Dynamic Kernel Function Fast Fourier Transform with Variable Truncation Scheme for Wideband Signal Detection

Dual thresholding for digital wideband receivers with variable truncation scheme

“…Hamming window is applied to the input data to improve the frequency selective characteristics of the FFT. Utilizing the variable truncation scheme proposed in [6], depending on whether any of the 128-point 10-bit input data exceed the full scale value of 8-bits, the most significant 8-bits (MSB) or the least significant 8-bits (LSB) are selected and sent to the 8-bit inputs of the first stage butterflies. Variable truncation scheme is also applied after stage one of the FFT to further improve on the dynamic range of the receiver.…”

“…The 8-bit 128-point fixed-point fixed-precision FFT design methodology in [6] is followed with three levels of folding. Dynamic kernel function is used to represent the twiddle factors needed for FFT computation.…”

“…A single threshold is sufficient for the above designs because they are limited to one operational dynamic range for signal detection. The application of automatic gain control or variable truncation scheme (VTS) [6] increases the overall dynamic range of the receiver, where signal detection may now operate in different regions of the dynamic range depending on the status signal of the VTS comparators. Numerous adaptive thresholding techniques are discussed in [7] requiring the mean and variance of the noise floor in the interested frequency spectrum to determine the threshold.…”

Adaptive thresholding for high dual-tone signal instantaneous dynamic range in digital wideband receiver

“…Substantial improvement on the dualtone signal IDR using the super resolution and the compensation techniques to remove side lobes and spurs in frequency domain after the FFT was first proposed in [2][3]. Techniques to minimize the FFT hardware while maintaining a good IDR were proposed in [4] using the dynamic kernel function and in [5] using a smaller FFT point size with a set number of digital channel filter banks. The effects of a fixed resolution between FFT stages on the detection of signals at various thresholds were studied in [5].…”

“…The effects of a fixed resolution between FFT stages on the detection of signals at various thresholds were studied in [5]. The common approach of employing a single fixed threshold is applied to the above design with approximately 80-90 % detection rates of weak signals [2][3][4][5]. A single threshold may be sufficient if signals processed in the FFT stages are limited to one operational dynamic range.…”