Multiple Concurrent Recursive Least Squares Identification

A new algorithm, multiple concurrent recursive least squares (MCRLS) is developed for parameter estimation in a system having a set of governing equations describing its behavior that cannot be manipulated into a form allowing (direct) linear regression of the unknown parameters. In this algorithm, the single nonlinear problem is segmented into two or more separate linear problems, thereby enabling the application of existing powerful linear regression algorithms such as recursive least squares (RLS). The individual linear sub-problems contain unknown parameters other than those that are identified; said parameters are initially set at their nominal values, and are subsequently updated by the other concurrently running identification (ID) processes. With all sub-problems sharing their results following each update, the results rival those of more computationally intensive nonlinear optimization algorithms. This algorithm was developed to address, and has been tested on a spacecraft on-line mass-property ID application. Beginning with reasonably accurate initial parameter estimates, the approximation error for this spacecraft mass-property ID example is negligible as compared to errors created by other un-modeled system parameters.