A method for detecting perturbation of a physical system from a reference state associated with a reference parameter (omega0) to a perturbed state associated with a perturbed parameter (omega) includes firstly deriving the reference parameter (omega0). A reference vector (F ) is then derived which describes the system's state at the reference parameter (omega0.)A measurement-related vector (Z) associated with a perturbed state of the system is then subtracted from the reference vector (F ) to provide an error vector (E). The error vector members are summed and normalised by division by a summation of elements of a vector (F') representing a derivative (f'(omega0,deltaomega)e) of a reference function (f(omega0,deltaomega)e) itself represented by the reference vector (F), the derivative (f'(omega0,deltaomega)e) being evaluated at the reference parameter (omega0). This provides a result equal to the difference ( omega - omega0 ) between the perturbed parameter (omega) and the reference parameter (omega0)
