CN-121279235-B - Intelligent layout and connection system and method for DSP module

CN121279235BCN 121279235 BCN121279235 BCN 121279235BCN-121279235-B

Abstract

The invention relates to the technical field of electric digital data processing, in particular to an intelligent layout and connection system and method of a DSP module, wherein the system comprises a DSP dragging connection unit, a DSP collision avoidance unit and a DSP automatic layout unit, the DSP drags the connection unit background to sense the dragging track in real time and automatically matches proper connection objects in candidate pins of the original layout. The DSP collision avoidance unit executes collision detection on the dragging algorithm module and the original layout and other modules, judges the collision state, and automatically offsets the module position according to the collision direction once the overlap is found. The DSP automatic layout unit calculates the hierarchical sequence and the maximum hierarchical number of each module by utilizing a depth-first search algorithm based on the latest topology updating pin state after the connection is completed, and automatically arranges input, processing and output modules according to the rules from left to right to generate an integral layout with clear structure and regular alignment, thereby obviously reducing the burden of manual alignment and connection arrangement.

Inventors

ZHU JIAN

Assignees

深圳同创音频技术有限公司

Dates

Publication Date: 20260508
Application Date: 20251208

Claims (9)

The intelligent layout and connection system of the DSP module is characterized by comprising the following components: The DSP dragging connection unit is used for generating an original layout of a fixed connection on the visual configuration interface based on the fixed configuration of the DSP equipment, and a dragging track of the background sensing DSP algorithm module automatically searches for the connection between pins and the original layout and enters the DSP collision avoidance unit; The DSP collision avoidance unit is used for marking the dragging DSP algorithm module as a current dragging module, performing collision detection on the current dragging module, judging the collision state of the current dragging module, automatically avoiding the module if collision exists, and entering the DSP automatic layout unit; the DSP automatic layout unit is used for updating pins in the dragging process, the current dragging module is used for connecting wires based on the updated pins, and after the wires are completed, a depth-first search algorithm is used for carrying out automatic layout of the DSP algorithm module; Updating pins in the dragging process, and connecting the current dragging module based on the updated pins, wherein the specific analysis process is as follows: the method comprises the steps of intelligently searching modules to be connected in an original layout, and marking the modules to be connected as target connection modules; After collision detection and automatic avoidance are completed, calculating shortest distance pins between each pin of the target connection module and the pin of the current dragging module based on a shortest distance pin calculation formula, setting the shortest distance pins to be in a to-be-connected state, changing pin colors, and recording the shortest distance pins in a pin hash set; Updating the module to be connected in the real-time dragging process, updating the shortest distance pins of the module to be connected based on a shortest distance pin calculation formula, recording the updating result of each pin by the pins Ha Xiji, removing the last shortest distance pin, and restoring the color of the pins; When the dragging is finished, extracting the shortest distance pins from the pin hash set, connecting the current dragging module with a module in the original layout corresponding to the shortest distance pins, and restoring the colors of the pins after the connection is finished; the DSP algorithm module is an independent functional unit which is formed by packaging equalization, filtering, automatic sound mixing, voltage limiting, time delay, acoustic echo cancellation and background music management in a digital signal processing system, and externally presents a block with a plurality of input/output pins, configurable gain, frequency points, threshold values, time constants and fixed algorithm logic.
2. The intelligent layout and connection system for DSP modules according to claim 1, wherein the visual configuration interface comprises a left side input module group layout, a middle processing module group layout and a right side output module group layout; The left input module group is arranged as a component column, and optional DSP algorithm modules are displayed in a list or icon mode; The intermediate processing module group is arranged as a topology editing area, and the original layout of the fixed connection lines is displayed; the right side output module group layout is the DSP algorithm module and original layout DSP algorithm combination layout, and the processed signals are sent to the corresponding physical or network output channels.
3. The intelligent layout and connection system of the DSP module according to claim 2, wherein the generating an original layout of a fixed connection line on the visual configuration interface comprises the following specific analysis processes: the original layout comprises a local sound amplifying mode original layout, a local-remote mode original layout and a background music system mode original layout; The original layout of the local sound expansion mode comprises a left input module, an automatic mixing module, a middle matrix mixing module and a right output module; The local-remote mode original layout comprises a left input module, a matrix mixing module, an acoustic echo cancellation module, a middle matrix mixing module and a right output module; The original layout of the background music system mode comprises a left input module, a background music management module, a middle matrix sound mixing module and a right output module.
4. The intelligent layout and connection system for DSP modules according to claim 1, wherein said collision detection is performed on the current drag module to determine the collision status of the current drag module, and the specific analysis process is as follows: after the original layout is completed, a dragging track of a DSP algorithm module is perceived based on a background of a left input module group to which the visual configuration interface belongs, the DSP algorithm module is generated through a factory mode, and is a module dragged on an intermediate processing module group to which the visual configuration interface belongs, and the DSP algorithm module is set as a current dragging module; Setting pins of a current dragging module to be in a state to be connected, changing the colors of the pins, and recording the positions of the current dragging module before dragging in a hash set; and performing collision detection when the current dragging module moves, searching all modules which can collide by using a breadth-first search algorithm, detecting whether the current dragging module is overlapped with the modules in the original layout, and judging the collision state of the current dragging module based on the detection result.
5. The intelligent layout and connection system for a DSP module according to claim 4, wherein the automatic avoiding of the module is performed if a collision exists, and the specific analysis process is as follows: Based on collision codes, intersecting detection is carried out on a style matrix which the current dragging module belongs to and a style matrix which the original layout module belongs to, and the detection result is expressed as that if the style matrix which the current dragging module belongs to is intersected with the style matrix which the original layout module belongs to, the collision state of the current dragging module is judged to be overlapped, otherwise, the collision state of the current dragging module is judged to be non-overlapped; If the collision states of the current dragging modules are overlapped, the module currently collided in the original layout is recorded as a collision module, the original position of the collision module is recorded to Ha Xiji, the moving direction is judged according to the center point of the current dragging module and the collision module, the moving distance is calculated based on a moving distance formula, and automatic avoidance of the module is completed.
6. The intelligent layout and connection system for a DSP module according to claim 5, wherein said automatically avoiding the module if a collision occurs further comprises: Extracting a collision module recorded in the hash table, performing collision detection with the current dragging module again, if a certain collision module marked in the hash table still overlaps with the current dragging module, automatically executing avoidance by the current dragging module, and enabling the collision feedback effect of the collision module to be unchanged, otherwise, restoring the collision module to an original position, and removing Ha Xiji the collision module.
7. The intelligent layout and connection system for DSP modules according to claim 1, wherein the automatic layout of the DSP algorithm modules is performed by using a depth-first search algorithm, and the specific analysis process is as follows: After connection is completed, automatic layout is performed from left to right, starting from each pin of each left input module, traversing all reachable connection paths between the left input module and the middle matrix audio mixing module by using a depth-first search algorithm, counting the number of modules passing through each path, finding out the path with the largest number of modules, and taking the maximum module layer number corresponding to the path as a layout reference; Dividing a plurality of equidistant layout grids in the X-axis direction according to the maximum module layer number to obtain uniform module layer-to-layer horizontal distance, sequentially placing each left input module and the middle matrix mixing module at the corresponding X-axis grid position according to the layer index of the pins of the left input module and the middle matrix mixing module to finish the left-to-right alignment layout between the left input module and the middle matrix mixing module; The left-to-right alignment layout between the middle matrix mixing module and the right output module is accomplished using a depth-first search algorithm.
8. The intelligent layout and connection system for a DSP module according to claim 4, wherein said collision detection is performed on the current drag module, and determining the collision status of the current drag module further comprises: If the plurality of DSP algorithm modules are copied and pasted, the plurality of DSP algorithm modules are marked as a dragging module group, and dragging and layout processing is carried out on the dragging module group: After the copying operation is triggered, sequencing the modules in the dragging module group according to the sequence from left to right and from top to bottom, extracting the modules positioned at the leftmost upper part and the rightmost upper part based on the sequencing result, and performing external optimal pin searching on the pins of the modules to match with the pins of the target connection module; When the dragging is finished, a connection relation among the modules is established in the dragging module group according to the sequencing result, then the dragging module group is automatically connected with the external optimal pins, connection between the dragging module group and the original layout is completed, and an automatic layout process is triggered after the internal connection and the external connection are completed.
An intelligent layout and connection method for a DSP module, applied to the intelligent layout and connection system for a DSP module according to any one of claims 1 to 8, comprising: s1, generating an original layout of a fixed connection line on a visual configuration interface based on fixed configuration of a DSP device, automatically searching the connection line between pins and the original layout by a dragging track of a background sensing DSP algorithm module, and entering S2; S2, marking the dragging DSP algorithm module as a current dragging module, performing collision detection on the current dragging module, judging the collision state of the current dragging module, and if collision exists, performing automatic module avoidance and entering S3; s3, updating pins in the dragging process, connecting wires by the current dragging module based on the updated pins, and automatically laying out the DSP algorithm module by using a depth-first search algorithm after connecting wires are completed.

Description

Intelligent layout and connection system and method for DSP module Technical Field The invention relates to the technical field of electric digital data processing, in particular to an intelligent layout and connection system and method of a DSP module. Background DSP (DIGITAL SIGNAL Processing) refers to a technology for digitally Processing collected continuous signals (such as sound, image, sensor data, etc.). The method processes, optimizes or analyzes the signals by utilizing various mathematical algorithms (such as filtering, fourier transformation, compression, feature extraction and the like) after converting the analog signals into digital signals, thereby extracting useful information, removing noise, compressing data or realizing other functions. The DSP module layout and connection refers to selecting and setting proper processing modules, algorithm parameters and hardware parameters for the digital signal processing system according to specific application requirements. This includes determining the processing flow of the signal, setting the filter order, sampling rate, quantization accuracy, etc. algorithm parameters, and configuring hardware resources such as sampling frequency, bit width, buffer size, hardware registers, etc. Through configuration, the DSP system is ensured to efficiently and accurately execute tasks, and simultaneously, the resource and performance requirements are met. The existing connection scheme ensures that the transmission path of signals from an input end to an output end is clear and efficient by connecting a plurality of DSP algorithm modules according to the signal flow and the processing sequence. In this process, parameters (such as sampling rate, cut-off frequency, gain value, etc.) of each module need to be set according to the system requirements, and meanwhile, synchronization of data transmission timing among the modules needs to be ensured. Finally, through optimization and inspection, the system is ensured to meet the design requirements in terms of resources, time sequence, performance and the like. An operation system based on a DSP chip array is disclosed in chinese patent publication No. CN114185599B, and comprises a DSP chip array and an array support unit connected by a hardware link, wherein the DSP chip array includes a plurality of DSP chips, the hardware link includes a number of computation data flow channels equivalent to the plurality of DSP chips and a number of loading maintenance data flow channels equivalent to the plurality of DSP chips, each DSP chip in the DSP chip array is connected to the array support unit by a corresponding computation data flow channel and loading maintenance data flow channel, and the array support unit includes a loading maintenance data flow channel for completing programming and program guidance, and a computation data flow channel for completing data sharing and algorithm acceleration between the plurality of DSP chips. For example, a signal processing SIP design circuit based on multiple DSPs disclosed in Chinese patent publication No. CN118095193B relates to the technical field of microelectronics; the circuit comprises N pieces of multi-core DSP chips, N pieces of FLASH chips, a storage module, N pieces of DDR memory interfaces, N pieces of voltage regulating chips, N pieces of clock chips and an ABF substrate, wherein the multi-core DSP is used for connecting the storage module and calculating and caching, the storage module is used for loading programs and storing data, the SPI interface is connected with the FLASH chips and the multi-core DSP chips, the voltage regulating chips are used for providing a working power supply, the clock chips are used for providing clocks, the chips and the modules are distributed on the ABF substrate in a peripheral arrangement mode, matrix data stored in the N pieces of DSP are scheduled in an EDMA moving mode and are integrated into a final required matrix, and the purpose of accelerating algorithm is achieved. In the prior art, most of the solutions are based on a DSP chip array, a multi-DSP package or a two-dimensional array deployment solution with fixed topology and static connection, or only basic drag-and-drop and manual connection of processing modules are provided in a graphical interface, and the actual connection process of the DSP modules often needs to frequently insert, delete and replace algorithm modules, dynamically adjust signal paths and coordinate the overall layout of a plurality of groups of modules, so that the connection is clear and orderly, and the resource and time sequence constraint is satisfied. Because the complexity is usually supported only through the interface operation which is relatively primary in the existing configuration software, the number of modules is continuously increased, the dragging and copying and pasting are required to keep real-time response, various operations are highly coupled in terms of space position and topolo