PMAC: probabilistic multimodality adaptive control

ABSTRACT

This paper develops a probabilistic multimodal adaptive control approach for systems that are characterised by temporal multimodality where the system dynamics are subject to abrupt mode switching at arbitrary times. In this framework, the control objective is redefined such that it utilises the complete probability distribution of the system dynamics. The derived probabilistic control law is thus of a dual type that incorporates the functional uncertainty of the controlled system. A multi-modal density model with prediction error-dependent mixing coefficients is introduced to effect the mode switching. This approach can deal with arbitrary noise distributions, nonlinear plant dynamics and arbitrary mode switching. For the affine systems focussed upon for illustration in this paper the approach has global stability. The theoretical architecture constructs are verified by validation on a simulation example.

KEYWORDS:

• Multi-modal density model
• temporal multimodality
• switching control
• probabilistic multimodal adaptive control
• operator Riccati equation
• multiobjective probabilistic control

Disclosure statement

No potential conflict of interest was reported by the authors.