Cooperative Algorithms in Adaptive Signal Processing

• Published in: IEEE Green Energy and Systems Conference pp. 1 - 6
• Main Author: Jalali, Sammuel
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2018
• Subjects: Adaptive systems; Channel equalization; Computational efficiency; Conferences; Convergence; Cooperative algorithms; Delays; Diversity in algorithm; Equalizers; Signal processing algorithms
• ISSN: 2640-0138
• DOI: 10.1109/IGESC.2018.8745524

Advanced adaptive signal processing algorithms commonly used in optimization, filtering, and estimation are subjects of seeking innovations and advancements. The important performance factors of algorithms are adaptation or convergence (the probability of, and the speed of), and computational cost (proportional to the algorithm complexity). The improvement in convergence often comes with higher computational needs. In this report we exploit the diversity in algorithms by using cooperative processes. An odd number of cooperative processes running simultaneously participate in a simple majority-vote decision method to improve the convergence speed of the corresponding application. The iterative algorithms concurrently progressing require higher computational resources, however only the most computationally-efficient algorithm remains active after the initial convergence is achieved. Additional technique is essential to synchronize the process delays of participating algorithms.