Non-uniform Information Transmission for Minimum Distortion in Wireless Networks
This paper considers average input-output distortion minimization through joint optimization of source index assignment and modulation design. First, we derive the general optimization criterion, and discuss the possibility of simultaneous minimization of bit-error-rate (BER) and distortion. Secondly, we propose a novel source-aware information transmission approach by exploiting the non-uniformity in Gray-coded constellations. The proposed approach makes it possible for simultaneous BER and distortion minimization and outperforms the existing schemes with big margin when the original geometric structure of the quantization codebook can not be maintained in information transmission. The simplicity and power efficiency of the proposed source-aware non-uniform information scheme make it particularly attractive for systems with tight power constraints, such as wireless sensor networks and space communications.
Unable to display preview. Download preview PDF.
- 5.Masnick, B., Wolf, J.: On linear unequal error protection codes, vol. 13, pp. 600–607 (October 1967)Google Scholar
- 10.Zeger, K., Gersho, A.: Vector quantization design for memoryless noisy channels. In: Proc. IEEE Int. Conf. Commun, Philadelphia, PA, June 1988, pp. 1593–1597 (1988)Google Scholar
- 12.Forney, G.D., Lee-Fang, W.: Multidimensional constellations - Part I: Introduction, figures of merit, and generalized cross constellations. IEEE Trans. Inform. Theory 7, 877–892 (1989)Google Scholar