Predictive Compression Transmission for Low-Latency Real-Time Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transmission methods in remote intelligent diagnosis and monitoring systems face challenges in achieving real-time visibility and high-density, high-precision data collection while maintaining efficient bandwidth utilization, as they are hindered by inefficient data compression and excessive delay due to frame header confirmations.
Innovation Solution
A method combining real-time and non-real-time communication, utilizing predictive compression features based on a self-coding network and fully connected neural network, where real-time data is compressed and decompressed using encoder and decoder parameters, and non-real-time data is cached and compressed, optimizing bandwidth usage and reducing delays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If frame header confirmation protocol is executed for each data transmission, then data transmission reliability is improved, but data transmission efficiency deteriorates due to increased confirmation overhead
Solution Approach 1:
The patent segments the transmission process into two distinct regions: a first region for real-time data transmission with frame header confirmations, and a second region for non-real-time data transmission without confirmations. This segmentation allows the system to maintain reliability for critical real-time data while improving overall efficiency by avoiding unnecessary confirmations for non-critical data.
Solution Approach 2:
The patent dynamically adjusts the transmission strategy based on data characteristics and timing. By dividing transmission periods into different regions and using predictive compression features, the system can adaptively decide whether to use confirmation protocols based on the specific transmission context, balancing reliability and efficiency dynamically.
2Measurement precision
If real-time data is transmitted with frame header confirmations, then data validation accuracy is improved, but transmission delay increases
Solution Approach 1:
The patent segments transmission into real-time and non-real-time regions. For non-real-time data in the second region, the system transmits without frame header confirmations, significantly reducing transmission delay while maintaining acceptable validation accuracy through alternative verification mechanisms.
Solution Approach 2:
The patent uses predictive compression features generated in advance to predict future data states. This preliminary action allows the system to prepare compression models before actual transmission occurs, reducing the time needed for real-time processing and validation.
3Loss of time
If business data volume is reduced for each frame transmission, then transmission delay is reduced, but data completeness deteriorates
Solution Approach 1:
The patent segments data into real-time critical data and non-real-time data. By transmitting only essential real-time data with confirmations while sending non-critical data without confirmations, the system reduces transmission delay without significantly compromising overall data completeness.
Solution Approach 2:
The patent extracts and separates non-real-time data from real-time data transmission streams. This extraction allows the system to transmit only the most critical real-time data with full validation when needed, while sending less critical data through a faster, confirmation-free channel.
4Quantity of substance
If full bandwidth is occupied by business unrelated data, then data collection coverage is improved, but real-time transmission capability deteriorates
Solution Approach 1:
The patent segments the transmission timeline into real-time and non-real-time regions. By dedicating the first region to real-time data with confirmations and the second region to non-real-time data, the system ensures real-time transmission capability is maintained while still collecting comprehensive data through the non-real-time channel.
Data Source
AI summary
A compression transmission method for real-time data is provided, when both a sending terminal and a receiving terminal have accumulated historical data, that is, there is business knowledge. The business knowledge is used by the receiving terminal to construct and train a self-coding network and a fully connection neural network, to thereby construct an output prediction model of an encoder. After the model converges, parameters of the model and a predicted result of an intermediate layer of the self-coding network are sent to the sending terminal by the receiving terminal with a nonreal-time method. After verifying by the sending terminal, the sending terminal decides to send the real-time data or not to send the real-time data to the receiving terminal, and the real-time data is decompressed into simulated real-time data by the receiving terminal according to the predictive result of the intermediate layer of the self-coding network.