AIP Conference Proceedings 151 on Neural Networks for Computing
Self-organization and associative memory: 3rd edition
Self-organization and associative memory: 3rd edition
| Hi-index | 0.98 |
The problems of remote sensing of physical parameters from satellites is an important inverse, or system identification problem. Progress has been made in several phases. During the 1960's basic equations of radiative transfer in various types of media (inhomogeneous, emitting, anisotropic, polarized, etc.) were derived for the treatment of direct problems, i.e., problems in which the parameters were known. In order to avoid instabilities in numerically obtaining solutions, nonlinear Cauchy problems for systems of ordinary differential equations were developed. This formulation aided in the treatment of inverse problems, where the parameters are to be estimated based on observations of external or internal radiant energy. System identification techniques such as quasilinearization were found to be effective, when good preliminary estimates were available. Recently, a systematic procedure for obtaining the needed initial estimates has been investigated. This approach, based on the concept of associative memories, has already produced excellent computational results.