A Channel Coding Approach for Random Access Communication with Bursty Sources

We extend Information Theoretic analysis to time-slotted packet random access communication with bursty sources. A new channel coding approach for coding within each packet is proposed with built-in support for bursty sources phenomena, such as message underflow, and for random access mechanisms, such as packet collision detection. The coding approach does not require joint communication rate determination either between the transmitters or between the transmitters and the receiver. Its performance limitation is characterized by an achievable region defined in terms of communication rates, such that reliable packet recovery is supported for all rates within the region and reliable collision detection is supported for all rates outside the region. For random access communication over a discrete-time memoryless channel using a class of random coding schemes, it is shown that the maximum achievable rate region of the introduced coding approach equals the Shannon information rate region without a convex hull operation.