Summary:
UCLA researchers have developed a new algorithm to rapidly decode transmitted data packets with low latency.
Background:
Convolutional codes (CCs) are used extensively in wireless communication systems to improve data transmission reliability over noisy signals. CCs must be decoded following transmission to be accurately interpreted by receiving computer systems. The Viterbi Algorithm is the standard mathematical method to decode these messages, however, it is not efficient enough for ultra-low latency applications. Ultra-low latency applications are networks that are designed to process large quantities of data with nanosecond level delay. Applications of this technology span a wide array, including financial trading, telecommunications, defense and autonomous vehicle navigation. To enhance data transmission for high-speed communication architectures like 5G systems, there is a need to develop improved decoding algorithms.
Innovation:
Researchers in the Department of Electrical and Computer Engineering have developed a low Viterbi decoding implementation that is compatible with graphical processing units (GPUs) and other parallel processing systems to rapidly decode CCs. They improved the Viterbi Algorithm, which is commonly used for decoding transmitted data, by merging data structures instead of processing them one by one. This massively improves computational speed. This innovation has many applications for data processing, electronic signaling, and other high-speed communication systems.
Potential Applications:
• High-speed communications
• Autonomous vehicles
• Wireless networks
• High-frequency financial trading (HFT)
• Satellite communications
• Military/defense
• Cloud Computing
Advantages:
• Reduced latency
• Increased data processing throughput
• Improved error correction
Development-To-Date:
Researchers have validated the algorithm provides significant reduction in latency compared to the standard Viterbi Algorithm.
Related Papers (from the inventors only)
N/A
Reference:
UCLA Case No. 2024-171
Lead Inventor:
Professor Rick Wesel, PhD, Department of Electrical and Computer Engineering