RFI modeling and mitigation
Wireless receivers suffer degradation in communication performance due to radio frequency interference (RFI) generated by both human-made and natural sources. Human-made sources of RFI include uncoordinated wireless devices transmitting in the same frequency band, devices communicating in adjacent frequency bands and computational platform subsystems radiating clock frequencies and their harmonics. RFI is also generated from environmental sources such as atmospheric noise and electrical discharge. In typical communication receiver design and performance analysis, interference is usually modelled as having a Gaussian distribution. While the Gaussian distribution is a good model for thermal noise at the receiver, RFI has predominantly non-Gaussian statistics and is well modelled by distributions such as alpha-stable and Middleton Class A distribution.
Recent wireless communication standards and research have focused on the use of multiple antennas at the transmitter and receiver, a.k.a. multi input multi output (MIMO) communications, to increase data rate and communication reliability. . Wireless receivers with multiple antennas are being deployed in RFI rich network environments. While interference alignment and cancellation techniques offer some reduction in RFI at the receiver, they often require coordination among multiple users. Residual interference may be present due to uncoordinated users (out-of-cell interference) or users in co-existing networks. This motivates the need to derive accurate statistical models of RFI observed in multi antenna wireless receivers. Prior work on statistical models of interference in multiple antenna systems includes multidimensional extensions of Middleton Class A or symmetric alpha stable model. Such extensions are incomplete in that are not derived from a statistical-physical model of interference generation. I have derived statistical models of RFI generated from randomly distributed, uncoordinated interfering sources around a multi antenna receiver. This study has yielded joint statistics of RFI observed at multiple antennas in a field of randomly distributed interferers.
There is significant prior work on using transmit and receive spatial diversity to increase MIMO communication performance in fading wireless channels. The proposed transceiver algorithms differ in their computational complexity and practicality, motivating the need to study the relative communication performance of the various transmit and receive strategies. Knowledge of interference statistics is key to performance analysis of transmit and receive strategies in MIMO communications, however, statistical models of multi-antenna RFI used for such analysis are usually not derived from physical principles of interference generation. Based on the approximations used in statistical modeling, the actual communication performance of these diversity strategies may differ significantly from their expected performance. I use accurate statistical models of multi antenna RFI in analyzing communication performance of transmit and receive diversity techniques for wireless systems operating in the presence of RFI.
Another important property of wireless communications is the multipath nature of the communication medium between the transmitter and receiver. The frequency selective nature of wireless channels can cause signal distortion and is typically compensated at the receiver. Prior work on mitigating RFI in wireless transceivers operating in presence of frequency selective channels, shows communication performance gains are achievable if accurate statistical models of interference are assumed during receiver design. However, there are few statistical models of multi antenna RFI that are based on physical principles of interferer placement and RFI generation. Using statistical-physical models of RFI, linear and non-linear receiver algorithms may be designed to achieve improved receiver performance in presence of RFI in multipath fading wireless channels. My research involves studying novel linear and non-linear filters to improve performance of single-carrier MIMO wireless communication in frequency selective fading channels and study the communication performance vs. implementation complexity tradeoffs for these methods.