Distributed Turbo Coding in the Presence of Inter-User Channel Impairment

We propose a distributed turbo coding (DTC) scheme, where two users cooperate in a two-phase cooperation scenario constituted by the exchange of data between the users in the first phase followed by the uplink transmission of the two users´ data in the second phase of cooperation. Following the information exchange phase, the data transmitted from the second user´s transmitter during the second phase is an interleaved version of the first user´s data. Each user´s transmitter is equipped with a recursive systematic convolutional (RSC) code, an interleaver and a sphere packing (SP) mapper. At the receiver side, interference cancellation is employed, which outputs two branches of decoded data corresponding to the data transmitted from the two users´ transmitters. Iterative detection is employed by exchanging extrinsic information between the RSC decoder and the SP demapper in each decoding branch, as well as between the RSC decoders in the two branches. The proposed DTC scheme is benchmarked against another cooperative scheme, where the two users do not exchange their data, they rather transmit their data simultaneously. At the receiver of the benchmark scheme, iterative detection is carried out by exchanging extrinsic information between the SP demapper and the RSC decoder of each user, but no iterations are invoked between the RSC decoders in the two branches. The proposed scheme attains an E_b/N_o gain of about 25 dB at BER of 10^-5, when compared to the benchmark scheme, while considering a perfect inter-user channel (IUC) and employing I_sys=5 system iterations in conjunction with an interleaver depth of D_int=80, 000 bits. Additionally, we study the effects of both Gaussian as well as Ricean IUC on the achievable BER performance of the uplink DTC scheme.