Brian Richards

Brian Richards received the B.S. degree in Electrical Engineering from the California Institute of Technology in 1983, and the M.S. degree in Electrical Engineering and Computer Science from the University of California, Berkeley in 1986. From 1986, he joined the research staff at the University of California, Berkeley, where he worked on large scale digital system design projects including speech recognition, full-custom ASIC design for image processing, and the Infopad portable wireless multimedia terminal. He is a founding member of the BWRC, maintaining and continuing the development of several ASIC and FPGA system design CAD tools and related libraries.  Current projects include supporting various research efforts related to prototyping and implementing wireless and low-power systems at the Berkeley Wireless Research Center.

BWRC Publications

System Design of Cooperative Wireless Networks

Project Members: Milos Jorgovanovic, Sameet Ramakrishnan, Matthew Weiner, Kathy Sun, Prof. Borivoje Nikolic, Prof. David Tse

Cooperative relaying has been envisioned as a promising technique for improving spectral efficiency in wireless networks. One way to increase the data rate of todays dense wireless networks is by utilizing the proximity of the nearby terminals. When a wireless terminal is allowed to cooperate with the nearby wireless terminals, it can utilize this proximity gain to “borrow” the antennas from these devices and form a virtual multiple-antenna transceiver. This is the basic principle of cooperative Multiple-Input Multiple-Output (MIMO) transmission, the technique that resembles that of the standard MIMO systems found in modern wireless standards, such as Wifi 802.11n and 4G LTE.

In this project we are exploring physical-layer cooperation in the uplink scenario (mobile station to base station) from a system design perspective. The project focusses on the network configuration presented on Figure 1, with one single-antenna transmit mobile terminal (source), one multiple-antenna access point (AP) or base station (BS) (destination), and many single-antenna mobile terminals (relays) close to the source and available for physical-layer cooperation. The relays are assumed to be half-duplex, i.e. they can either be in the transmit (Tx) or the receive (Rx) mode.

We have looked into several aspects of how to design such cooperative network. First, we simulated what is the capacity gain for a different number of relays and typical channel conditions. We demonstrated around 3x increase in QMF achievable rate with up to 5 relays. Second, we looked at efficient relay scheduling (when relays should be in Tx and when in Rx mode) algorithms. We have suggested a simple local scheduling algorithm that can operate in real-time and achieve an optimal rate for small number of relays, Figure 2. Third, we are looking into relay interference issue. With other nearby relays transmitting at the same time, some amount of interference from other relays is inevitable. We have suggested an efficient algorithms that can mitigate the effect of relay interference. Lastly, we would like to show proof-of-concept with a hardware implementation of the simplified version of this cooperative network. The hardware prototype has a single source, a single relay, and a two-antenna destination. We are building the complete baseband using the National Instruments (NI) platform for design and testing of wireless communication systems. The system will run on FPGA chips and use standard radio boards supplied by NI, Figure 3.

Figure 1: cooperative wireless network

Figure 2: data rate scaling with the number of relays

Figure 3: cooperative MIMO hardware prototype