matching relationship between the renormalized low-frequency signal and the renormalized high-frequency signal and the seat control command based on the user adjustment data; and determining a new seat control command based on the newly determined renormalized low-frequency signal, the renormalized high-frequency signal, and the adjusted matching relationship to control the seat.

[0076] Since different users may have personalized needs regarding the range of seat adjustments, the system can proactively inquire about the user's adjustment experience after executing the seat control command, and obtain user adjustment data based on user feedback. Alternatively, the system can detect the user's seat adjustment actions; for example, if the user performs seat adjustment control immediately after executing a control command, this control data can be used as user adjustment data. User adjustment data is used to characterize the user's evaluation of the seat adjustment performed using the method provided in this embodiment, or data indicating a need for readjustment of the seat.

[0077] Seat vibration intensity feedback data characterizes the user's evaluation of the seat vibration amplitude, such as feeling the vibration amplitude is too large. Seat vibration mode feedback data characterizes the user's evaluation of the seat vibration frequency or mode, such as the vibration speed being too fast. Audio source feedback data indicates whether the user wants related adjustments to the seat while playing music or videos. For example, a user might want to avoid related seat controls when listening to a certain type of music; in this case, subsequent seat controls will be based on the user's personalized needs, and related seat controls will not be activated when playing that type of music.

[0078] The above methods provide users with a way to personalize their settings and also offer a self-learning mechanism for seat adjustments, which can better meet user needs and improve the user experience.

[0079] The seat control method proposed in the above embodiments acquires the audio signal to be played, extracts low-frequency coefficients and high-frequency coefficients at different scales, performs multi-scale entanglement renormalization on the audio signal, all low-frequency coefficients and high-frequency coefficients, obtains renormalized low-frequency signals and renormalized high-frequency signals, and determines the seat control command. When the audio signal is played, the seat is controlled to execute the seat control command corresponding to the audio signal. When this method is applied to a vehicle, it can realize the vibration and other actions of the seat based on the audio signal while the in-vehicle entertainment system performs its functions, further meeting the user's needs for immersive entertainment and improving the user experience.

[0080] This method offers a multi-scale feature extraction approach. Wavelet transform can capture local features of a signal, and MERA further optimizes these features through renormalization, making it more suitable for low-frequency signal analysis. It also enables dimensionality reduction and resource optimization; MERA can compress complex signals into low-dimensional representations, reducing computational resource consumption. Furthermore, this method is highly adaptable, suitable for non-stationary signals (such as music and movie audio), and can effectively extract high and low frequency signals while reducing noise interference.

[0081] By applying the above method to vehicles, a linkage between seat control and in-vehicle entertainment system can be provided, offering users an immersive entertainment experience. The seat vibration function and the in-vehicle entertainment system can work together, avoiding resource waste and a fragmented user experience. Furthermore, the embodiments provided by this method fully consider users' personalized needs and can dynamically adjust the vibration effect according to user preferences or entertainment content.

[0082] Please see Figure 6 , Figure 6 A specific flowchart illustrating a seat control method provided in an embodiment of this application is shown below. Figure 6 As shown, taking the application of this method to a vehicle as an example, the steps are as follows: The in-vehicle entertainment system sends a corresponding signal, which can be the audio signal to be played. The signal processing module further processes the wavelet-transformed signal based on the MERA network. First, the audio signal is subjected to wavelet transformation, and then, according to the network structure of the MERA network, the wavelet-transformed signal is processed into a coefficient tree with the same network structure as the MERA network, so that the MERA network can process it. Then, the control unit determines the corresponding seat control command based on the received signal (based on the data processed by the MERA network) and controls the working mode of the vibration module. The in-vehicle seat is equipped with a vibration module. The user can customize the vibration intensity, vibration mode, and linkage content based on their own feelings, and feed this customized data back to the control unit so that the control unit can use it when determining subsequent seat control commands.

[0083] In one embodiment, a seat control device is provided for performing the seat control method provided in any of the above embodiments. See also... Figure 7 , Figure 7 A schematic diagram of the structure of a seat control device provided in an embodiment of this application is shown below. Figure 7 As shown, the seat control device 700 includes: an audio signal acquisition module 710 for acquiring an audio signal to be played; a coefficient extraction module 720 for extracting low-frequency coefficients and high-frequency coefficients at different scales of the audio signal; a multi-scale entanglement renormalization simulation processing module 730 for performing multi-scale entanglement renormalization simulation processing on the audio signal, all low-frequency coefficients, and high-frequency coefficients to obtain renormalized low-frequency signals and renormalized high-frequency signals; a seat control command generation module 740 for determining seat control commands based on the renormalized low-frequency signals and renormalized high-frequency signals; and a control module 750 for controlling the seat to execute the seat control commands corresponding to the audio signal when the audio signal is played.

[0084] In one embodiment, the seat control device further includes a self-learning module, which, after controlling the seat to execute a seat control command corresponding to an audio signal, acquires user adjustment data, including at least one of seat vibration intensity feedback data, seat vibration mode feedback data, and audio source feedback data; adjusts the matching relationship between the renormalized low-frequency signal and the renormalized high-frequency signal and the seat control command based on the user adjustment data; and determines a new seat control command based on the newly determined renormalized low-frequency signal, the renormalized high-frequency signal, and the adjusted matching relationship to control the seat.

[0085] For specific limitations regarding the seat control device, please refer to the limitations on the seat control method above, which will not be repeated here. Each module in the aforementioned seat control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the electronic device in hardware form, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0086] In this embodiment, the seat control device is essentially equipped with multiple modules to execute the seat control method in any of the above embodiments. The specific functions and technical effects can be referred to in the above embodiments, and will not be repeated here.

[0087] In one embodiment, a seat control system is provided for executing the seat control method provided in any of the above embodiments. See also... Figure 8 , Figure 8 A schematic diagram of a seat control system provided in one embodiment of this application is shown below. Figure 8As shown, the seat control system 800 includes a vehicle 810, a wavelet decomposition module 820, and a multi-scale entanglement renormalization simulation processing network 830. The vehicle 810 includes a seat 811, a seat controller 812, and an audio playback device 813. Specifically: the wavelet decomposition module 820 extracts low-frequency coefficients and high-frequency coefficients at different scales from the audio signal to be played by the audio playback device; the multi-scale entanglement renormalization simulation processing network 830 performs multi-scale entanglement renormalization simulation processing on the audio signal, all low-frequency coefficients, and high-frequency coefficients to obtain renormalized low-frequency and high-frequency signals; the seat controller 812 determines seat control commands based on the renormalized low-frequency and high-frequency signals; the audio playback device 813 plays the audio signal; and the seat controller 812 also controls the seat 811 to execute the seat control commands corresponding to the audio signal when the audio signal is played.

[0088] It should be noted that the wavelet decomposition module and / or the multi-scale entanglement renormalization proposed processing network can be deployed on the vehicle or in the cloud, depending on the needs of those skilled in the art.

[0089] In one embodiment, the vehicle further includes a vibration module installed in the seat back, cushion, and headrest to generate vibration feedback. The vibration module is controlled by a seat controller to execute seat control commands. For example, multiple vibration motors are installed in the seat back, cushion, and headrest to simulate vibration effects in different directions. The vibration motors are designed for low power consumption and low noise to ensure comfort.

[0090] This method supports multiple vibration modes, such as continuous vibration, pulsed vibration, and gradual vibration. The vibration intensity can be automatically adjusted according to user preferences.

[0091] In one embodiment, the vehicle also includes a control unit for receiving signals from the in-vehicle entertainment system.

[0092] In one embodiment, the seat controller includes a signal processing module for converting audio, video, or event signals from entertainment content (such as games, movies, and music) into seat control commands such as vibration instructions. Low-frequency signals from music or movies are extracted using audio analysis technology and converted into vibration commands. For game content, dynamic vibration effects are generated based on game events (such as collisions or accelerations).

[0093] In one embodiment, the vehicle also includes a user interface that allows users to customize vibration intensity, mode, and associated content. For example, vibration settings options can be provided on a central control screen or a mobile app, allowing users to select associated content (such as movies, games, and music) and vibration modes.

[0094] For specific limitations regarding the seat control system, please refer to the limitations on the seat control method above, which will not be repeated here. Each module in the aforementioned seat control system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the electronic device in hardware form or independent of it, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0095] In this embodiment, the seat control system is essentially configured with multiple modules to execute the vehicle-side execution method in any of the above embodiments of the seat control method. The specific functions and technical effects can be referred to in the above embodiments, and will not be repeated here.

[0096] See Figure 9 , Figure 9 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown below. Figure 9 As shown, this embodiment of the invention also provides an electronic device 900, including a processor 901, a memory 902, and a communication bus 903; the communication bus 903 is used to connect the processor 901 and the memory 902; the processor 901 is used to execute a computer program stored in the memory 902 to implement the method described in any of the above embodiments.

[0097] This invention also provides a computer-readable storage medium having a computer program stored thereon, the computer program being used to cause a computer to perform the method provided in any of the above embodiments.

[0098] This application also provides a non-volatile readable storage medium storing one or more modules (programs) that, when applied to a device, enable the device to execute the instructions included in the steps provided in this application.

[0099] This application also provides a computer program product, including a computer program that, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0100] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0101] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0102] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0103] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0104] It should be understood that the terms "first," "second," etc., used in this application are used to distinguish similar objects and do not necessarily indicate a specific order or sequence. The technical features to which these terms are used can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that shown in the figures or text.

[0105] It should be understood that although the flowcharts provided in the embodiments of this application indicate the various steps with arrows, the order indicated by the arrows does not necessarily limit the implementation order of these steps. Those skilled in the art can perform these steps in other orders according to different implementation scenarios and requirements.

[0106] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A seat control method, characterized in that, The method includes: Obtain the audio signal to be played; Extract the low-frequency coefficients and high-frequency coefficients at different scales of the audio signal; The audio signal, all the low-frequency coefficients and high-frequency coefficients are subjected to multi-scale entanglement renormalization simulation processing to obtain the renormalized low-frequency signal and the renormalized high-frequency signal. Seat control commands are determined based on the renormalized low-frequency signal and the renormalized high-frequency signal; When the audio signal is played, the control seat executes the seat control command corresponding to the audio signal.

2. The seat control method as described in claim 1, characterized in that, Extracting low-frequency coefficients and high-frequency coefficients at different scales from the audio signal, including: The audio signal is decomposed by wavelet to obtain low-frequency coefficients and high-frequency coefficients at different scales.

3. The seat control method as described in claim 1, characterized in that, Extracting low-frequency coefficients and high-frequency coefficients at different scales from the audio signal, including: Obtain continuous-time variables, translation parameters, scaling functions, and wavelet basis functions; The low-frequency coefficients at different scales are determined based on the continuous-time variable, audio signal, scaling function, and translation parameters. The high-frequency coefficients at different scales are determined based on the continuous time variable, audio signal, wavelet basis function, and translation parameters.

4. The seat control method according to any one of claims 1-3, characterized in that, The audio signal, all the low-frequency coefficients, and the high-frequency coefficients are subjected to multi-scale entanglement renormalization to obtain the renormalized low-frequency signal and the renormalized high-frequency signal, including: Obtain the scale of the audio signal decomposition; The audio signal, all the low-frequency coefficients and high-frequency coefficients are divided into a number of levels according to the scale to obtain a coefficient tree. The audio signal is located at the bottom level of the coefficient tree. The data at each level of the coefficient tree represents the low-frequency coefficients and high-frequency coefficients of the audio signal at the corresponding scale. The coefficient tree data at each level is subjected to unitary transformation and isometric transformation alternately according to the level to realize the multi-scale entanglement renormalization simulation process, and the renormalized low-frequency signal and renormalized high-frequency signal are obtained. The bottom level data is subjected to unitary transformation.

5. The seat control method as described in claim 4, characterized in that, The coefficient tree data at each level are subjected to alternating unitary and isometric transformations according to the level to achieve the proposed multi-scale entanglement renormalization process, resulting in renormalized low-frequency and high-frequency signals, including: The coefficient tree is input into a pre-trained multi-scale entanglement renormalization simulation network, and the multi-scale entanglement renormalization simulation process is implemented through the multi-scale entanglement renormalization simulation network to obtain the renormalized low-frequency signal and the renormalized high-frequency signal. The multi-scale entanglement renormalization proposed network includes multiple unentanglers and isometric transformation modules. The unentanglers that are not at the bottom layer are used to perform unitary transformation on the layer data corresponding to the unentangler and the data after isometric transformation. The unentanglers at the bottom layer are used to perform unitary transformation on the layer data corresponding to the unentangler. The isometric transformation module is used to perform isometric transformation on the data after unitary transformation and the layer data of the layer where the isometric transformation module is located. The network layer of the multi-scale entanglement renormalization proposed network is equal to the scale. Each grid point of the multi-scale entanglement renormalization proposed network is used to store high-frequency coefficients or low-frequency coefficients at the corresponding position and scale.

6. The seat control method as described in claim 5, characterized in that, The training methods for the proposed multi-scale entanglement renormalization network include: Acquire sample data, which includes low-frequency signal after sample renormalization, high-frequency signal after sample renormalization, low-frequency signal, and high-frequency signal; The sample low-frequency signal and sample high-frequency signal are processed by an initial multi-scale entanglement renormalization simulation network to obtain the processed renormalized low-frequency signal and the processed renormalized high-frequency signal. The loss value is determined based on the low-frequency signal after sample renormalization, the high-frequency signal after sample renormalization, the low-frequency signal after renormalization, and the high-frequency signal after processing and renormalization. Based on the loss value, the parameters of the initial multi-scale entanglement renormalization simulated network are updated to obtain the trained multi-scale entanglement renormalization simulated network.

7. The seat control method according to any one of claims 1-3, characterized in that, After the control seat executes the seat control command corresponding to the audio signal, the method further includes: Acquire user adjustment data, which includes at least one of seat vibration intensity feedback data, seat vibration mode feedback data, and audio source feedback data; The matching relationship between the renormalized low-frequency signal and the renormalized high-frequency signal and the seat control command is adjusted based on the user adjustment data. New seat control commands are determined based on the newly determined renormalized low-frequency signal, the renormalized high-frequency signal, and the adjusted matching relationship, in order to control the seat.

8. A seat control device, characterized in that, The device includes: The audio signal acquisition module is used to acquire the audio signal to be played. The coefficient extraction module is used to extract low-frequency coefficients and high-frequency coefficients at different scales of the audio signal; The multi-scale entanglement renormalization simulation processing module is used to perform multi-scale entanglement renormalization simulation processing on the audio signal, all the low-frequency coefficients and high-frequency coefficients to obtain the renormalized low-frequency signal and the renormalized high-frequency signal. A seat control command generation module is used to determine seat control commands based on the renormalized low-frequency signal and the renormalized high-frequency signal; The control module is used to control the seat to execute the seat control command corresponding to the audio signal when the audio signal is played.

9. A seat control system, characterized in that, The seat control system includes a vehicle, a wavelet decomposition module, and a multi-scale entangled renormalization simulation processing network. The vehicle includes a seat, a seat controller, and an audio playback device, wherein: The wavelet decomposition module is used to extract low-frequency coefficients and high-frequency coefficients at different scales of the audio signal to be played by the audio playback device. The multi-scale entanglement renormalization simulation processing network is used to perform multi-scale entanglement renormalization simulation processing on the audio signal, all the low-frequency coefficients and high-frequency coefficients to obtain the renormalized low-frequency signal and the renormalized high-frequency signal. The seat controller is used to determine seat control commands based on the renormalized low-frequency signal and the renormalized high-frequency signal; The audio playback device is used to play the audio signal; The seat controller is also used to control the seat to execute the seat control command corresponding to the audio signal when the audio signal is played.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 7.

Images