Learn More
—We present and analyze a joint source-channel coding strategy for the transmission of a Gaussian source across a Gaussian channel in n channel uses per source symbol. Among all such strategies, the scheme presented here has the following properties: i) the resulting mean-squared error scales optimally with the signal-to-noise ratio, and ii) the scheme is(More)
We study the fusion of data collected by multiple heterogeneous sensors that work cooperatively to achieve a common goal. This paper presents fast algorithms to fuse the sensor data. We map the problem into a graphical model and then develop a fast message-passing scheme to fuse the data. We simulate scenarios with 150 sensors and 200 targets that are(More)
An analog source is to be transmitted across a Gaussian channel in more than one channel use per source symbol. This paper derives a lower bound on the asymptotic mean squared error for a strategy that consists of repeatedly quantizing the source, transmitting the quantizer outputs in the first channel uses, and sending the remaining quantization error(More)
We consider source coding with a fidelity criterion, channel coding with a channel input constraint, and the combined problem of reproducing a source across a noisy channel. All three cases face a similar tradeoff between resource and performance, and the operating point with the highest performance per resource is of particular interest. In the case of(More)
  • 1