Work

Artificial Intelligence for Iteration Count Prediction in Real-Time CORDIC Processing

Year: 2025

Type: article

Abstract: The first research attempt to dynamically optimize the CORDIC algorithm’s iteration count using artificial intelligence is presented in this paper. Conventional approaches depend on a certain number of iterations, which frequently results in extra calculations and longer processing times. Our method drastically reduces the number of iterations without compromising accuracy by using machine learning regression models to predict the near-best iteration value for a given input angle. Overall efficiency is increased as a result of reduced computational complexity along with faster execution. We optimized the hyperparameters of several models, including Random Forest, XGBoost, and Support Vector Machine (SVM) Regressor, using Grid Search and Cross-Validation. Experimental results show that the SVM Regressor performs best, with a mean absolute error of 0.045 and an R2 score of 0.998. This AI-driven dynamic iteration prediction thus offers a promising route for efficient and adaptable CORDIC implementations in real-time digital signal processing applications.

Source: Mathematics

DOI: 10.3390/math13243957

Authors: Ratheesh Sudheerbabu, Lekshmi Chandrika Reghunath, Valentina Franzoni, Alfredo Milani, Cristian Randieri

Institutions: Amrita Vishwa Vidyapeetham, University of Perugia, Link Campus University, Università degli Studi eCampus

Language: English

Topic: Numerical Methods and Algorithms

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $3,137

Funder: Ministry of Science and Higher Education of the Republic of Kazakhstan