Upper bounds on the capacity of binary intermittent communication

This paper focuses on obtaining upper bounds on the capacity of a special case of intermittent communication, introduced in [1], in which the channel is binary-input binary-output noiseless with i.i.d. number of zeros inserted in between the binary codeword symbols. Upper bounds are obtained by providing the encoder and the decoder with various amounts of side-information, and calculating or upper bounding the capacity of this genie-aided system. The results suggest that the linear scaling of the receive window with respect to the codeword length considered in the system model is relevant since the upper bounds imply a tradeoff between the capacity of the channel and the intermittency rate.