Design of Spectrum Processing Chiplet Based on FFT Algorithm

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• The increasing complexity of electronic information and computer science necessitates efficient digital signal processing (DSP) techniques.
• General-purpose spectrum chips and FPGAs often face limitations in speed, precision, and resource utilization for spectrum processing tasks.
• Fast Fourier Transform (FFT) is a fundamental algorithm in DSP, critical for analyzing spectral signals.

Purpose of the Study:

• To present a chiplet design method for spectrum processing that overcomes the limitations of existing solutions.
• To improve the speed, precision, and resource utilization in spectrum processing.
• To enable efficient spectral analysis for applications like wireless communication.

Main Methods:

• Implementation of a Radix-2 4096-point FFT algorithm with a pipeline structure for spectral signal processing.
• Integration of a windowing module to optimize input data and mitigate spectrum leakage.
• Utilization of a clock gating unit (CGU) for low-power management of the clock reset.

Main Results:

• The designed chiplet successfully completes a 4096-point frequency sweep in 0.368 ms at a clock frequency of 61.44 MHz.
• Demonstrated significant improvements in both speed and accuracy for spectrum processing tasks.
• The chiplet design shows high potential for practical applications in wireless communication.

Conclusions:

• The proposed chiplet design method offers a substantial advancement in spectrum processing capabilities.
• The optimized FFT algorithm and integrated modules lead to enhanced performance and efficiency.
• This chiplet design holds significant promise for future wireless communication systems requiring high-performance spectral analysis.