UC Berkeley

The proposed triple-channel UWB-based cognitive radio exploits spectral crowding and coexistence of other wireless devices as the number of sensors and wearable computing devices increases in 3.1GHz to 10.6GHz ISM band to achieve energy efficient 1Gb/s short-range wireless communication. A dual-resolution analog wavelet-based spectrum performs bandwidth-and frequency-agile band pass filter (BPF) to detect narrowband and wideband interferers with low power consumption. A charge-pump-based triangular waveform generators and a source follower type low pass filter (LPF) generates basis function for the spectrum sensing with 132MHz sensing resolution. A Low power integer-N QPLL with reduced reference spur by digital calibration on mismatch of the charge pump current supports the tuning frequencies with a linear tuned wide range two stage ring-VCO and a low power programmable true-single-phase-clock (TSPC) divider. The proposed Triple-channel UWB-based cognitive radio was fabricated in 1V 65nm CMOS GP process. The test chip size is 2.3×2 mm2, and the active area is 2.1mm2. The data rate by using triangular shaped BPSK data is 1Gb/s at 1m communication. The lowest FoM of the energy/bit is 61pJ/bit, and the highest FoM is 102pJ/bit. It achieves BER from 9.2×10-7 to 1.1×10-4 according to frequency allocation of the triple-channels. The triple-channel UWB-based cognitive radio can provide energy efficient high-data rate wireless communication even with over 20% channel occupation.