Robust procedures for communication data
Abstract
The qualitative theory of robustness in parameter estimation was modified and extended to include robust filtering. The qualitative properties of robust filters were studied. The qualitative analysis of robust filtering and smoothing provided some specific sufficient conditions regarding the filtering and smoothing operations. A nonlinear such operation is included. Robust linear prediction and interpolation for convex and compact families of vector stationary processes were studied. The problem as a stochastic game with saddle point solution was formalized and it is found that this solution applies to a number of processed families. Such families include the family of linearity contaminated nominal processes, and the class of processes whose total energy is fixed within prespecified quantities. Robust linear filtering for convex and compact families of vector stationary information and noise processes is studied. Robust solutions for various such families are found and their performance was studied for efficiency, performance deviation, and breakdown points or curves. Properties of two dimensional random processes as applied to imaging are examined. The multiple access problem where a finite number of packet transmitting data users access a single channel while they sense the feedback only when blocked is considered. A random access, collision resolution protocol for this case is analyzed.
- Publication:
-
NASA STI/Recon Technical Report N
- Pub Date:
- August 1983
- Bibcode:
- 1983STIN...8418493P
- Keywords:
-
- Limits (Mathematics);
- Multiple Access;
- Robustness (Mathematics);
- Stochastic Processes;
- Telecommunication;
- Game Theory;
- Interpolation;
- Prediction Analysis Techniques;
- Vector Analysis;
- Communications and Radar