Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Interference identification in cellular networks via adaptive projected subgradient methods

: Oltmann, K.; Cavalcante, R.L.G.; Staczak, S.; Kasparick, M.


Matthews, M.B. ; Naval Postgraduate School -NPS-, Monterey/Calif.; IEEE Signal Processing Society:
Asilomar Conference on Signals, Systems and Computers 2013. Vol.3 : Pacific Grove, California, USA, 3 - 6 November 2013
Piscataway, NJ: IEEE, 2013
ISBN: 978-1-4799-2391-5
ISBN: 978-1-4799-2390-8
ISBN: 978-1-4799-2388-5 (Print)
Asilomar Conference on Signals, Systems and Computers <47, 2013, Pacific Grove/Calif.>
Fraunhofer HHI ()

We develop an adaptive algorithm to estimate a channel gain matrix in cellular heterogeneous networks. This algorithm has the objective of providing important information to interference coordination and management schemes, a crucial functionality of 'beyond 2020 networks'. In more detail, we pose the estimation problem as a set-theoretic adaptive filtering problem. In the proposed scheme, the channel gain matrix is tracked with the adaptive projected subgradient method (APSM), a powerful iterative tool that can seamlessly use prior information and information gained by measurements. More precisely, we construct multiple closed convex sets, each of which containing estimates that are consistent with a piece of information about the channel gain matrix. The intersection of these sets corresponds to estimates that are consistent with all available information. In particular, we use the following information to construct the sets: i) physical upper and lower bounds of the path gains, ii) interference bounds for the downlink and uplink communication, and iii) received signal received power (RSRP) measurements. The algorithm produces a sequence of estimates where each term is an estimate that approaches the intersection of the multiple sets available at a given time instant. Simulations show that the proposed algorithm is able to track the channel gain matrix in scenarios with mobile users, and it outperforms standard adaptive filters that do not use prior information.