I want to improve the range accuracy in the "cascade_radar_object_detect usecase" by applying zeropadding in the fft algorithm. Note that the zero-padding which is actually implemented is not what I need, as it is applied to reach the next power of two on the number of adc samples. For example, with 256 valid adc samples I want to compute the fft over 1024 samples (256 valid samples and 768 zeros). Is this possible? Which source files should I modify (I guess I need to modify fft parameterers and buffer dimensions).
Thank you,
Sara Salvador