ON NORMAL APPROXIMATIONS TO THE ERROR PROBABILITY FOR A CORRELATION RECEIVER IN A HARD-LIMITED CHANNEL.
INSTITUTE FOR DEFENSE ANALYSES ARLINGTON VA RESEARCH AND ENGINEERING SUPPORT DIV
Pagination or Media Count:
The advent of communication satellites has stimulated interest in signal designs allowing access to a repeater by many independent carriers. To maximize the power efficiency of the satellite, consideration has been given to the use of a repeater which can be characterized as a hard-limited channel. A technique for achieving multiple access to a hard-limited channel is to use constant-envelope, pseudo-noise phase-coded spread-spectrum carriers p-n carriers with correlation-type receivers. The interferring signals at the limiter input are modeled as a statistically independent zero mean stationary Gaussian noise process. This paper examines the approximation to Normal of the error probability when binary symmetric bit-to-bit independent message sources are received by a coherent stored reference correlator after passing through a hard-limited channel. The approach used here follows that of Shannon and Gallagher for computing the probability tail with Chernoff bounding techniques. This method avoids the usual objections raised to the Edgeworth series when used to compute an error probability. Subject to restrictions of weak carrier-to-noise ratios at the channel input and error rates less than 10 to the -2, the conclusion reached is that the error probability can be computed, to within the accuracy of the mathematical idealization of a real channel, as if the correlator output were Normal. Author
- Unmanned Spacecraft
- Non-Radio Communications