Sampling and Bandwidth Expanding Communications
Thursday, Sep. 19, 2002
2:00-3:30 p.m.
Hughes Room
We consider the problem of designing low-complexity digital receivers for UWB and CDMA systems operating over channels with single or multiple propagation paths. We extend some of our recent sampling results for certain classes of non-bandlimited signals and develop a method that takes advantage of transform techniques to perform channel estimation and signal detection from a low-dimensional subspace of a received signal, that is, by sampling below the traditional Nyquist rate. A critical sampling rate in our framework is the signal innovation rate. By lowering the sampling rate, we can reduce computational requirements compared to existing solutions, allow for slower A/D converters and significantly reduce power consumption of digital receivers. In certain cases we demonstrate performance exceeding that of methods using Nyquist rate sampling. In effect, we use an algorithmic solution to improve hardware specifications in terms of complexity and power.

This is joint work with M. Vetterli (EPFL/UCB), J. Kusuma (MIT) and K. Ramchandran (UCB)

UC Berkeley Networking
Ashwin Pananjady and Orhan Ocal
Last Modification Date: Wednesday, February 10, 2016