Predicting burst error statistics of digital wireless systems with HARQ

Hybrid Automatic Retransmission reQuest (HARQ) is an effective technique to improve the reliability of wireless communication systems by detecting, correcting, and retransmitting the erroneous packets. Packet-level error sequences obtained from physical layer wireless communication systems are important for the design and performance evaluation of high layer protocols, e.g., HARQ. In this paper, we utilize the open source Vienna long-term evolution (LTE) simulator to study the impact of HARQ on the burst error statistics of packet-level error sequences. Moreover, we propose a generative model that can generate packet-level error sequences with predicted burst error statistics similar to those of error sequences obtained from wireless systems with HARQ. Simulation results demonstrate that the proposed generative model is accurate and efficient in predicting the behavior of HARQ in terms of a set of burst error statistics rather than predicting the packet error rate (PER) only.