Hierarchical operation method and system and application thereof

Abstract

The invention relates to the field of complex system calculation processing, and discloses a hierarchical operation method and system and application thereof.The hierarchical operation method specifically comprises the following steps that a three-layer operation structure is established, and an operation unit of a complex system is decomposed into an application layer, a module layer and an operator layer; constructing a functional module oriented to a final user in an application layer; constructing an algorithm module required for realizing functions of the application layer in the module layer; constructing an operation unit required by the support module layer in the operator layer; operation is executed in a top-to-bottom decomposition and bottom-to-top combination mode that the application layer calls the module layer, and the module layer calls the operator layer, so that hierarchical processing is realized. The maintainability, the expandability and the resource utilization rate of the system can be improved.

Description

Technical Field

[0001] The present invention relates to the field of complex system computing and processing, and particularly to a hierarchical operation method, system and its application. Background Art

[0002] A complex system refers to a system composed of a large number of components with complex interaction relationships among the components. Its operations usually involve multi-level and multi-dimensional computing tasks, and have characteristics such as high non-linearity, multi-variables, and multi-objectives. Complex systems widely exist in application scenarios such as digital signal processing, audio processing, image processing, machine learning, and communication systems. In these fields, the system needs to process a large amount of data, execute multiple algorithms, and achieve efficient computing under limited resource constraints.

[0003] With the continuous growth of application requirements, the computing complexity of the system has also increased, posing higher requirements for operation efficiency and resource utilization. Especially in embedded systems and real-time processing scenarios, how to efficiently implement complex functions under limited computing resources and energy consumption constraints has become a key challenge for technological development.

[0004] Existing complex system operations usually adopt an integrated design, integrating various functions and algorithms into a single module. Although this design method is relatively direct when implementing specific functions, it has obvious deficiencies in system maintenance and function expansion. When a certain function needs to be modified or a new function needs to be added, it often requires re-designing and validating the entire system, resulting in an extended development cycle and serious waste of resources.

[0005] Existing complex system operations also generally have the problem of low code reuse rate. Due to the lack of a unified design architecture, similar algorithms often need to be repeatedly implemented in different scenarios, increasing the development workload and the difficulty of code maintenance. This situation is particularly prominent in resource-constrained embedded systems. For systems with limited memory and computing resources, redundant code will significantly affect system performance.

[0006] Therefore, there is an urgent need to propose a hierarchical operation method, system and its application to solve the above problems. Summary of the Invention

[0007] The object of the present invention is to propose a hierarchical operation method, system and its application, which can improve the maintainability, scalability and resource utilization rate of the system.

[0008] To solve the above technical problems, the present invention provides a hierarchical operation method, which specifically includes the following: Establish a three-layer operation structure, and decompose the operation unit of the complex system into an application layer, a module layer and an operator layer; Construct a function module for the end user in the application layer; Build algorithm modules required to implement the functions of the application layer in the module layer; Build operation units required to support the module layer in the operator layer; Execute operations through the top-down decomposition and bottom-up combination methods of the application layer calling the module layer and the module layer calling the operator layer to achieve hierarchical processing.

[0009] Furthermore, the process of building function modules for end users in the application layer includes: Analyze user requirements and determine the functions that the function modules need to implement; Decompose the determined functions into sub-functions that can be independently implemented; Define standardized input parameters and output parameters for each function module.

[0010] Furthermore, when applied to an audio accelerator, the application layer construction includes at least one function module among an equalizer module, a dynamic range control module, a loudness control module, a time-domain smoothing processing module, and an anti-saturation superposition module.

[0011] Furthermore, the process of building algorithm modules required to implement the functions of the application layer in the module layer includes: Analyze the basic algorithms required for the implementation of the function modules in the application layer; Abstract each basic algorithm into a reusable algorithm module; Define a standardized interface compatible with the function modules in the application layer for each algorithm module.

[0012] Furthermore, when applied to an audio accelerator, the module layer construction includes at least one algorithm module among a linear-to-exponential domain conversion module, an exponential-to-linear domain conversion module, a moving average filter module, a gain slow change control module, a finite impulse response filter module, and an infinite impulse response filter module.

[0013] Furthermore, the process of building operation units required to support the module layer in the operator layer includes: Analyze the operation operations required for the implementation of the algorithm modules in the module layer; Abstract each operation operation into an independent operation unit; Define a standardized operation interface compatible with the algorithm modules in the module layer for each operation unit.

[0014] Furthermore, when applied to an audio accelerator, the operator layer construction includes at least one operation unit among a multiplier unit, an accumulator unit, a rounding unit, a saturation processing unit, a normalization unit, and a shift unit.

[0015] Further, the process of the application layer calling the module layer and the module layer calling the operator layer to perform operations in a top-down decomposition and bottom-up combination manner includes: Receiving a call request for the application layer function module; The application layer function module decomposes the algorithm combination required for function implementation according to the request; The application layer function module calls the corresponding algorithm module of the module layer; The module layer algorithm module decomposes the operation sequence required for algorithm implementation; The module layer algorithm module calls the corresponding operation unit of the operator layer; The operator layer performs the corresponding operation and returns the result to the module layer; The module layer aggregates the operation results and returns them to the application layer; The application layer integrates the processing results and outputs the final result.

[0016] In addition, the present invention also proposes a hierarchical operation system for implementing the hierarchical operation method as described above, including: An application layer for constructing function modules for end users; A module layer for constructing algorithm modules required to implement the functions of the application layer; An operator layer for constructing basic operation units required to support the module layer; and An inter-layer interface structure for realizing standardized information exchange between layers.

[0017] In addition, the present invention also proposes an application of the hierarchical operation method or the hierarchical operation system as described above in the field of audio processing.

[0018] Through the above technical solutions, the present invention has the following beneficial effects: By establishing a three-layer operation structure, the complex system is decomposed into an application layer, a module layer, and an operator layer, realizing hierarchical decomposition and combination of system functions, significantly improving the modularity and maintainability of the system. This hierarchical structure enables clear division of labor for each layer's functions, standardizes the interfaces, reduces the coupling degree between parts of the system. When modifying or expanding the system functions, only local adjustments need to be made at the corresponding layer without redesigning the entire system, greatly improving the system development efficiency and scalability. At the same time, by performing operations in a top-down decomposition and bottom-up combination manner, the system can flexibly combine resources of each layer according to specific tasks, improving the utilization rate of computing resources. Description of the Drawings

[0019] Figure 1 It is a flowchart of the hierarchical operation method in an embodiment of the present invention; Figure 2 A block diagram of a hierarchical operation system according to an embodiment of the present invention. Detailed implementation manners

[0020] Based on the revelation of this specification and without generating technical contradictions, those skilled in the art can form new technical solutions through cross-combination of different embodiments, and such variations should be regarded as falling within the protection scope of this patent.

[0021] The following will describe a hierarchical operation method, system and its application of the present invention in more detail with reference to the accompanying drawings, in which the preferred embodiments of the present invention are shown. It should be understood that those skilled in the art can modify the present invention described herein while still achieving the advantageous effects of the present invention. Therefore, the following description should be understood as broad knowledge for those skilled in the art and not as a limitation to the present invention.

[0022] In the following paragraphs, the present invention will be described more specifically by way of example with reference to the accompanying drawings. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the accompanying drawings are all in very simplified forms and use non-precise scales, only for the purpose of facilitating and clearly assisting in explaining the purpose of the embodiments of the present invention.

[0023] As Figure 1 - Figure 2 shown, an embodiment of the present invention proposes a hierarchical operation method, which specifically includes the following steps: S1. Establish a three-layer operation structure, and decompose the operation units of the complex system into an application layer, a module layer, and an operator layer; S2. Build a function module for the end user in the application layer; S3. Build an algorithm module required to implement the functions of the application layer in the module layer; S4. Build operation units required to support the module layer in the operator layer; S5. Execute operations through the top-down decomposition and bottom-up combination methods of the application layer calling the module layer and the module layer calling the operator layer to achieve hierarchical processing.

[0024] In this embodiment, by establishing a three-layer operation structure, the functions and operations of the complex system are decomposed in an orderly manner. Each layer has a clear function positioning and responsibility boundary, reducing the coupling degree between parts of the system. When it is necessary to modify or expand the system functions, only local adjustments need to be made at the corresponding layer without re-designing the entire system, improving the development efficiency and scalability of the system.

[0025] It should be noted that the hierarchical operation method proposed by the present invention is universal and can be applied to the design and implementation of various complex systems, including but not limited to image processing systems, speech recognition systems, machine learning systems, communication protocol stacks, industrial control systems, and other fields that require complex calculations. This method modularizes complex systems through a hierarchical structure, improves the cohesion of each part of the system, and reduces the coupling degree, making the system easier to develop, maintain, and expand. Hereinafter, the implementation manner of the present invention will be described in detail by taking the audio processing field as an example, but the application scope of the present invention is not limited thereto.

[0026] In step S2, the process of constructing the function modules for the end users in the application layer includes: analyzing the user requirements to determine the functions that the function modules need to implement; decomposing the determined functions into sub-functions that can be independently implemented; and defining standardized input parameters and output parameters for each function module.

[0027] In a specific example, the definition of the function modules in the application layer can adopt the object-oriented design method, and each function module is designed as a class with a clear interface. The methods of the class correspond to the function operations required by the user, and the attributes of the class correspond to the states and parameters of the functions. This design method enhances the cohesion and reusability of the function modules. It is known to those skilled in the art that the specific design method of the function modules can be selected according to actual requirements, and other design methods in addition to this embodiment are also included.

[0028] For example, in an image processing system, the application layer may include an image filtering module, an edge detection module, an image segmentation module, etc.; in a machine learning system, the application layer may include a data preprocessing module, a model training module, a prediction analysis module, etc.; in an industrial control system, the application layer may include a signal acquisition module, a control algorithm module, an actuator control module, etc. Each field constructs the corresponding function modules in the application layer according to its specific requirements, but the construction methods and principles are the same.

[0029] Since in the audio processing field, audio processing involves various algorithms such as filtering, transformation, and dynamic range control. These algorithms are both independent and related to each other. In the traditional audio processing field, these algorithms are usually implemented as independent modules, resulting in inconsistent interfaces between the algorithms, low data transmission efficiency, and difficulty in supporting complex audio processing requirements. In addition, when processing high-sampling-rate or multi-channel audio, the system resource consumption increases significantly, and the real-time processing requirements cannot be met. Therefore, this method can be applied to the audio processing field as follows.

[0030] In this embodiment, when applied to an audio accelerator, the application layer construction includes at least one functional module among an equalizer module, a dynamic range control module, a loudness control module, a time-domain smoothing processing module, and an anti-saturation superposition module. These modules directly correspond to specified functions in audio processing, enabling users to intuitively operate and control audio effects and improving the usability of the system.

[0031] Among them, the equalizer module is used to adjust the gain of the audio signal in different frequency bands to achieve frequency equalization of the sound; the dynamic range control module is used to compress or expand the dynamic range of the audio signal to make the sound more balanced or prominent; the loudness control module adjusts the audio signal according to the differences in the human ear's perception of sounds with different frequencies to make the listening experience consistent; the time-domain smoothing processing module is used to reduce mutations in the audio signal and improve the listening comfort; the anti-saturation superposition module is used to prevent signal overload when multiple audio signals are mixed. The setting of these modules enables the audio accelerator to support rich audio processing functions and meet the requirements of different application scenarios.

[0032] In step S3, the process of constructing the algorithm modules required to implement the functions of the application layer in the module layer includes: analyzing the basic algorithms required for the implementation of the application layer functional modules; abstracting each basic algorithm into a reusable algorithm module; and defining a standardized interface compatible with the application layer functional modules for each algorithm module. This design decouples the algorithm modules from the functional modules, and the same algorithm module can be shared and used by multiple functional modules, improving the code reuse rate and system efficiency.

[0033] In various complex systems, the design principles of the module layer remain consistent, but the specific implementation will vary according to the domain characteristics. For example, in an image processing system, the module layer may include a convolution operation module, a Fourier transform module, an interpolation algorithm module, etc.; in a machine learning system, the module layer may include a gradient descent optimization module, a feature extraction module, an activation function module, etc. These algorithm modules all follow the modular design principle and interact with the application layer through standardized interfaces.

[0034] In this embodiment, the interface design of the algorithm module can adopt a parameterized configuration method to implement different algorithm variants through different parameter combinations, increasing the flexibility of the algorithm module. For example, the filter module can implement different filter types and characteristics through parameter configuration. Those skilled in the art know that the specific interface design of the algorithm module can be adjusted according to actual needs and also includes other design methods besides this embodiment.

[0035] In a specific example, when applied to an audio accelerator, the construction of the module layer includes at least one algorithm module among a linear-to-exponential domain conversion module, an exponential-to-linear domain conversion module, a moving average filter module, a gain fade control module, a finite impulse response filter module, and an infinite impulse response filter module. These algorithm modules are the core computing units in audio processing. Through reasonable combination, various complex audio effects can be achieved, improving the computing efficiency and processing accuracy of the system.

[0036] Among them, the linear-to-exponential domain conversion module and the exponential-to-linear domain conversion module are used for the conversion of audio data between different representation domains, facilitating certain calculations in the exponential domain to improve efficiency; the moving average filter module is used to smooth the audio signal and reduce the influence of noise; the gain fade control module is used to achieve smooth transitions of volume or gain, avoiding popping caused by sudden changes; the finite impulse response filter module and the infinite impulse response filter module are the most basic filter types in audio processing, used to achieve frequency-selective filtering. The setting of these modules enables the modular implementation of audio processing algorithms, enhancing the maintainability and scalability of the system.

[0037] In step S4, the process of constructing the operator layer to support the operation units required by the module layer includes: analyzing the operation operations required for the implementation of the algorithm modules in the module layer; abstracting each operation operation into an independent operation unit; and defining a standardized operation interface compatible with the algorithm modules in the module layer for each operation unit.

[0038] In a specific example, when applied to an audio accelerator, the construction of the operator layer includes at least one operation unit among a multiplier unit, an accumulator unit, a rounding unit, a saturation processing unit, a normalization unit, and a shift unit. These operation units are the most basic computing components in audio processing. By optimizing these basic operations, the performance of the entire system can be significantly improved.

[0039] Among them, the multiplier unit is used to achieve the scaling and modulation of audio signals; the accumulator unit is used to achieve the accumulation and averaging of signals; the rounding unit is used to control numerical precision and bit-width conversion; the saturation processing unit is used to prevent distortion caused by numerical overflow; the normalization unit is used to adjust the signal amplitude to the standard range; and the shift unit is used to quickly implement multiplication and division operations of powers of 2. The optimized design of these basic operation units enables the efficient execution of audio processing algorithms, improving the real-time processing ability of the system.

[0040] In step S5, the process of the application layer calling the module layer and the module layer calling the operator layer to perform operations in a top-down decomposition and bottom-up combination manner includes: receiving a call request for the application layer function module; the application layer function module decomposing the algorithm combination required for function implementation according to the request; the application layer function module calling the corresponding algorithm module of the module layer; the module layer algorithm module decomposing the operation sequence required for algorithm implementation; the module layer algorithm module calling the corresponding operation unit of the operator layer; the operator layer performing the corresponding operation and returning the result to the module layer; the module layer summarizing the operation result and returning it to the application layer; and the application layer integrating the processing result and outputting the final result. This way of layer-by-layer decomposition and combination decomposes complex operations into simple steps, improving the comprehensibility and maintainability of the system.

[0041] This top-down decomposition and bottom-up combination execution method is applicable to various complex systems, whether it is a real-time control system, a data analysis system, or a multimedia processing system. Through a unified hierarchical call mechanism, systems in different fields can adopt the same architecture design principle, but implement different function modules, algorithm modules, and operation units according to specific requirements.

[0042] In a specific example, asynchronous processing mechanism can be adopted for inter-layer calls, so that the upper layer does not have to wait for the lower layer to complete after sending a request and can continue to process other tasks, improving the parallel processing ability of the system. At the same time, a caching mechanism can be set to cache frequently used intermediate results, reduce repeated calculations, and improve processing efficiency. Those skilled in the art know that the specific implementation method of inter-layer calls can be designed according to system requirements and also includes other implementation methods besides this embodiment.

[0043] In addition, this embodiment also proposes a hierarchical operation system for implementing the hierarchical operation method as described above, including: an application layer for constructing function modules for end users; a module layer for constructing algorithm modules required to implement the functions of the application layer; an operator layer for constructing basic operation units required to support the module layer; and an inter-layer interface structure for realizing standardized information exchange between layers.

[0044] In addition, this embodiment also proposes an application of the hierarchical operation method or the hierarchical operation system as described above in the field of audio processing. Specifically as follows: In addition, this embodiment also proposes an audio accelerator, including: a data storage unit for storing audio data and intermediate calculation results; an address generation unit for generating and managing the storage addresses of audio data; a data operation unit implemented by using the hierarchical operation system as described above; and a control unit for coordinating the data flow and processing flow between each unit to realize hierarchical processing of audio data.

[0045] In addition, this embodiment also proposes an audio processing method, which uses the audio accelerator as described above, and specifically includes the following steps: S100. By establishing a three-layer operation structure, decompose audio processing into an application layer, a module layer, and an operator layer, where the application layer constructs an audio processing function module for users, the module layer constructs algorithm modules required to implement the functions of the application layer, and the operator layer constructs operation units required to support the module layer; S200. Generate an address sequence through an address generation unit and load audio data into a data storage unit; S300. Use a data operation unit to process and operate on the audio data in a top-down decomposition and bottom-up combination manner in which the application layer calls the module layer and the module layer calls the operator layer; and S400. Output the processed audio data from the data storage unit.

[0046] In this embodiment, when this method is applied to the field of audio processing, first, the control unit receives processing instructions and parameters from the main control processor; then, the address generation unit generates an access address for audio data and loads the data from the main memory into the data storage unit; then, the data operation unit executes an audio processing algorithm according to the three-layer structure, and the processing result is stored in the data storage unit; finally, the processed data is written back to the main memory under the control of the address generation unit to complete the entire processing process. This pipelined processing method enables each unit to work together and improves the overall processing efficiency.

[0047] This method can be applied to various fields that require complex calculations. By means of a three-layer operation structure, a complex system is decomposed into an application layer, a module layer, and an operator layer, realizing hierarchical decomposition and combination of system functions, and significantly improving the modularity, maintainability, and scalability of the system. The application of this method in the field of audio processing is only one of its many application scenarios, and its basic principle and design concept can be extended to the design of complex systems in other fields, providing an effective solution for improving system efficiency, flexibility, and maintainability.

[0048] In summary, a hierarchical operation method, system, and its application proposed by the present invention have the following advantages: By establishing a three - layer operation structure, the complex system is decomposed into an application layer, a module layer, and an operator layer, realizing the hierarchical decomposition and combination of system functions, significantly improving the modularity and maintainability of the system. This hierarchical structure enables clear division of labor for each layer's functions, standardizes the interfaces, reduces the coupling degree between different parts of the system. When modifying or expanding the system functions, only local adjustments need to be made at the corresponding layer without redesigning the entire system, greatly improving the system's development efficiency and scalability. At the same time, by performing operations in a top - down decomposition and bottom - up combination manner, the system can flexibly combine resources of each layer according to specific tasks, improving the utilization rate of computing resources.

[0049] In addition, in the specific implementation applied to an audio accelerator, the three - layer structure specifically realizes the hierarchical decomposition from specific functions (such as equalizers, dynamic range controls, etc.) to algorithm modules (such as filters, conversion modules, etc.) and then to basic operation units (such as multipliers, accumulators, etc.), enabling complex audio processing tasks to be efficiently completed. Especially through the standardized interface design, seamless connection between various functional modules, algorithm modules, and operation units is achieved, greatly improving the code reuse rate and the flexibility of the system. The data operation unit is implemented using a hierarchical system, which, in cooperation with other hardware units such as data storage units, address generation units, etc., forms a complete audio accelerator system that can efficiently process various audio tasks and meet the complex computing requirements in resource - constrained environments.

[0050] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A hierarchical operation method, characterized in that, Specifically, it includes the following: Establish a three - layer operation structure, and decompose the operation units of a complex system into an application layer, a module layer, and an operator layer; Construct function modules for end - users in the application layer; Construct algorithm modules required to implement the functions of the application layer in the module layer; Construct operation units required to support the module layer in the operator layer; Execute operations through the top - down decomposition and bottom - up combination methods where the application layer calls the module layer and the module layer calls the operator layer, to achieve hierarchical processing.

2. The hierarchical operation method according to claim 1, wherein The process of constructing function modules for end - users in the application layer includes: Analyze user requirements and determine the functions that the function modules need to implement; Decompose the determined functions into sub - functions that can be independently implemented; Define standardized input parameters and output parameters for each function module.

3. The hierarchical operation method according to claim 2, wherein, When applied to an audio accelerator, the construction of the application layer includes at least one of the function modules such as an equalizer module, a dynamic range control module, a loudness control module, a time - domain smoothing processing module, and an anti - saturation superposition module.

4. The hierarchical operation method according to claim 1, characterized in that, The process of constructing algorithm modules required to implement the functions of the application layer in the module layer includes: Analyze the basic algorithms required for the implementation of the function modules in the application layer; Abstract each basic algorithm into a reusable algorithm module; Define standardized interfaces compatible with the function modules in the application layer for each algorithm module.

5. The hierarchical operation method according to claim 4, wherein When applied to an audio accelerator, the construction of the module layer includes at least one of the algorithm modules such as a linear - to - exponential domain conversion module, an exponential - to - linear domain conversion module, a moving average filter module, a gain slow - change control module, a finite impulse response filter module, and an infinite impulse response filter module.

6. The hierarchical operation method according to claim 4, wherein The process of constructing operation units required to support the module layer in the operator layer includes: Analyze the operation operations required for the implementation of the algorithm modules in the module layer; Abstract each operation operation into an independent operation unit; Define standardized operation interfaces compatible with the algorithm modules in the module layer for each operation unit.

7. The hierarchical operation method according to claim 6, wherein When applied to an audio accelerator, the construction of the operator layer includes at least one of the operation units such as a multiplier unit, an accumulator unit, a rounding unit, a saturation processing unit, a normalization unit, and a shift unit.

8. The hierarchical operation method according to claim 1, wherein, The process of executing operations through the top - down decomposition and bottom - up combination methods where the application layer calls the module layer and the module layer calls the operator layer includes: Receive a call request for the function module in the application layer; The function module in the application layer decomposes the algorithm combination required for function implementation according to the request; The function module in the application layer calls the corresponding algorithm module in the module layer; The algorithm module in the module layer decomposes the operation sequence required for algorithm implementation; The algorithm module in the module layer calls the corresponding operation unit in the operator layer; The operator layer executes the corresponding operation operations and returns the results to the module layer; The module layer aggregates the operation results and returns them to the application layer; The application layer integrates the processing results and outputs the final results.

9. A hierarchical operation system for implementing the hierarchical operation method according to any one of claims 1-8, characterized in that, It includes: An application layer, which is used to construct function modules for end - users; A module layer, which is used to construct algorithm modules required to implement the functions of the application layer; An operator layer, which is used to construct basic operation units required to support the module layer; And An inter - layer interface structure, which is used to achieve standardized information exchange between layers.

10. Use of a hierarchical operation method as described in any one of claims 1-8 or a hierarchical operation system as described in claim 9 in the field of audio processing.

Images