Coding for Turbulent Channels. Volume III. Turbulent Channel Data Analysis and Evaluation of Coding Techniques.
Abstract
The data processing and analysis operations performed on the channel data recordings of Volume II are described, and representative numerical results are discussed. Some of these results are based on defining error bursts in terms of error gaps, instead of error densities (or as the absence of an error-free interval). This basis for defining bursts was found to be advantageous. One important result was the finding that when bursts defined in this way contain only short gaps, the bursts have only limited lengths but are quite closely spaced. Another result is that for the channel data tapes processed, the signal quality parameters are relatively unreliable indicators of demodulator output errors. One result of the evaluation of coding techniques for turbulent channels indicates that multiple-interleaved cascade F-multi-codes are advantageous. Another one of these results is that in the present state of the art, these multicodes should be based on cyclic, shortened cyclic, or extended cyclic subcodes, and the inner subsystem should use a bit error detecting F-singlecode with a bit EDAC augmented digital decoder. If the error control performance requirements are sufficiently high, and if the channel bit rate is sufficiently low (e.g., less than 9600 bps), then a hybrid hardware/software decoder implementation may be advantageous if otherwise permissible. In all other cases, a completely hardware implementation is preferable. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1970
- Accession Number
- AD0869942
Entities
People
- L. E. Colley
- M. E. Mitchell
Organizations
- General Electric