Shannon Theory
Capacity Computation

Computable Lower Bounds for Capacities of Input-Driven Finite-State Channels

V. Arvind Rameshwar, Navin Kashyap

Date & Time

01:00 am – 01:00 am


This paper studies the capacities of input-driven finite-state channels, i.e., channels whose current state is a time-invariant deterministic function of the previous state and the current input. We lower bound the capacity of such a channel using a dynamic programming formulation of a bound on the maximum reverse directed information rate. We show that the dynamic programming-based bounds can be simplified by solving the corresponding Bellman equation explicitly. In particular, we provide analytical lower bounds on the capacities of (d,k)-runlength-limited input-constrained binary symmetric and binary erasure channels.


V. Arvind Rameshwar

Indian Institute of Science

Navin Kashyap

Indian Institute of Science
Private Conversation

Reach out to the speaker privately

Questions & Answers

Post a publicly available question

No questions have been asked.

Session Chair

Tobias Koch

Universidad Carlos III de Madrid