My Research Interests and Activities

Research Themes I have investigated or are investigating include: -

 


Femtocells are wireless access points for mobile broadband cellular technology.  Femtocells are self-deployed and therefore the first step on the path to building an ad-hoc wireless network. Femtocells have many unanswered research problems associated with them.  A major industry concern with femtocells is related to the interference aspects of femtocells.  My research team conducted extensive work into understanding the interference effects of a femtocell underlay system in WCDMA/3G.  An extensive world-first system simulation of hundreds of femtocells, thousands of users and tens of macro cells was configured to better understand the interaction effects between the macro and femtocell system.  From these numerical results both residential and enterprise system results have been developed which highlight the benefits of femtocell interference cancellation.  The benefits to users include improved throughput and range, while reducing the impact on the macro cell uplink throughput.

  We have looked extensively at the signal parameter acquisition problems associated with OFDM, these include coarse timing synchronisation, integer carrier frequency offset (ICFO) and fractional carrier frequency offset (FCFO) problems. The research looked at practical solutions with both stationary and mobile channels, as well as deriving analysis based on the Cramer-Rao bounds. Finally work has been conducted on the application of multiuser detection and interference cancellation for the contention based OFDMA ranging channel. Existing literature treats the MAI as noise, so the performance degrades dramatically as the number of ranging subscriber stations increases. Our proposed method successively detects the channel paths of active ranging signals and cancels their interference for further detection. This approach significantly suppresses the MAI between ranging subscriber stations and improves the ranging detector’s performance. With reconstructed ranging signals in the time domain, we are also able to effectively mitigate the interference from the ranging channel to data subscriber stations.  The proposed technique allows ranging subscriber stations to complete the ranging processes in a shorter time with smaller residual timing and frequency offsets.
 
We performed detailed analysis and investigation into a novel iterative channel estimation for WCDMA, IDMA, and OFDM systems.  We performed statistical analysis on this and showed that this technique could overcome the packet loss associated with high-speed mobility.  We realised this algorithm in a real-time WiMAX modem. We demonstrated through outdoor trials the packet throughput benefits of the modem. From laboratory trials at radio frequency with real-time Jakes channel models we showed at 60km/hr that the technique provided packet throughput improvements of 50% and 100% at speeds of 60km/hr and 200km/hr, respectively.
IDMA is a new modulation method that assumes the use of iterative detection at the receiver.  We have performed world first research work into IDMA systems showing how the system model can be extended to also include the IDMA interleaver.  This model allows the derivation of the optimal detector and provides additional insights into the IDMA principle. Bit error rate and EXIT charts were used to study the performance of both parallel and serial interference cancellation.  We showed that a bit interleaved DS-CDMA system is essentially a special case of IDMA and therefore an IDMA receiver can be used to detect multi-user DS-CDMA signals.  We have also performed other work on asynchronous IDMA and timing acquisition and tracking for IDMA. This work is continuing with one of my research students where we are investigating methods to determine, analytically, the optimal combination of channel coding and IDMA to achieve near capacity performance.
Initially with wireless modems the challenge is to provide the best performance possible, however, in practical systems the challenge is to achieve the best complexity vs. performance point, or the minimal cost (chip die area) for a given level of performance.  In multi-user systems the resources can be shared among users and a significant amount of optimisation can be performed to minimise the overall cost.  We came up with world-first results in this area and showed for a multiple-user DS-CDMA base station system loosely based on the 3GPP standard that we could reduce the overall computational complexity by 50% for a given performance metric (Bit error rate performance).  The techniques are based on utilising EXIT analysis dynamically and in real-time, to constantly monitor modem performance and resource allocation.  Inclusion of optimal power levels for each user can improve receiver sensitivity by an average of 8dB.
 
Students that I co-supervise have cast the compressed sensing (CS) problem for sparse binary and trinary vectors as a communication  problem, and use a generalised inter-leave division multiple access (IDMA) receiver to solve it. The result is that for the same sparsity, the receiver has a higher recovery probability than any of the previously reported techniques to date, including the popular `1 minimisation. The performance gain is substantial for binary data in particular, but less pronounced for trinary data. In binary scenarios, where the traditional CS algorithms reported a 0% chance of recovery, the receiver may still achieve a 90% success rate. For dim 128 256 measurement matrices this also holds true in the presence of AWGN, whenever the SNR is higher than 22dB.
Investigated interference cancellation for CPM using initially the super state detector and then Laurent decomposition.  Extended research to ISI channels.  Also performed comparison of optimal and sub-optimal iterative equalisation techniques for full-response CPM, as well as investigated the addition of iterative channel estimation within a serially concatenated CPM system.  Analysed the systems using EXIT charts and included pre-filtering to reduce the effect of the ISI channels.  Developed doubly iterative receivers that worked well over harsh ISI Channels0.

Author: Mark C. Reed, Last Revised : 17 Dec 2009