Multiuser Detection with Base Station Diversity IEEE International
Multiuser Detection with Base Station Diversity IEEE International Conference on Universal Personal Communications Florence, Italy October 9, 1998 VIRGINIA POLYTECHNIC INSTITUTE & STATE UNIVERSITY MPRG MOBILE & PORTABLE RADIO RESEARCH GROUP Matthew C. Valenti and Brian D. Woerner Mobile and Portable Radio Research Group Virginia Tech Blacksburg, Virginia 10/9/98
Outline of Talk n n Outline n 10/9/98 n n Multiuser detection for TDMA systems. Macrodiversity combining for multiuser detected TDMA. The Log-MAP MUD algorithm. Simulation results for fading channels. Extensions to coded systems.
Multiuser Detection for the TDMA Uplink n For CDMA systems: MUD for TDMA u u u n For TDMA systems: u u u n Cochannel interference comes from other cells. Cochannel users do not have distinct spreading codes. Small number of cochannel interferers. MUD can still improve performance for TDMA. u u 10/9/98 Resolvable interference comes from within the same cell. Each cochannel user has a distinct spreading code. Large number of cochannel interferers. Signals cannot be separated based on spreading codes. Delay, phase, and signal power can be used.
Macrodiversity Combining for the TDMA Uplink Macrodiversity n 10/9/98 BS 1 n MS 1 BS 3 n MS 3 MS 2 BS 2 n In TDMA systems, the cochannel interference comes from adjacent cells. Interferers to one BS are desired signals to another BS. Performance could be improved if the base stations were allowed to share information. If the outputs of the multiuser detectors are log-likelihood ratios, then adding the outputs improves performance.
System Model MAI Channel Model n Received signal at base station m: n Where: u F u u n 10/9/98 a is the signature waveform of all users. Assumed to be a rectangular pulse. k, m is a random delay of user k at receiver m. Pk, m[i] is power at receiver m of user k’s ith bit. Matched filter output for user k at base station m:
Proposed System Model n 10/9/98 Each of M base stations has a multiuser detector. u Each MUD produces a log-likelihood ratio of the code bits. u The LLR’s are added together prior to the final decision. Multiuser Estimator #1 Multiuser Estimator #M
The Log-MAP MUD Algorithm n Optimal MUD uses the Viterbi algorithm u Log-MAP MUD u n Verdu, 1984 This algorithm produces hard bit decisions. The proposed system requires a multiuser estimation algorithm that produces LLR’s. u The symbol-by-symbol MAP algorithm can be used. F u The Log-MAP algorithm is performed in the Log domain, F n Robertson, Hoeher, Villebrun, 1997. The complexity of Log-MAP MUD is O(2 K). u 10/9/98 Bahl, Cocke, Jelinek, Raviv, 1974. u This is too complex for CDMA. However for TDMA, K is small, and this is reasonable.
Log-MAP MUD Algorithm: Setup 10/9/98 n Place y and b into vectors: n Place the fading amplitudes into a vector: n Compute cross-correlation matrix for each BS: u Assuming rectangular pulse shaping.
Log-MAP MUD Algorithm: Execution S 3 Log-MAP MUD S 2 S 1 S 0 i=1 Jacobian Logarithm: Branch Metric: 10/9/98 i=2 i=3 i=4 i=5 i=6
Simulation Parameters n The uplink of a TDMA system was simulated. u Simulation u 10/9/98 120 degree sectorized antennas. 3 cochannel interferers in the first tier F F u u u K=3 users M=3 base stations. Fully-interleaved Rayleigh flat-fading. Assume perfect channel estimation. No error correction coding.
Performance for Constant C/I n n n 10/9/98 C/I = 7 d. B Performance improves with MUD at one base station. An additional performance improvement obtained by combining the outputs of the three base stations.
Performance for Constant Eb/No n n n Performance as a function of C/I. u Eb/No = 20 d. B. For conventional receiver, performance is worse as C/I gets smaller. Performance of single-base station MUD is invariant to C/I. u n 10/9/98 Near-far resistant. For macrodiversity combining, performance improves as C/I gets smaller.
Macrodiversity Combining for Coded TDMA Systems n Coded Systems n n n Each base station has a multiuser estimator. Sum the LLR outputs of each MUD. Pass through a bank of Log-MAP channel decoder. Feed back LLR outputs of the decoders. Multiuser Estimator #1 Bank of K SISO Channel Decoders 10/9/98 Multiuser Estimator #M
Performance for Constant C/I n n n TDMA uplink. u K=3 mobiles. u M=3 base stations. C/I = 7 d. B Convolutionally coded. u u n Log-MAP algorithm. u u n MUD. Channel decoder. Iterative processing. u 10/9/98 Constraint length 3. Code rate 1/2. LLR from decoder fed back to MUD’s.
Conclusion and Future Work n Conclusions n n MUD can improve the performance of TDMA system. Performance can be further improved by combining the outputs of the base stations. This requires that the output of the MUD be in the form of a log-likelihood ratio. u n n FEC-decoders can provide a priori information to the base stations (see paper in Globecom CTMC). The study assumes perfect channel estimates. u u 10/9/98 Log-MAP MUD algorithm. The effect of channel estimation should be considered. Decision directed estimation should be possible. F Output of each base station can assist estimation at the others.
- Slides: 15