Convergence in mean for ELS-based adaptive control (Q1070211)

The convergence in mean for the adaptive control algorithm studied by \textit{R. Kumar} and \textit{J. B. Moore} [IEEE Trans. Autom. Control AC-27, 146-153 (1982; Zbl 0469.93069)] is investigated. In the article loc. cit. almost sure convergence of this algorithm has been shown. In the paper the assumptions are formulated under which the algorithm converges in p- th mean, where \(p=2-\eta\), \(\eta >0\). As the almost sure convergence does not imply the p-th mean one, both convergence results are stronger than each one of them, separately. The considerations in the paper show that the combination of these two forms of convergence results makes the stability results more robust and potentially of more practical significance.