Reliable Packet Detection in Random Access Networks: Analysis, Benchmark, and Optimization

Advanced vehicular communication systems extensively employ random access protocols to facilitate massive machine-type communications (MTC) with burst transmissions. Recent investigations have significantly reduced the decoding error rates in MTC. However, for random access, before a receiver decodes a packet, it has to first detect its arrival, since the presence of the packet at a particular point in time is not known beforehand. Packet detection, as opposed to packet decoding, could become a significant bottleneck, especially when considering that the widely used Schmidl-and-Cox (S&C) detection algorithm dates back several decades. This paper re-evaluates the S&C algorithm and presents several key improvements. First, we introduce a new “compensated autocorrelation” packet-detection metric that enables precise analytical expressions of false alarm and missed-detection probabilities, enhancing the rigor of performance assessment. Second, we establish a Pareto comparison principle that allows for a comprehensive and fair evaluation of packet detection algorithms by simultaneously accounting for both false alarms and missed detections. Third, we show that the performance of S&C can be significantly improved by considering only the real component of the autocorrelation, giving rise to a new real-part S&C (RP-S&C) scheme. Fourth, and perhaps most importantly, we leverage the newly introduced compensated autocorrelation to extend the single-antenna S&C algorithm for application in the multi-antenna setting. Two optimization problems, minimizing false-alarm and missed detection probabilities respectively, are formulated and addressed with rigorous solutions. Overall, this study holds considerable implications for the design and deployment of packet-detection schemes in vehicular random-access networks.