Intelligent set top box system based on multi-channel synchronous processing

Abstract

The invention provides an intelligent set top box system based on multi-channel synchronous processing, and relates to the technical field of set top box processing systems.The system architecture comprises a hardware layer, a software layer and a user interaction layer, the hardware layer conducts special core group division according to different links of multi-channel signal processing, and the software layer conducts core group division according to different links of multi-channel signal processing; decoding tasks can be reasonably distributed according to channel priorities and core loads, smooth playing of high-priority channels is guaranteed, a decision tree is constructed according to a large number of off-line signal samples based on a decision tree and signal feature analysis method to accurately judge interference types, measures are taken in a targeted mode, and the interference quality is improved. The intelligent resource management subsystem sets different resource thresholds, monitors system loads and task resource requirements of all channels in real time, establishes multi-dimensional content tags, constructs user interest portraits in combination with multi-dimensional data such as user watching history, mines potential associations among different channel programs, generates personalized recommendation lists and knowledge expansion, and provides a personalized recommendation system for the users. And the viewing field of the user is expanded.

Description

Technical Field

[0001] The present invention relates to the technical field of set-top box processing systems, in particular to an intelligent set-top box system based on multi-channel synchronous processing. Background Art

[0002] According to a method for fast switching channels under multiple IPTV platforms disclosed in China Publication No. "CN107682718A", the method relates to the technical field of set-top boxes. The method is as follows: when receiving a signal from a user to switch channels, the set-top box requests a unicast media stream from an FCC server corresponding to the channel information according to the channel information of the new channel; the FCC server sends the latest media stream content cached by the corresponding new channel to the set-top box in the form of a unicast media stream; after receiving the unicast media stream, the set-top box starts decoding and displays the picture; the FCC server sends a command to join the multicast media stream to the set-top box; the set-top box joins the multicast group, receives the multicast media stream sent by the IPTV platform, analyzes the characteristics of the unicast media stream sent by the FCC server corresponding to the channel information, and realizes switching from the unicast media stream to the multicast media stream. The present invention also discloses a system for fast switching channels under multiple IPTV platforms.

[0003] The above patent documents and prior art have the following technical problems when used:

[0004] Problem 1: Existing set-top boxes usually adopt a single-core or simple multi-core architecture. When faced with multi-channel simultaneous processing tasks, there is a lack of refined core allocation strategies for different channel processing links. Decoding tasks are often allocated according to a fixed order or simple rules, and it is difficult to dynamically adjust according to the actual priority of the channel and the real-time load of the system. This leads to decoding delays, freezes, and even system crashes when the number of channels increases, the signal quality fluctuates, or the user switches channels, which cannot meet the user's needs for smooth viewing of multi-channel programs.

[0005] Problem 2: Traditional set-top boxes usually use relatively fixed and single filtering or error correction technology when dealing with signal interference. They lack the ability to accurately identify different types of interference signals and cannot perform targeted processing based on the specific characteristics of the interference. For example, when faced with complex and changeable adjacent channel interference and pulse interference, they cannot effectively and dynamically adjust filter parameters or use appropriate signal reconstruction technology, which results in a decrease in signal quality and affects the user's viewing experience. Problems such as snow and horizontal stripes may appear on the screen and noise may occur on the sound.

[0006] Problem 3: The previous set-top box resource management solution was relatively rigid. Regardless of the system load status, the allocation and operation mode of hardware resources were basically fixed. It was impossible to quickly and effectively allocate resources to meet different user load requirements, resulting in system performance degradation, unsmooth playback, slow screen loading, and other problems, affecting users' normal use of set-top box functions.

[0007] Question 4: The recommendation function of existing set-top boxes is mostly based on simple program classification or popular recommendations within a single channel. It does not fully explore the potential connections between programs on different channels, lacks in-depth analysis of users' multi-dimensional viewing data and integration with rich external information sources, and is unable to build accurate user interest portraits. It is difficult to provide personalized recommendation content across channels, resulting in poor interactivity between users and set-top boxes and reduced user stickiness. Summary of the invention

[0008] Technical issues solved

[0009] In view of the shortcomings of the prior art, the present invention provides an intelligent set-top box system based on multi-channel synchronous processing, which solves the following problems:

[0010] 1. Address the problem of users not being able to watch smoothly due to insufficient multi-channel processing efficiency and stability;

[0011] 2. The set-top box system has limited ability to deal with signal interference, resulting in reduced signal quality;

[0012] 3. Address the problem of insufficient adaptability of resource management and inability to effectively allocate resources;

[0013] 4. Address the issues of lack of multi-dimensional analysis and insufficient interactive depth in content recommendations.

[0014] Technical Solution

[0015] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent set-top box system based on multi-channel synchronous processing, the system architecture includes a hardware layer, a software layer and a user interaction layer, wherein:

[0016] The hardware layer is centered on a distributed multi-core processor module, which includes core groups such as signal processing, decoding, and graphics processing, which are responsible for the tasks of preliminary processing of multi-channel signal reception, decoding, and image rendering layout optimization. The storage unit provides support for data temporary storage and system software and related data storage, and the communication interface module ensures the connection with external devices and networks.

[0017] The software layer includes a multi-channel synchronous decoding subsystem, a multi-channel signal interference elimination subsystem, an intelligent resource management subsystem, and a multi-channel content association and intelligent recommendation subsystem. The multi-channel synchronous decoding subsystem uses an algorithm based on priority queues and dynamic load balancing strategies to allocate decoding tasks according to channel priorities and core load conditions, and has error detection and correction capabilities. The multi-channel signal interference elimination subsystem adopts a method based on decision trees and signal feature analysis to determine the type of interference and eliminate it in a targeted manner by building a decision tree. The intelligent resource management subsystem uses a threshold-based resource monitoring and allocation algorithm to dynamically adjust hardware resources according to system load. The multi-channel content association and intelligent recommendation subsystem adopts an algorithm based on content tags and collaborative filtering to achieve cross-channel content association and personalized recommendation by building tags, analyzing user data and collaborative filtering.

[0018] The user interaction layer includes a traditional interaction interface module and an intelligent interaction interface module. The traditional interaction interface module provides basic operating functions and works in conjunction with some subsystems of the software layer. The intelligent interaction interface module uses VR and AR technologies and gesture and voice interaction functions to provide users with an immersive multi-channel interaction experience, and uses interaction data feedback for recommended content optimization.

[0019] Preferably, the hardware layer includes a distributed multi-core processor module, a storage unit and a communication interface module. The distributed multi-core processor module is composed of multiple core groups, including a signal processing core group, a decoding core group, and a graphics processing core group. The signal processing core group is responsible for the reception, filtering and preliminary processing of multi-channel signals. It is equipped with a high-speed signal receiver and an advanced filter unit, and can process signal inputs of multiple channels at the same time to ensure stable reception and preliminary purification of signals.

[0020] Preferably, the multi-channel synchronous decoding subsystem of the software layer sets up a priority queue for each channel, and the priority is determined based on the user's viewing history, channel subscription status, and current signal quality factors. For example, the channel being watched and the channels that the user often watches and pays for have higher priority. The system then monitors the load of the decoding core group in real time and assigns the channel decoding tasks at the head of the queue to the decoding core with a lighter load. If the load of all cores is high, the decoding tasks of the low-priority channels are appropriately delayed or the frame rate is reduced based on the priority to ensure smooth playback of the high-priority channels. At the same time, the subsystem also has a decoding error detection and correction function. When an error occurs in the decoding of a channel, it can try to switch to different decoding strategies or use redundant information to repair the error to ensure the stability of multi-channel decoding.

[0021] Preferably, the multi-channel signal interference elimination subsystem of the software layer utilizes the construction of a decision tree to obtain the relationship between different interference types and signal characteristics based on the analysis of a large number of offline signal samples. When a multi-channel signal is received, the system extracts the time domain and frequency domain characteristics of the signal, such as signal amplitude changes, frequency distribution, etc., and inputs these characteristics into the decision tree to determine the type of interference. For adjacent channel interference, instructions are sent to the signal processing core group according to the interference range determined by the decision tree, and the filter parameters are dynamically adjusted to accurately filter out the interference frequency components. For pulse interference, it works in coordination with the decoding core group, and utilizes signal reconstruction technology to restore the signal of the interfered part according to the correlation between the previous and next signals, thereby ensuring the quality of the multi-channel signal.

[0022] Preferably, the intelligent resource management subsystem of the software layer pre-sets thresholds for different resources, such as core frequency, memory usage, etc., and monitors the resource requirements of each channel processing task in each core group of the hardware layer and the load of the overall system in real time. When the system is in a low-load state, for example, the user only watches a few standard-definition channels and has no complex interactive operations, if the resource usage is lower than the set threshold, an instruction is sent to the distributed multi-core processor module of the hardware layer to reduce the operating frequency and voltage of some unused or low-load cores, or even shut down some unnecessary hardware modules, such as idle decoding chip channels or graphics processing special effects units to save energy. When the system load suddenly increases, such as the user switches to multiple high-definition channels or performs complex interactive operations, such as multi-screen interaction, virtual reality interaction, etc., if the resource usage exceeds the threshold, the dormant resources are quickly awakened and the resources are reallocated to each task. For example, more decoding core resources and graphics processing resources are allocated to the newly switched high-definition channel to ensure that the system performance is not affected. At the same time, the allocation strategy of the cache memory is reasonably adjusted to give priority to the data storage needs of the current active channel.

[0023] Preferably, the multi-channel content association and intelligent recommendation subsystem of the software layer establishes multi-dimensional content tags for each channel program, including program type, theme, actor, and director information, and then builds a user interest portrait by collecting and analyzing multi-dimensional data such as user viewing history, viewing time, channel stay time, and operating habits. At the same time, data interaction is performed with content databases and social platforms on the Internet to obtain more program information and social interaction data. When a user is watching a program on a certain channel, the subsystem analyzes the program content of other channels in the background, matches the user interest portrait according to the content tags, and mines the potential associations between programs on different channels. For example, when a user is watching a science fiction movie on a movie channel, the subsystem searches for related science fiction-themed content on news channels, technology channels, documentary channels, etc., such as movie special effects technology analysis in news reports, science fiction concept popular science programs on science fiction channels, and science fiction film production behind-the-scenes documentaries on documentary channels. Then, according to the collaborative filtering algorithm, these related contents are sorted in combination with the viewing behavior of other similar users to generate a personalized recommendation list, and the recommended content is displayed on the user interface in an intuitive manner to guide users to discover more multi-channel content of interest.

[0024] Preferably, the traditional interactive interface module of the user interaction layer provides basic channel list display, program information query, playback control, play, pause, fast forward, rewind, volume adjustment and other functions. The user performs conventional operations such as channel switching and program selection in the traditional interactive interface by operating the buttons on the remote control. The module works closely with the multi-channel synchronous decoding subsystem and the intelligent resource management subsystem of the software layer. When the user switches channels, the relevant subsystems are notified in time to allocate resources and adjust decoding tasks.

[0025] Preferably, the intelligent interaction interface module of the user interaction layer includes an immersive interaction interface based on virtual reality (VR) and augmented reality (AR) technology, as well as gesture and voice interaction functions. In the immersive interaction interface, the user views the real-time images of multiple channels in a three-dimensional and dynamic form through a head-mounted display device (VR) or a mobile device such as a mobile phone (AR). Through a gesture recognition device or a sensor of a mobile phone, the user can switch between different channels and adjust the size and layout of the channel screen. The voice interaction function allows the user to perform operations such as channel search, program play, and volume adjustment through voice commands. For example, the user can say commands such as "switch to the sports channel" and "play the next recommended program". The intelligent interaction interface module parses the voice commands and passes them to the corresponding subsystem of the software layer for execution. The module can also feed back the user's behavior data during the intelligent interaction process to the multi-channel content association and intelligent recommendation subsystem to further optimize the recommended content.

[0026] Beneficial Effects

[0027] The present invention provides an intelligent set-top box system based on multi-channel synchronous processing. It has the following beneficial effects:

[0028] 1. The present invention uses the system hardware layer to divide the different links of multi-channel signal processing into special core groups. For example, the signal processing, decoding and graphics processing core groups work together to efficiently process tasks such as signal reception, decoding and picture rendering layout optimization of multiple channels at the same time. The multi-channel synchronous decoding subsystem of the software layer is based on the priority queue and dynamic load balancing strategy. It can reasonably allocate decoding tasks according to channel priority and core load, ensure the smooth playback of high-priority channels, and flexibly adjust low-priority channels when the core load is high, effectively improving the efficiency and stability of the overall multi-channel synchronous processing, ensuring that users will not experience freezes or delays when switching and watching multiple channels, and greatly improving the user viewing experience.

[0029] 2. The present invention adopts a system that utilizes a multi-channel signal interference elimination subsystem and adopts a method based on decision trees and signal feature analysis. A decision tree is constructed based on a large number of offline signal samples to accurately determine the type of interference. Whether it is adjacent channel interference or pulse interference, targeted measures can be taken, such as dynamically adjusting filter parameters in the case of adjacent channel interference, and using signal reconstruction technology in collaboration with the decoding core group to restore the signal in the case of pulse interference, effectively removing interference, ensuring high-quality transmission and reception of multi-channel signals, making the picture of each channel clear and the sound smooth, reducing problems such as picture snow and noisy sound caused by signal interference, and providing users with stable and reliable multi-channel audio-visual services.

[0030] 3. The present invention adopts the intelligent resource management subsystem in the system to set different resource thresholds to monitor the system load and the resource requirements of each channel task in real time. Under low-load conditions, such as when the user watches a small number of standard-definition channels and has no complex operations, the frequency and voltage of unused or low-load cores can be reduced, and even unnecessary hardware modules can be turned off to achieve energy saving and consumption reduction. In high-load scenarios, such as switching to multiple high-definition channels or performing complex interactive operations, dormant resources can be quickly awakened and reasonably allocated to ensure that system performance is not affected, while optimizing the cache memory allocation strategy. This intelligent resource allocation mechanism not only improves system resource utilization, but also reduces energy consumption, extends the service life of the set-top box, and reduces user usage costs. It conforms to modern energy-saving and environmental protection concepts and ensures the stable operation of the system under different working conditions.

[0031] 4. The present invention uses a system to establish multi-dimensional content tags for programs through multi-channel content association and intelligent recommendation subsystems, build user interest portraits in combination with multi-dimensional data such as user viewing history, and interact with Internet content databases and social platforms to obtain more information. Through content tag matching and collaborative filtering algorithms, it can explore potential associations between programs on different channels and generate personalized recommendation lists and knowledge expansion. For example, for costume dramas, the historical background and life experience of characters can be identified according to the content, enriching the entire viewing core, guiding users to discover more content of interest, and expanding the user's viewing perspective. At the same time, the intelligent interactive interface module of the user interaction layer uses VR / AR technology and gesture and voice interaction functions to provide an immersive multi-channel interactive experience, and uses interactive data feedback for recommended content optimization, further enhancing the depth of interaction between users and the set-top box system, so that the set-top box is no longer just a playback device, but a personalized multimedia entertainment and information acquisition center for users, meeting the increasingly diversified and personalized needs of users, and improving user stickiness and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a system architecture diagram of the present invention;

[0033] Figure 2 It is a diagram of the system operation steps of the present invention;

[0034] Figure 3 It is a system data trend relationship diagram of the present invention;

[0035] Figure 4 A bar graph showing channel priority and decoding core allocation of the present invention;

[0036] Figure 5 It is a line graph of the data block checksum change of the present invention;

[0037] Figure 6 It is a scatter plot of the relationship between the signal characteristics and the interference type of the present invention;

[0038] Figure 7 A line graph showing the relationship between system load and core frequency adjustment of the present invention;

[0039] Figure 8 This is a bar chart of the comprehensive recommendation scores of the programs of the present invention for user 1. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Specific embodiment one:

[0042] like Figures 1 to 8 As shown, the intelligent set-top box system based on multi-channel synchronous processing, the system architecture includes a hardware layer, a software layer and a user interaction layer, wherein:

[0043] The hardware layer is centered on a distributed multi-core processor module, which includes core groups such as signal processing, decoding, and graphics processing, which are responsible for the preliminary processing of multi-channel signal reception, decoding, and image rendering layout optimization. The storage unit provides support for data temporary storage and system software and related data storage, and the communication interface module ensures the connection with external devices and networks.

[0044] The hardware layer includes a distributed multi-core processor module, a storage unit and a communication interface module. The distributed multi-core processor module is composed of multiple core groups, including a signal processing core group, a decoding core group and a graphics processing core group. The signal processing core group is responsible for the reception, filtering and preliminary processing of multi-channel signals. It is equipped with a high-speed signal receiver and an advanced filter unit. It can process signal inputs from multiple channels at the same time to ensure stable reception and preliminary purification of signals. The decoding core group contains multiple cores specifically used to decode video signals of different formats and encoding types, and performs efficient decoding operations for video encodings of different channels (such as H.264, H.265, etc.). The graphics processing core group focuses on channel screen rendering, special effects processing, and synthesis and layout optimization of multi-channel split-screen images. It has powerful graphics processing capabilities to provide high-quality visual output.

[0045] The storage unit includes a cache memory and a large-capacity storage medium. The cache memory is used to temporarily store the multi-channel data being processed to increase the data reading speed and reduce the core waiting time. The large-capacity storage medium is used to store system software, channel list information, user preference data, some pre-loaded channel content, and algorithm-related data (such as model parameters for multi-channel synchronous decoding algorithms and signal interference elimination algorithms).

[0046] The communication interface module has multiple communication interfaces, such as the network interface (wired Ethernet interface and Wi-Fi module) for connecting to the Internet, realizing data interaction with the content provider's server, obtaining channel program information, updating system software, and realizing social interaction functions, etc. The HDMI interface is used to connect to the display device and transmit high-definition video and audio signals. The USB interface can be used for external storage devices, cameras (for image acquisition of virtual reality or augmented reality interactive interfaces), voice input devices, etc., to expand the functions of the set-top box.

[0047] The entire system uses the distributed multi-core processor module of the above-mentioned hardware layer to divide the different links of multi-channel signal processing into special core groups. For example, the signal processing, decoding and graphics processing core groups work together to efficiently handle tasks such as signal reception, decoding and screen rendering layout optimization of multiple channels at the same time. The multi-channel synchronous decoding subsystem of the software layer is based on priority queues and dynamic load balancing strategies. It can reasonably allocate decoding tasks according to channel priority and core load to ensure smooth playback of high-priority channels, and flexibly adjust low-priority channels when the core load is high, effectively improving the efficiency and stability of the overall multi-channel synchronous processing, ensuring that users will not experience freezes or delays when switching and watching multiple channels, greatly improving the user viewing experience.

[0048] The software layer includes a multi-channel synchronous decoding subsystem, a multi-channel signal interference elimination subsystem, an intelligent resource management subsystem, and a multi-channel content association and intelligent recommendation subsystem. The multi-channel synchronous decoding subsystem uses an algorithm based on priority queues and dynamic load balancing strategies to allocate decoding tasks according to channel priorities and core load conditions, and has error detection and correction capabilities. The multi-channel signal interference elimination subsystem adopts a method based on decision trees and signal feature analysis to determine the type of interference and eliminate it in a targeted manner by building a decision tree. The intelligent resource management subsystem uses a threshold-based resource monitoring and allocation algorithm to dynamically adjust hardware resources according to system load. The multi-channel content association and intelligent recommendation subsystem uses an algorithm based on content tags and collaborative filtering to achieve cross-channel content association and personalized recommendations by building tags, analyzing user data and collaborative filtering.

[0049] The multi-channel synchronous decoding subsystem of the software layer sets a priority queue for each channel. The priority is determined based on the user's viewing history, channel subscription status, and current signal quality factors. For example, the channel being watched and the channels that the user often watches and pays for have higher priority. Then the system monitors the load of the decoding core group in real time and assigns the channel decoding tasks at the head of the queue to the decoding core with a lighter load. If the load of all cores is high, the decoding tasks of the low-priority channels are appropriately delayed or the frame rate is reduced according to the priority to ensure the smooth playback of the high-priority channels. At the same time, the subsystem also has the decoding error detection and correction function. When an error occurs in the decoding of a channel, it can try to switch to different decoding strategies or use redundant information to repair the error to ensure the stability of multi-channel decoding.

[0050] The multi-channel signal interference elimination subsystem of the software layer uses the construction of decision trees to obtain the relationship between different interference types and signal characteristics based on a large number of offline signal samples. When receiving multi-channel signals, the system extracts the time domain and frequency domain characteristics of the signal, such as signal amplitude changes, frequency distribution, etc., and inputs these characteristics into the decision tree to determine the type of interference. For adjacent frequency interference, according to the interference range determined by the decision tree, instructions are sent to the signal processing core group to dynamically adjust the filter parameters and accurately filter out the interference frequency components. For pulse interference, it works in coordination with the decoding core group and uses signal reconstruction technology to restore the signal of the interfered part according to the correlation between the previous and next signals. To ensure the quality of multi-channel signals, the multi-channel signal interference elimination subsystem adopts a method based on decision trees and signal feature analysis. It builds a decision tree based on a large number of offline signal samples to accurately determine the type of interference. Whether it is adjacent channel interference or pulse interference, targeted measures can be taken, such as dynamically adjusting filter parameters in the case of adjacent channel interference, and coordinating with the decoding core group to use signal reconstruction technology to restore the signal in the case of pulse interference, effectively removing interference, ensuring high-quality transmission and reception of multi-channel signals, making the picture clear and the sound smooth on each channel, reducing problems such as picture snow and noisy sound caused by signal interference, and providing users with stable and reliable multi-channel audio-visual services;

[0051] The intelligent resource management subsystem of the software layer pre-sets thresholds for different resources, such as core frequency and memory usage, and monitors in real time the resource requirements of each channel processing task in each core group of the hardware layer and the load of the entire system. When the system is in a low-load state, for example, the user only watches a few standard-definition channels and has no complex interactive operations, if the resource usage is lower than the set threshold, instructions are sent to the distributed multi-core processor module of the hardware layer to reduce the operating frequency and voltage of some unused or low-load cores, or even shut down some unnecessary hardware modules, such as idle decoding chip channels or graphics processing special effects units to save energy. When the system load suddenly increases, such as when the user switches to multiple high-definition channels or performs complex interactive operations, such as multi-screen interaction, virtual reality interaction, etc., if the resource usage exceeds the threshold, the dormant resources are quickly awakened and the resources are reallocated to various tasks. For example, more decoding cores are allocated to the newly switched high-definition channels. Resources and graphics processing resources to ensure that system performance is not affected. At the same time, the cache memory allocation strategy is reasonably adjusted to give priority to the data storage needs of the current active channels. The intelligent resource management subsystem monitors the system load and the resource requirements of each channel task in real time by setting different resource thresholds. In low-load conditions, such as when users watch a small number of standard-definition channels without complex operations, the frequency and voltage of unused or low-load cores can be reduced, and even unnecessary hardware modules can be turned off to achieve energy saving and consumption reduction; in high-load scenarios, such as switching to multiple high-definition channels or performing complex interactive operations, dormant resources can be quickly awakened and reasonably allocated to ensure that system performance is not affected. At the same time, the cache memory allocation strategy is optimized. This intelligent resource allocation mechanism not only improves the system resource utilization, but also reduces energy consumption, extends the service life of the set-top box, and reduces user usage costs. It is in line with modern energy-saving and environmental protection concepts and ensures the stable operation of the system under different working conditions.

[0052] The multi-channel content association and intelligent recommendation subsystem of the software layer establishes multi-dimensional content tags for each channel program, including program type, theme, actor, and director information, and then builds a user interest portrait by collecting and analyzing multi-dimensional data such as user viewing history, viewing time, channel stay time, and operating habits. At the same time, it interacts with content databases and social platforms on the Internet to obtain more program information and social interaction data. When a user is watching a program on a certain channel, the subsystem analyzes the program content of other channels in the background, matches the content tags with the user's interest portrait, and explores the potential associations between programs on different channels. For example, when a user is watching a science fiction movie on the movie channel, the subsystem searches for related science fiction content on news channels, science channels, documentary channels, etc., such as news reports. The system can analyze the special effects technology of movies in the TV channel, popular science programs on science fiction concepts in the science and technology channel, and documentaries behind the scenes of science fiction film production in the documentary channel, etc. Then, based on the collaborative filtering algorithm and combined with the viewing behavior of other similar users, these related contents are sorted to generate a personalized recommendation list, and the recommended content is displayed in an intuitive way on the user interface to guide users to discover more multi-channel content of interest. The multi-channel content association and intelligent recommendation subsystem establishes multi-dimensional content tags for the programs, combines multi-dimensional data such as user viewing history to build user interest portraits, and interacts with Internet content databases and social platforms to obtain more information. Through content tag matching and collaborative filtering algorithms, it can explore the potential connections between programs on different channels and generate personalized recommendation lists, guiding users to discover more content of interest and expanding their viewing horizons.

[0053] The user interaction layer includes the traditional interaction interface module and the intelligent interaction interface module. The traditional interaction interface module provides basic operating functions and works with some subsystems of the software layer. The intelligent interaction interface module uses VR and AR technology as well as gesture and voice interaction functions to provide users with an immersive multi-channel interaction experience, and uses interaction data feedback for recommendation content optimization.

[0054] The traditional interactive interface module of the user interaction layer provides basic channel list display, program information query, playback control, play, pause, fast forward, rewind, volume adjustment and other functions. Users can perform routine operations such as channel switching and program selection in the traditional interactive interface by pressing buttons on the remote control. This module works closely with the multi-channel synchronous decoding subsystem and intelligent resource management subsystem of the software layer. When the user switches channels, the relevant subsystems are notified in time to allocate resources and adjust decoding tasks;

[0055] The intelligent interactive interface module of the user interaction layer includes an immersive interactive interface based on virtual reality (VR) and augmented reality (AR) technology, as well as gesture and voice interaction functions. In the immersive interactive interface, users view real-time images of multiple channels in a three-dimensional and dynamic form through head-mounted display devices (VR) or mobile devices such as mobile phones (AR). Through gesture recognition devices or mobile phone sensors, users can switch between different channels and adjust the size and layout of channel images. The voice interaction function allows users to search channels, play programs, adjust volume, and other operations through voice commands. For example, users can say commands such as "switch to the sports channel" and "play the next recommended program". The intelligent interactive interface module of the user interaction layer includes an immersive interactive interface based on virtual reality (VR) and augmented reality (AR) technology, as well as gesture and voice interaction functions. The interactive interface module parses the voice commands and passes them to the corresponding subsystem of the software layer for execution. The module can also feed back the user's behavioral data during the intelligent interaction process to the multi-channel content association and intelligent recommendation subsystem to further optimize the recommended content. The intelligent interactive interface module of the user interaction layer uses VR / AR technology and gesture and voice interaction functions to provide an immersive multi-channel interactive experience, and uses the interaction data feedback to optimize the recommended content, further enhancing the depth of interaction between the user and the set-top box system, making the set-top box no longer just a playback device, but a personalized multimedia entertainment and information acquisition center for users, meeting the increasingly diversified and personalized needs of users and improving user stickiness and satisfaction. Specific embodiment 2:

[0057] like Figures 1 to 8 As shown, according to the contents in the above specific embodiments, the following contents are further disclosed:

[0058] The algorithm content in the entire system architecture is as follows:

[0059] The algorithm based on priority queue and dynamic load balancing strategy in the multi-channel synchronous decoding subsystem is as follows:

[0060] Let the channel set be ,

[0061] Each channel The corresponding priority is , The value range is set to 0 to 1. The larger the value, the higher the priority. Its value is determined based on factors such as user viewing history, channel subscription status, and current signal quality.

[0062] The decoding core group has cores, denoted as ,

[0063] Per core The current load condition is Indicates that the load is measured by indicators such as the resource occupancy rate of the tasks assigned to the core. The value range is from 0 to 1, 0 means no load and 1 means full load:

[0064] The decision function for assigning decoding tasks can be simply expressed as:

[0065]

[0066] Among them: In the formula Indicates channel Whether the decoding task is assigned to the core , 1 means allocation, 0 means no allocation, This part means that the channel is considered comprehensively Priorities and core The remaining load capacity, Indicates the remaining load space. By comparing the product values ​​corresponding to all cores, the channel decoding task is assigned to the core that makes the value the largest, that is, the task is preferentially assigned to the core with light load and high channel priority:

[0067] Implementation steps:

[0068] Sp1: First, for each channel ,determine their priority according to established rules, such as viewing history;

[0069] Sp2: Real-time monitoring of each decoding core Load conditions :

[0070] Sp3: For each channel The decoding task is calculated with each core of value:

[0071] Sp4: Find the The largest core and the channel The decoding task is assigned to the core. If all cores are highly loaded, for low-priority channels, the task allocation strategy is further adjusted according to the priority, such as appropriately delaying tasks or reducing frame rates. This can be judged by setting additional threshold conditions. For example, when all When the load is greater than a certain high load threshold, corresponding actions are taken.

[0072] Operation logic: The system continuously executes the above steps in a loop. As the channel conditions (such as changes in signal quality, changes in priority caused by user operations, etc.) and core load conditions change dynamically, the decoding tasks are continuously reallocated to ensure that high-priority channels have priority in obtaining sufficient decoding resources to achieve smooth playback.

[0073] Beneficial effects: By reasonably allocating decoding tasks, the resources of each decoding core are fully utilized, avoiding the situation where some cores are overloaded while some cores are idle, effectively improving the efficiency of overall multi-channel synchronous decoding, and ensuring that channels of different priorities can be played as smoothly as possible under different system load conditions, reducing the occurrence of freezes and delays when users watch multiple channels, and improving the viewing experience.

[0074] The decoding error detection and correction function related algorithms further include the following:

[0075] Set the received video stream data ,Will Divide into several data blocks ,

[0076] For each data block Calculate Checksum There are many ways to calculate the checksum, such as the common cyclic redundancy check CRC;

[0077] set up ,in, Represents a data block The Bytes of data, It is a data block The byte length of is a chosen prime number:

[0078] When a channel decoding error is detected, there is a redundant data block , by comparing the error data block (data blocks where checksum mismatch was detected) and redundant data blocks Correlation of , calculate the similarity function

[0079]

[0080] in, and They are the byte data at the corresponding positions of the error data block and the redundant data block, is the length of the data block, find the redundant data block with the highest similarity To repair the error block ;

[0081] First, checksum The calculation is used to preliminarily determine whether the data block has errors. The checksum calculation uses the cumulative modulus operation of the byte data in the data block. If the checksum recalculated by the receiving end is inconsistent with the checksum attached by the sending end, it is determined that the data block has errors, and the similarity function It measures the similarity between the erroneous data block and the redundant data block in content, and corrects the decoding error by finding the most similar redundant data block to replace or repair the erroneous data block;

[0082] Operation logic: During the entire decoding process, data blocks are checked and detected in real time. Once an error is found, the error repair process is started, and redundant information is continuously used to try to restore the correct data content to ensure the accuracy and stability of decoding;

[0083] Beneficial effects: It can timely discover and repair errors that occur during the decoding process, reduce problems such as screen distortion and abnormal sound caused by interference and data loss during signal transmission, improve the reliability of multi-channel decoding, and further ensure the stability of the user's viewing experience.

[0084] The method based on decision tree and signal feature analysis in the multi-channel signal interference elimination subsystem is as follows:

[0085] Assume that the time domain and frequency domain feature vectors of multi-channel signals are ,

[0086] in, It can be specific characteristic values ​​such as signal amplitude change, frequency distribution, etc.

[0087] Decision Tree Model It is obtained through training with a large number of offline signal samples. Its internal nodes are divided based on different features, and leaf nodes correspond to different interference types. ;

[0088] The classification function of the decision tree can be expressed as , that is, the input signal feature vector , through the decision tree The interference type corresponding to the judgment output ;

[0089] Among them, here Represents the extracted signal feature set, decision tree It is built based on the analysis of a large number of signal samples under different interference conditions. It has a series of judgment branches based on conditions such as feature thresholds. After the input is entered, it will be judged along the branches of the decision tree and finally reach a leaf node. The corresponding leaf node represents the type of interference It is the interference type of the current signal determined;

[0090] Implementation steps:

[0091] Sp1: In the offline stage, a large number of multi-channel signal samples with different interference types are collected, and these samples are analyzed in time domain, frequency domain and other features to extract feature values ​​such as signal amplitude change and frequency distribution, and construct feature vectors;

[0092] Sp2: Use these sample data with feature vectors and interference type annotations to train a decision tree model , by selecting appropriate features as the basis for dividing the internal nodes of the decision tree (such as selecting the features with the largest information gain for division, and other common decision tree construction methods), branches and leaf nodes are continuously constructed, so that the final decision tree can accurately judge the interference type based on the input feature vector;

[0093] Sp3: When receiving multi-channel signals in real time, extract the time domain and frequency domain features of the signals to form feature vectors ;

[0094] Sp4: The feature vector Input to the trained decision tree In the above equation, we can get the corresponding interference type. If the result is adjacent frequency interference, send instructions to the signal processing core group according to the interference range determined by the decision tree to dynamically adjust the filter parameters and set the filter parameter adjustment function , according to the interference type Determine the specific adjustment amount:

[0095] For adjacent channel interference , adjust the filter's cutoff frequency and other parameters to make Accurately filter out interfering frequency components;

[0096] For pulse interference, it works with the decoding core group and uses signal reconstruction technology (assuming the signal reconstruction function is , according to the signal before and after interference and The signal of the interfered part is restored by the correlation, for example, by interpolation, prediction and other algorithms).

[0097] Beneficial effects: Regardless of common interference situations such as adjacent channel interference or pulse interference, it can accurately judge and take effective measures to eliminate interference, ensure high-quality transmission and reception of multi-channel signals, make the picture clear and the sound smooth on each channel, reduce problems such as picture snow and noisy sound caused by signal interference, and provide users with stable and reliable multi-channel audio-visual services.

[0098] The threshold-based resource monitoring and allocation algorithm of the intelligent resource management subsystem is as follows:

[0099] Assume that there are multiple types of system resources, based on core frequency and memory usage For example, the threshold range of the core frequency is preset as , the threshold range of memory usage is ; For core frequency adjustment, set the current system load to Indicates (the load can be comprehensively evaluated based on the resource usage of the running channel processing tasks, with a value range of 0 to 1), and the core frequency adjustment function is

[0100]

[0101] in, and It is the low and high threshold limit of the load, which is used to divide different frequency adjustment intervals:

[0102] For memory resource allocation, let the channel task set be , each channel task The memory requirement is , the total system memory is , memory allocation function:

[0103]

[0104] Allocate memory according to the ratio of each channel task memory requirement to the total requirement, and set the cache memory to be allocated to the channel. The ratio is , determined by factors such as channel priority, the higher the priority, the larger the total cache memory capacity is , cache allocation function :

[0105] Core frequency adjustment function Depending on system load Dynamically adjust the core frequency according to different situations:

[0106] When the load is high (greater than ) to increase the core frequency to the maximum value To ensure system performance;

[0107] When the load is in the middle range arrive ) when the frequency is adjusted linearly in proportion;

[0108] When the load is low (less than ) when reducing the frequency to the minimum value To save energy:

[0109] Memory allocation functions It is based on the proportion of memory demand of each channel task to the total demand to fairly distribute the total system memory. , ensuring that each task has reasonable memory resources, cache allocation function The ratio is determined by factors such as channel priority. To allocate the total cache memory capacity , giving priority to ensuring the data storage needs of important channels;

[0110] Operation logic: The system continuously monitors resource demand and load conditions, and dynamically adjusts the allocation of various hardware resources based on set thresholds and corresponding allocation functions to adapt to system performance assurance and energy-saving requirements under different user operation scenarios (such as changes in the number of viewing channels, channel types, etc.). It is a real-time feedback and adjustment cycle.

[0111] Beneficial effects: Under low-load conditions, such as when users watch a small number of standard-definition channels without complex operations, the frequency and voltage of unused or low-load cores can be reduced, and even unnecessary hardware modules can be shut down to achieve energy saving and consumption reduction; in high-load scenarios, such as switching to multiple high-definition channels or performing complex interactive operations, dormant resources can be quickly awakened and reasonably allocated to ensure that system performance is not affected, while optimizing the cache memory allocation strategy. This intelligent resource allocation mechanism not only improves system resource utilization, but also reduces energy consumption, extends the service life of the set-top box, and reduces user costs. It complies with modern energy-saving and environmental protection concepts and ensures the stable operation of the system under different working conditions.

[0112] The specific contents of the algorithm based on content labeling and collaborative filtering of the multi-channel content association and intelligent recommendation subsystem are as follows:

[0113] Assume the program set is, ;

[0114] For each program Building a multi-dimensional content label vector ;

[0115] in, The label value represents a specific dimension such as program type, theme, actor, director, etc. It can be a discrete category value or a numerical value, such as Indicates the type of program, with values ​​such as "movie" or "TV series". Indicates the theme, with values ​​such as "science fiction" and "romance";

[0116] Let the user set be ;

[0117] For Users , by collecting and analyzing multi-dimensional data such as viewing history, viewing time, channel stay time, operation habits, etc., to build a user interest portrait vector ,For example Content label vector that can represent the degree of preference for different types of programs ,in The label value represents a specific dimension such as program type, theme, actor, director, etc. It can be a discrete category value or a numerical value, such as Indicates the type of program, with values ​​such as "movie" or "TV series". Indicates the theme, with values ​​such as "science fiction" and "romance";

[0118] Let the user set be , for users , by collecting and analyzing multi-dimensional data such as viewing history, viewing time, channel stay time, operation habits, etc., to build a user interest portrait vector ,like It can indicate the degree of preference for different types of programs, with a value range of 0 to 1, etc. The higher the value, the more you like the type of program;

[0119] The function for calculating the matching degree between a program and a user's interest can be expressed as:

[0120]

[0121] It is the weight value set between different dimension labels and interest portrait dimensions, which is used to reflect the importance of each dimension in the matching calculation:

[0122] For the collaborative filtering part, let Similar user sets are (Similar users can be determined by calculating the similarity of interest portrait vectors between users, such as using cosine similarity and setting a similarity threshold to filter out similar users). About the show The rating of (can be converted into a rating value based on factors such as viewing time and whether it is a favorite, etc.) ), calculate the program The recommendation score function based on collaborative filtering is:

[0123]

[0124] in, It is the weight value set for different similar users, which can be determined based on factors such as the similarity between users. The higher the similarity, the greater the weight:

[0125] For Users The comprehensive recommendation score function can be expressed as:

[0126]

[0127] Among them, is a weight coefficient ranging from 0 to 1, which is used to balance the proportion of content tag matching and collaborative filtering score in comprehensive recommendation;

[0128] First, the interest matching function The cosine similarity between the program content label vector and the user interest portrait vector is calculated to measure the degree of fit between the program content features and the user interest. The importance of each dimension can be adjusted according to the actual business situation, and the collaborative filtering recommendation score function The recommendation score function is to use the ratings of similar users on the program and comprehensively consider the weights of different similar users to evaluate the recommendation value of the program based on the behavior of other similar users. Combining content tag matching and collaborative filtering, the weight coefficient is used to reasonably allocate the proportion of the two, so as to obtain a more comprehensive and accurate program recommendation score for a specific user.

[0129] Operation logic: The system will regularly update program content tags (such as when new programs are launched or program information changes, etc.) and user interest portraits (as users' viewing behaviors continue to accumulate and change). Whenever a user enters a recommendation-related interface or there is a suitable recommendation trigger scenario, the above recommendation algorithm process will be run in real time to dynamically generate a personalized recommendation list that meets the user's current interests. The recommendation results can be further optimized based on the user's subsequent feedback on the recommended content (such as clicking to watch, ignoring, etc.), forming a recommendation cycle of continuous learning and improvement.

[0130] Beneficial effects: By combining content tag matching with collaborative filtering algorithms, the potential connections between programs on different channels can be fully explored. It is not limited to recommendations within the channels that users have already watched, but it can discover more content that meets the user's interests across channels, generate personalized recommendation lists and display them to users, guide users to discover more content of interest, expand users' viewing horizons, and increase users' opportunities to discover high-quality programs, thereby improving users' satisfaction and stickiness with the set-top box system, making users more willing to use the system for a long time to obtain multimedia entertainment and information services.

[0131] like Figure 4The channel priority and decoding core allocation bar chart shown in the figure shows the priority values ​​corresponding to each channel. The blue bar chart shows the priority values ​​corresponding to each channel, which highly intuitively reflects the priority of each channel. The value range is between 0 and 1. Different channels have different priorities, representing different levels of importance. They may be formed based on factors such as user viewing history, channel subscription status, and current signal quality. The red scattered points are marked above the corresponding channel number, and the vertical axis value represents the decoding core number assigned to the channel. Since 1 is added, the scattered points will be at the integer position corresponding to the core number for easy observation.

[0132] By observing the graph, we can clearly see the differences in the priorities of different channels and the decoding cores to which they are assigned. For example, if a channel has a high priority (corresponding to a higher blue column), and the load of the core number corresponding to the red scattered point above it is relatively low (in more complex scenarios, it can be further analyzed in combination with the core load data), it means that the system tends to assign the decoding tasks of high-priority channels to lightly loaded cores according to the algorithm of the priority queue and dynamic load balancing strategy to ensure smooth playback. Overall, it can intuitively present the execution effect of the algorithm in terms of channel and decoding core resource allocation, helping to understand whether the resource allocation is reasonable and whether it meets the expected goal of ensuring priority processing of high-priority channels;

[0133] like Figure 5 As shown in the figure, the data block checksum change line graph shows the change of the checksum value corresponding to the original data block. The horizontal axis is the data block number, and the vertical axis is the checksum value. Different data blocks have their own corresponding checksums, which together form a continuously changing line, reflecting the distribution of the data block checksums under normal circumstances. The red line represents the change of the checksum values ​​obtained by recalculating all data blocks after simulating an error in a data block. At the location of the erroneous data block (corresponding to a point on the horizontal axis), its checksum value is significantly different from the blue line (original checksum), causing the two lines to separate here.

[0134] The main function of this figure is to show the effect of decoding error detection. Under normal circumstances, the blue line smoothly reflects the original checksum of each data block. When a data block has an error, the red line changes at the corresponding position. This change is easily observed by comparing the two lines. From this, we can intuitively determine which data block has a checksum mismatch problem, that is, a decoding error has been detected. Based on this, we can then take corresponding error correction measures, such as using redundant information to repair erroneous data blocks. This overall illustrates the key role of error detection in the entire decoding process and the intuitive form of detection.

[0135] like Figure 6As shown in the figure, a scatter plot of the relationship between signal features and interference types is shown. Each scatter point in the figure represents a signal sample, the horizontal axis is the first eigenvalue of the signal (for example, one of the specific features such as signal amplitude change), and the vertical axis is the second eigenvalue (also a certain characteristic dimension of the signal). The color of the scatter point represents the corresponding interference type judgment result. Two colors are used here to distinguish different interference types (assuming that 1 represents one interference type, corresponding to one color; 0 represents another interference type, corresponding to another color).

[0136] This graph can intuitively show the relationship between signal features and interference types. By observing the distribution of scattered points and color distinction, we can see the approximate distribution pattern of signals of different interference types in the feature space. For example, if most of the scattered points in a certain area are of the same color (representing the same interference type), it means that within the range of feature values, the signal is more likely to have the corresponding interference type. This helps to understand how decision trees judge interference types based on these features in methods based on decision trees and signal feature analysis, that is, by analyzing the area where the signal features fall (from the graph, the feature value range corresponding to the scattered point distribution area) to determine the interference type, and then take targeted interference elimination measures, such as adjusting filter parameters for different types of interference or using different signal reconstruction techniques, etc., to provide intuitive feature basis for subsequent accurate interference elimination.

[0137] like Figure 7 As shown in the figure, there is a line graph of the relationship between system load and core frequency adjustment. The blue line represents the system load at different time points, and its value fluctuates between 0 and 1, reflecting the change in the degree of occupation of the overall system resources over time. For example, changes in user operation behavior (such as switching channels, performing interactive operations, etc.) will cause the load to increase or decrease accordingly. The red line shows the core frequency adjustment value calculated by the core frequency adjustment algorithm based on the system load. The vertical axis corresponds to the core frequency value (varying between the set minimum and maximum values), and the horizontal axis is also the time point, corresponding to the time series of the system load.

[0138] By observing this graph, we can clearly see the dynamic relationship between system load and core frequency adjustment. When the blue line (system load) rises above the set high threshold (such as 0.7 in the example), the red line (core frequency) will be adjusted to a higher value (such as the maximum value 5 in the example), which means that the system will increase the core frequency to ensure performance and ensure that each channel task can be processed normally in response to high load conditions; conversely, when the system load drops below the low threshold (such as 0.3 in the example), the core frequency will drop to a lower value (such as the minimum value 1 in the example) to achieve energy saving. Overall, it intuitively presents how the threshold-based resource monitoring and allocation algorithm of the intelligent resource management subsystem works dynamically according to the system load in terms of core frequency adjustment, and verifies the algorithm's effect of reasonably allocating resources under different working conditions to balance performance and energy saving.

[0139] like Figure 8 As shown in the figure, a bar chart of the comprehensive recommendation scores of programs for user 1 is shown. The horizontal axis is the program number, covering all simulated programs, and the vertical axis is the comprehensive recommendation score of the program for user 1. The height of the column corresponds to the specific score value. Different programs have different recommendation scores, which reflects the degree of recommendation for user 1.

[0140] From this bar chart, we can intuitively know the recommendation priority of each program relative to User 1. The higher the score of the program (the higher the corresponding bar), the more it is in line with User 1's interest preferences and is more worthy of being recommended to User 1 for viewing, based on the algorithm based on content tags and collaborative filtering, taking into account the matching degree between the program's own content tags and User 1's interest profile and the ratings of similar users on the program. This helps us understand how the recommendation algorithm integrates multiple factors to generate personalized recommendation results, and by observing the differences in scores of different programs, we can analyze which programs are more in line with the preferences of target users in terms of content features or public evaluation (reflected by similar user ratings), etc., providing an intuitive reference for further optimizing the recommendation algorithm and improving the accuracy of recommendations.

[0141] These result diagrams show from different angles the execution status of the key links of each subsystem algorithm, the input-output relationship, and the expected effects of resource allocation, error detection, interference elimination, recommendation generation, etc., helping to better understand the working principles and actual functions of each algorithm in the entire smart set-top box system. Specific embodiment three:

[0143] like Figures 1 to 8 As shown, according to the contents in the above specific embodiments, the following contents are further disclosed:

[0144] The operation steps of the whole system are as follows:

[0145] Sp1: Multi-channel signals are received, filtered and preliminarily processed by the signal processing core group in the distributed multi-core processor module of the hardware layer using the equipped high-speed signal receiver and advanced filter unit to ensure stable reception and preliminary purification of the signal;

[0146] Sp2: The signal after preliminary processing enters the decoding core group. The decoding core group contains multiple cores dedicated to decoding video signals of different formats and encoding types. For video encoding of different channels (such as H.264, H.265, etc.), efficient decoding operations are performed according to the priority queue and dynamic load balancing strategy set by the multi-channel synchronous decoding subsystem of the software layer. The strategy first sets a priority queue for each channel (the priority is determined based on the user's viewing history, channel subscription status, and current signal quality factors), and then monitors the load of the decoding core group in real time, and assigns the channel decoding tasks at the head of the queue to the decoding core with a lighter load. If the load of all cores is high, the decoding tasks of low-priority channels are appropriately delayed or the frame rate is reduced according to the priority. At the same time, the subsystem has the function of decoding error detection and correction. When an error occurs, it tries to switch to different decoding strategies or use redundant information to repair the error;

[0147] Sp3: The decoded signal is transmitted to the graphics processing core group, which focuses on channel image rendering, special effect processing, and synthesis and layout optimization of multi-channel split-screen images. It has powerful graphics processing capabilities to provide high-quality visual output;

[0148] Sp4: In the signal processing process, the multi-channel signal interference elimination subsystem of the software layer adopts a method based on decision tree and signal feature analysis. It uses a decision tree (built based on the relationship between different interference types and signal features obtained by analyzing a large number of offline signal samples). When receiving a multi-channel signal, it extracts the time domain and frequency domain features of the signal (such as signal amplitude change, frequency distribution, etc.), and inputs these features into the decision tree to determine the interference type. For adjacent frequency interference, according to the interference range determined by the decision tree, it sends instructions to the signal processing core group, dynamically adjusts the filter parameters, and accurately filters out the interference frequency components. For pulse interference, it works in coordination with the decoding core group, using signal reconstruction technology to restore the signal of the interfered part according to the correlation between the previous and next signals, to ensure the quality of the multi-channel signal.

[0149] Sp5: The intelligent resource management subsystem of the software layer pre-sets thresholds for different resources (such as core frequency, memory usage, etc.), monitors the resource requirements of each channel processing task in each core group of the hardware layer and the load of the overall system in real time, and sends instructions to the distributed multi-core processor module of the hardware layer in a low-load state (such as when the user only watches a few standard-definition channels and there are no complex interactive operations, if the resource usage is lower than the set threshold), reduces the operating frequency and voltage of some unused or low-load cores, and even shuts down some unnecessary hardware modules to save energy. When the system load suddenly increases (such as when the user switches to multiple high-definition channels or performs complex interactive operations), if the resource usage exceeds the threshold, it quickly wakes up the dormant resources, reallocates the resources to each task, and reasonably adjusts the cache memory allocation strategy to prioritize the data storage needs of the current active channels;

[0150] Sp6: The multi-channel content association and intelligent recommendation subsystem of the software layer first establishes multi-dimensional content tags for each channel program, and then builds a user interest profile by collecting and analyzing multi-dimensional data such as user viewing history, viewing time, channel stay time, and operating habits. At the same time, it interacts with content databases and social platforms on the Internet to obtain more program information and social interaction data. When a user is watching a program on a certain channel, the subsystem analyzes the program content of other channels in the background, matches the user interest profile according to the content tags, and explores the potential associations between programs on different channels. Then, based on the collaborative filtering algorithm and the viewing behavior of other similar users, it sorts these related contents, generates a personalized recommendation list, and displays the recommended contents in an intuitive way on the user interface.

[0151] Sp7: Users perform basic operations through the traditional interactive interface module of the user interaction layer, such as displaying the channel list, querying program information, playing control (play, pause, fast forward, rewind, volume adjustment, etc.), channel switching, selecting programs and other routine operations through the key operations on the remote control. This module works closely with the multi-channel synchronous decoding subsystem and intelligent resource management subsystem of the software layer, and notifies the relevant subsystems in time to allocate resources and adjust decoding tasks during operation;

[0152] Sp8: Users can also use the intelligent interactive interface module of the user interaction layer, with the help of an immersive interactive interface based on virtual reality (VR) and augmented reality (AR) technology, as well as gesture and voice interaction functions, to view real-time images of multiple channels in a three-dimensional and dynamic form through head-mounted display devices (VR) or mobile devices such as mobile phones (AR), switch between different channels, adjust the size and layout of channel images through gesture recognition devices or mobile phone sensors, or search channels, play programs, adjust the volume, etc. through voice commands (such as "switch to the sports channel" and "play the next recommended program". The intelligent interactive interface module parses the voice commands and passes them to the corresponding subsystem of the software layer for execution. The module will also feed back the user's behavior data during the intelligent interaction process to the multi-channel content association and intelligent recommendation subsystem to further optimize the recommended content. Specific embodiment four:

[0154] like Figures 1 to 8 As shown, according to the contents in the above specific embodiments, the following contents are further disclosed:

[0155] The data transmission relationship between the various levels of the entire system is as follows:

[0156] Hardware Layer

[0157] Distributed multi-core processor modules

[0158] Signal Processing Core Group:

[0159] Data source: multi-channel signal input from outside;

[0160] Input type: original analog or digital signals of different channels;

[0161] Processing method: Use the equipped high-speed signal receiver to receive the signal, filter it through the advanced filter unit, and perform reception, filtering and preliminary processing operations on multi-channel signals;

[0162] Output type: relatively pure and stable multi-channel signal after preliminary processing;

[0163] Output flow: Output to the decoding core group, and provide the original signal data required for signal feature analysis to the multi-channel signal interference elimination subsystem (for extracting time domain and frequency domain features, etc.).

[0164] Decoding core groups:

[0165] Data source: Multi-channel signals after preliminary processing by the signal processing core group;

[0166] Input type: video signal after preliminary processing (including different formats and encoding types, such as H.264, H.265, etc.);

[0167] Processing method: According to the priority queue and dynamic load balancing strategy set by the multi-channel synchronous decoding subsystem of the software layer, the channel decoding task at the head of the queue is allocated to the decoding core with lighter load for decoding. If the core load is high, the low-priority channel is adjusted and processed according to the rules. It also has the decoding error detection and correction function. If an error occurs, the strategy is switched or repaired by using redundant information.

[0168] Output type: decoded video signal data;

[0169] Output direction: Output to the graphics processing core group.

[0170] Graphics processing core group:

[0171] Data source: Video signal data decoded by the decoding core group;

[0172] Input type: decoded video signal data of each channel;

[0173] Processing method: Focus on channel screen rendering, special effects processing, and synthesis and layout optimization of multi-channel split-screen images, using its own powerful graphics processing capabilities for processing;

[0174] Output type: processed multi-channel image data (video data) with high-quality visual effects;

[0175] Output flow: Output to the HDMI interface in the communication interface module, and then transmitted to the display device for display through the interface; at the same time, the relevant picture data will also be obtained by the multi-channel content association and intelligent recommendation subsystem of the software layer to analyze the channel content characteristics, etc. (for example, visual related features such as program types in the picture assist in establishing content tags).

[0176] Storage Unit

[0177] Cache memory:

[0178] Data sources: temporary data during the processing of each core group in the hardware layer, such as multi-channel data being processed; temporary data related to some subsystems in the software layer (such as some cached data in the resource allocation process of the intelligent resource management subsystem, etc.);

[0179] Input type: temporary intermediate data generated by each link of multi-channel processing (data fragments in different formats, etc.);

[0180] Processing method: As a temporary storage area, temporarily store these data to increase data reading speed and reduce core waiting time;

[0181] Output type: Output corresponding temporary data according to the requirements of each core group and software layer subsystem.

[0182] Output flow: Output data to the corresponding core group (such as signal processing core group, decoding core group, etc.) or software layer related subsystem (such as intelligent resource management subsystem, etc.) as required to continue to participate in calculation and processing.

[0183] Mass storage media:

[0184] Data sources: system software, channel list information, user preference data imported during system initialization; some preloaded channel content; algorithm-related data (such as model parameters for multi-channel synchronous decoding algorithm and signal interference elimination algorithm, etc.);

[0185] Input types: system installation packages, text-formatted channel lists, configuration files for user settings preferences, preloaded channel content in video format, algorithm parameter files, and other data files;

[0186] Processing method: Store these data for a long time and make them available for the corresponding modules and subsystems to call during the system operation;

[0187] Output type: Output corresponding system software, channel information, user preferences and other stored data according to the requests of different modules and subsystems;

[0188] Output flow: Output system software data to the processor module of the hardware layer for system startup and operation support; output channel list information to the multi-channel content association and intelligent recommendation subsystem of the software layer to assist in recommendation analysis; output user preference data to the multi-channel content association and intelligent recommendation subsystem to build user interest portraits; algorithm-related data is output to the corresponding software layer algorithm processing subsystem (such as multi-channel synchronous decoding subsystem, multi-channel signal interference elimination subsystem, etc.) to assist in decoding, interference elimination and other algorithm operations.

[0189] Communication interface module

[0190] Network interface (wired Ethernet interface and Wi-Fi module):

[0191] Data source: The server of the content provider on the Internet, which obtains relevant data through network connection;

[0192] Input type: network data packets, including different types of data such as channel program information, system software update packages, and social interaction related data;

[0193] Processing method: Receive and parse these network data packets according to the network communication protocol;

[0194] Output type: parsed channel program information, updated system software files, social interaction related structured data, etc.

[0195] Output flow: Channel program information is output to the multi-channel content association and intelligent recommendation subsystem of the software layer to enrich the recommended content; system software update files are output to the large-capacity storage medium of the storage unit to update the system software; social interaction related data is output to the multi-channel content association and intelligent recommendation subsystem to assist in building user interest portraits and optimizing recommended content.

[0196] HDMI interface:

[0197] Data source: processed multi-channel image data (video data) with high-quality visual effects and corresponding audio signal data (if there is audio synchronization processing, it will be transmitted here together) output by the hardware layer graphics processing core group;

[0198] Input type: mixed data stream of video and audio signals (compliant with HDMI transmission standard format);

[0199] Processing method: According to the HDMI interface transmission protocol specification, the input data is transmitted to the external display device (such as a TV, monitor, etc.);

[0200] Output type: video and audio data stream output via HDMI cable;

[0201] Output flow: Transmitted to an externally connected display device for playback and display, allowing users to see multi-channel images and hear corresponding sounds.

[0202] USB interface:

[0203] Data sources: data transmitted by external devices such as external storage devices (such as mobile hard disks, USB flash drives, etc.), cameras (used for image acquisition in virtual reality or augmented reality interactive interfaces), and voice input devices.

[0204] Input type: file data transmitted by storage devices (such as multimedia files such as video, audio, and pictures, or configuration files, etc.); image data collected by cameras; voice command data obtained by voice input devices, and other different types of data;

[0205] Processing method: Identify and receive these different types of data, and process or forward them accordingly according to the functional requirements of the set-top box system;

[0206] Output type: Data from storage devices can be output to storage units for storage (such as importing video files and storing them in large-capacity storage media); image data collected by the camera is output to the intelligent interactive interface module of the user interaction layer for VR / AR related interactive operations (such as gesture recognition and other image-based interactive scenarios); voice command data from voice input devices is output to the intelligent interactive interface module for subsequent processing such as voice command analysis;

[0207] Output flow: According to the data type, it flows to the storage unit, intelligent interactive interface module and other corresponding modules for the next step of processing.

[0208] Software Layer

[0209] Multi-channel synchronous decoding subsystem

[0210] Data sources: multi-channel signal data after preliminary processing by the signal processing core group at the hardware layer; core load data fed back by the intelligent resource management subsystem at the software layer; user-related viewing data (collected, integrated and provided by the multi-channel content association and intelligent recommendation subsystem to determine channel priority);

[0211] Input type: multi-channel raw signals, core load value information, structured data related to user viewing (such as viewing history, channel subscription status, etc.);

[0212] Processing method: Use an algorithm based on priority queues and dynamic load balancing strategies to allocate decoding tasks according to channel priority and core load conditions. It also has the ability to detect and correct errors and perform corresponding repairs on channels with decoding errors.

[0213] Output types: decoding task allocation plan data (informs the decoding core group how to allocate tasks); decoding error report and repair status data (feedback to other related subsystems for monitoring system operation, etc.);

[0214] Output flow: The decoding task allocation plan is output to the decoding core group of the hardware layer to guide its decoding work; the decoding error report and repair status data are output to the intelligent resource management subsystem (to facilitate the allocation of resources according to the decoding situation) and the multi-channel content association and intelligent recommendation subsystem (to assist in analyzing the impact of channel playback stability on recommendations, etc.).

[0215] Multi-channel signal interference elimination subsystem

[0216] Data sources: multi-channel signal data transmitted by the hardware layer signal processing core group (for extracting features and analyzing interference); self-built model parameters based on decision tree and signal feature analysis (obtained from the algorithm-related data loaded from the large-capacity storage medium of the storage unit);

[0217] Input type: multi-channel original signals, model parameter data related to decision trees;

[0218] Processing method: Adopt a method based on decision tree and signal feature analysis, build a decision tree to determine the interference type, and take corresponding measures (such as dynamically adjusting filter parameters, using signal reconstruction technology in the collaborative decoding core group, etc.) to eliminate interference for different interference types such as adjacent channel interference and pulse interference;

[0219] Output type: Interference type judgment result data (feedback to the relevant core groups of the hardware layer to guide operations); signal quality evaluation data after interference elimination (provided to the intelligent resource management subsystem for overall system performance evaluation, etc.);

[0220] Output flow: The interference type judgment result is output to the signal processing core group of the hardware layer (to guide it to adjust the filter parameters for adjacent channel interference) and the decoding core group (to work together for pulse interference); the signal quality evaluation data after interference elimination is output to the intelligent resource management subsystem for reference in resource allocation decisions, and is also output to the multi-channel content association and intelligent recommendation subsystem to assist in judging the impact of channel picture quality on recommendations, etc.

[0221] Intelligent resource management subsystem

[0222] Data sources: resource usage data (such as core frequency, memory usage, etc.) fed back by each core group at the hardware layer; decoding task allocation data of the multi-channel synchronous decoding subsystem at the software layer; signal quality assessment data after interference elimination of the multi-channel signal interference elimination subsystem at the software layer; preset configuration data of different resource thresholds (loaded from relevant configuration files obtained from the large-capacity storage medium of the storage unit);

[0223] Input type: hardware resource usage information, decoding task allocation related data, signal quality assessment values, structured data for resource threshold settings;

[0224] Processing method: Threshold-based resource monitoring and allocation algorithm dynamically adjusts hardware resources according to system load, such as reducing the operating frequency and voltage of some unused or low-load cores and shutting down unnecessary hardware modules when the load is low, waking up dormant resources when the load is high, reallocating resources to various tasks, and adjusting the allocation strategy of cache memory;

[0225] Output type: resource allocation instruction data (sent to the distributed multi-core processor module of the hardware layer to guide its hardware resource adjustment); system load and resource allocation report data (output to the multi-channel synchronous decoding subsystem for adjusting the decoding strategy, output to the multi-channel signal interference elimination subsystem to assist it in further optimizing the interference elimination work, output to the multi-channel content association and intelligent recommendation subsystem for comprehensive analysis of the impact of system status on recommendation, etc.);

[0226] Output flow: Resource allocation instructions are output to the distributed multi-core processor module of the hardware layer; system load and resource allocation status report data are output to other related subsystems of the software layer (such as multi-channel synchronous decoding subsystem, multi-channel signal interference elimination subsystem, multi-channel content association and intelligent recommendation subsystem, etc.).

[0227] Multi-channel content association and intelligent recommendation subsystem

[0228] Data sources: Channel screen-related data output by the graphics processing core group of the hardware layer (to assist in analyzing program content features for establishing content tags, etc.); multi-dimensional data such as user viewing history, viewing time, channel stay time, and operating habits (collected and organized through the user interaction layer); program information and social interaction data obtained through the interaction of content databases and social platforms on the Internet (obtained through the network interface of the communication interface module); system operation-related data fed back by other subsystems of the software layer (such as decoding status, signal quality, resource allocation status, etc., used to comprehensively judge the timing and content of recommendation);

[0229] Input types: channel screen visual feature data, user behavior structured data, Internet program and social interaction related data, and operating status data fed back by other subsystems of the software layer;

[0230] Processing method: Adopt an algorithm based on content tags and collaborative filtering. By establishing multi-dimensional content tags for each channel program, combining user data to build a user interest profile, and exploring the potential associations between programs on different channels, the collaborative filtering algorithm is used to sort the associated content based on similar user viewing behaviors to generate a personalized recommendation list.

[0231] Output type: Personalized recommendation list data (including recommended channels, programs, and other information); user interest profile update data (for subsequent continuous optimization of recommendation accuracy);

[0232] Output flow: The personalized recommendation list data is output to the traditional interactive interface module and the intelligent interactive interface module of the user interaction layer, and displayed on the user interface in an intuitive way to guide users to watch; the user interest portrait update data is retained internally for subsequent continuous updating and improvement of the portrait, so as to further optimize the recommendation effect.

[0233] User Interaction Layer

[0234] Traditional interactive interface module

[0235] Data sources: Operation command data generated by users through remote control button operations (such as channel switching, playback control, etc.); channel decoding status data fed back by the multi-channel synchronous decoding subsystem of the software layer (such as the decoding status of the current playback channel, etc.); resource allocation status data fed back by the intelligent resource management subsystem of the software layer (to facilitate prompting users of system resource-related situations, such as resource allocation prompts when switching high-definition channels, etc.);

[0236] Input type: electrical signal data corresponding to the user's remote control key commands, and structured data of status information feedback from the software layer subsystem;

[0237] Processing method: After parsing the user's operation instructions, according to the set-top box system function logic, the corresponding instructions are passed to the corresponding subsystem of the software layer (such as passing the channel switching instructions to the multi-channel synchronous decoding subsystem, and displaying the resource allocation related prompts to the user, etc.);

[0238] Output type: operation instruction forwarding data (sent to the corresponding subsystem of the software layer); operation feedback and prompt information data displayed on the interface (displayed on the traditional interactive interface for users to view, such as operation success prompts, resource allocation prompts, etc.);

[0239] Output flow: The operation instruction forwarding data is output to the corresponding subsystems of the software layer, such as the multi-channel synchronous decoding subsystem and the intelligent resource management subsystem; the operation feedback and prompt information data displayed on the interface are displayed on the traditional interactive interface for users to view.

[0240] Intelligent interactive interface module

[0241] Data sources: interactive data generated by users through VR / AR devices, gesture recognition devices, voice input devices, etc. (such as gesture operation instructions, voice instructions, etc.); channel screen data output by the graphics processing core group of the hardware layer (used to be displayed in the immersive interactive interface); personalized recommendation list data output by the multi-channel content association and intelligent recommendation subsystem of the software layer (used to display recommended content); user behavior data during the intelligent interaction process (self-collected and recorded user operation behavior);

[0242] Input type: image data corresponding to gesture actions, audio data corresponding to voice commands, channel screen video data, recommendation list structured data, user behavior operation record data;

[0243] Processing method: Parse voice commands and convert them into commands that the system can recognize and execute; recognize gesture operations and convert them into corresponding channel switching, screen adjustment and other operation commands; use VR / AR technology to display channel screens to build an immersive interactive interface, display personalized recommendation lists, and collect user behavior data for subsequent analysis;

[0244] Output type: parsed operation command data (sent to the corresponding subsystem of the software layer for execution, such as the channel search command corresponding to the voice command is sent to the multi-channel synchronous decoding subsystem, etc.); displayed immersive interactive interface screen and recommended content data (presented to the user's viewing experience); user behavior data feedback (output to the multi-channel content association and intelligent recommendation subsystem for optimizing recommended content);

[0245] Output flow: The parsed operation instruction data is output to the corresponding subsystem of the software layer (such as the multi-channel synchronous decoding subsystem, the intelligent resource management subsystem, etc.); the displayed immersive interactive interface screen and recommended content data are displayed to the user; the user behavior data feedback is output to the multi-channel content association and intelligent recommendation subsystem of the software layer.

[0246] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprising a reference structure" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0247] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent set-top box system based on multi-channel synchronous processing, characterized in that: The system architecture includes a hardware layer, a software layer and a user interaction layer, wherein: The hardware layer is centered on a distributed multi-core processor module, which includes core groups such as signal processing, decoding, and graphics processing, which are responsible for the tasks of preliminary processing of multi-channel signal reception, decoding, and image rendering layout optimization. The storage unit provides support for data temporary storage and system software and related data storage, and the communication interface module ensures the connection with external devices and networks. The software layer includes a multi-channel synchronous decoding subsystem, a multi-channel signal interference elimination subsystem, an intelligent resource management subsystem, and a multi-channel content association and intelligent recommendation subsystem. The multi-channel synchronous decoding subsystem uses an algorithm based on priority queues and dynamic load balancing strategies to allocate decoding tasks according to channel priorities and core load conditions, and has error detection and correction capabilities. The multi-channel signal interference elimination subsystem adopts a method based on decision trees and signal feature analysis to determine the type of interference and eliminate it in a targeted manner by building a decision tree. The intelligent resource management subsystem uses a threshold-based resource monitoring and allocation algorithm to dynamically adjust hardware resources according to system load. The multi-channel content association and intelligent recommendation subsystem adopts an algorithm based on content tags and collaborative filtering to achieve cross-channel content association and personalized recommendation by building tags, analyzing user data and collaborative filtering. The user interaction layer includes a traditional interaction interface module and an intelligent interaction interface module. The traditional interaction interface module provides basic operating functions and works in conjunction with some subsystems of the software layer. The intelligent interaction interface module uses VR and AR technologies and gesture and voice interaction functions to provide users with an immersive multi-channel interaction experience, and uses interaction data feedback for recommended content optimization.

2. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The hardware layer includes a distributed multi-core processor module, a storage unit and a communication interface module. The distributed multi-core processor module is composed of multiple core groups, including a signal processing core group, a decoding core group, and a graphics processing core group. The signal processing core group is responsible for the reception, filtering and preliminary processing of multi-channel signals. It is equipped with a high-speed signal receiver and an advanced filter unit, and can process signal inputs of multiple channels at the same time to ensure stable reception and preliminary purification of signals.

3. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The multi-channel synchronous decoding subsystem of the software layer sets up a priority queue for each channel. The priority is determined based on the user's viewing history, channel subscription status, and current signal quality factors. The system then monitors the load of the decoding core group in real time and assigns the channel decoding tasks at the head of the queue to the decoding core with a lighter load. If the load on all cores is high, the decoding tasks of the low-priority channels are appropriately delayed or the frame rate is reduced based on the priority to ensure smooth playback of the high-priority channels.

4. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The multi-channel signal interference elimination subsystem of the software layer utilizes the construction of a decision tree based on the relationship between different interference types and signal characteristics obtained by analyzing a large number of offline signal samples. When a multi-channel signal is received, the system extracts the time domain and frequency domain characteristics of the signal, inputs these characteristics into the decision tree, and determines the interference type. For adjacent channel interference, according to the interference range determined by the decision tree, instructions are sent to the signal processing core group to dynamically adjust the filter parameters and accurately filter out the interference frequency components. For pulse interference, it works in coordination with the decoding core group and utilizes signal reconstruction technology to restore the signal of the interfered part according to the correlation between the previous and next signals to ensure the quality of the multi-channel signal.

5. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The intelligent resource management subsystem of the software layer pre-sets thresholds for different resources, and monitors in real time the resource requirements of each channel processing task in each core group of the hardware layer and the load of the overall system. When the system is in a low-load state, if the resource usage is lower than the set threshold, instructions are sent to the distributed multi-core processor module of the hardware layer to reduce the operating frequency and voltage of some unused or low-load cores, or even shut down some unnecessary hardware modules to save energy. When the system load suddenly increases, if the resource usage exceeds the threshold, the dormant resources are quickly awakened and the resources are reallocated to each task.

6. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The multi-channel content association and intelligent recommendation subsystem of the software layer establishes multi-dimensional content tags for each channel program, including program type, theme, actor, and director information, and then builds a user interest portrait by collecting and analyzing multi-dimensional data such as the user's viewing history, viewing time, channel stay time, and operating habits. At the same time, it interacts with content databases and social platforms on the Internet to obtain more program information and social interaction data. When a user is watching a program on a certain channel, the subsystem analyzes the program content of other channels in the background, matches the content tags with the user's interest portrait, and explores the potential associations between programs on different channels.

7. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The traditional interactive interface module of the user interaction layer provides basic channel list display, program information query, and playback control functions. Users perform routine operations such as channel switching and program selection in the traditional interactive interface by pressing buttons on the remote control. This module works closely with the multi-channel synchronous decoding subsystem and intelligent resource management subsystem of the software layer. When the user switches channels, the relevant subsystems are notified in time to allocate resources and adjust decoding tasks.

8. The intelligent set-top box system based on multi-channel synchronous processing according to claim 1, characterized in that: The intelligent interactive interface module of the user interaction layer includes an immersive interactive interface based on virtual reality (VR) and augmented reality (AR) technology, as well as gesture and voice interaction functions. In the immersive interactive interface, users view real-time images of multiple channels in a three-dimensional and dynamic form through head-mounted display devices (VR) or mobile devices such as mobile phones (AR). Through gesture recognition devices or mobile phone sensors, users can switch between different channels and adjust the size and layout of channel images.

Images