A joint adaptive rate control and channel estimation
algorithm based on
stochastic learning is presented. First, channel-matched quantizer
design is discussed. The quantizers are optimized empirically
based on simulation data. The application of linear reward inaction
learning to source rate control for video transmission is
studied. Some convergence properties of this method is
analyzed. The convergence of
the algorithm to the optimal channel matched quantizer is fast.
The learning delay and optimal quantization can be traded-off
using the learning parameter.
Simulation results are given for different channel bit error
probabilities and for various video sequences. The delay in the existing
schemes that employ pilot signaling for channel identification before
transmission is avoided in this method. Therefore, on-line rate control
is possible. The PSNR of the received video signal is shown to be
better using the proposed approach.
