We propose a hybrid ARQ scheme using recursive convolutional codes, the turbo principle and combining diversity. While spending the same energy for transmission, the proposed scheme presents a better throughput for most of the SNR range and a smaller decoding complexity than another well-known method.

We propose a hybrid ARQ scheme using recursive convolutional codes and the turbo principle. In the first transmission the system encodes the data through a systematic recursive convolutional encoder. In the second transmission data is interleaved prior to being encoded by a recursive convolutional encoder. Then, from the second transmission on the receiver operates as a turbo decoder. The proposed method is compared with a popular scheme known in the literature. It is shown that, while spending the same energy for transmission, the proposed scheme presents both smaller FER and decoding complexity.

A hybrid ARQ scheme for multiple transmit antennas systems is presented. Space-diversity is achieved through the use of random signal mapping (RSM). The use of RSM instead of other typical space-time coding schemes, as space-time trellis codes or space-time block codes, allows that more antennas may be added to the system with negligible complexity increase at the receiver. Coding gain is achieved with a standard parallel turbo code. Numerical results for frame error rate, throughput and number of required retransmissions for a varying number of transmit antennas are presented. The impact of the proposed scheme in the performance of a higher layer protocol, as TCP, is also investigated through computer simulations.