Data Fusion for Least Squares
Andrew Thompson · Defense Technical Information Center (DTIC) · 1991
The wide array of uses for least-squares estimation testifies to its effectiveness. The key to structuring a problem for at least-squares solution is finding a Markov representation of the problem. This representation defines a recursive approach to estimation. When multiple estimates are available at each time step, the processing time can be decreased by using data fusion networks to reduce the information to a single estimate. Hierarchical networks, using both parallel and serial combination of data, can be devised. The cases presented herein can be used to preprocess the for any recursive least-squares method.