NON-ITERATIVE PARAMETER IDENTIFICATION TECHNIQUES

Central to virtually all aspects of helicopter design and evaluation is an appropriate mathematical model. The model structure is chosen to adequately represent the physical behavior of the phenomena under study. These models are usually parametric in nature and so the analyst is faced with the task of determining appropriate values to be assigned to the relevant parameters. This is often done by observing input/output measurements and choosing the parameters that in some sense best fit the model to the observations. This procedure is known as parameter identification. In the following paper the model structure is restricted to being linear and time-invariant. Descriptions of three different non-iterative techniques that can be used to determine parameter estimates are given in this paper. The techniques are Deterministic (Recursive) Least Squares (DLS), the Extended Kalman Filter (EKE), and the Statistically Linearized Filter (SLF). The methods are then used to identify the parameters of two different systems. The first is a simple one state, two parameter system. The second is a four state, twenty four parameter system taken from a linearized longitudinal model of an Advanced Light Helicopter (ALH) in forward flight.