On the visual systems & control on matrix space

This paper proposes a new modeling and control of dynamical systems on matrix space. By extending our previous approach on polynomial space (Sakai and Sato, 2014), and also replacing the POD basis by the standard basis (pixel by pixel basis) on matrix space, high resolution modeling and control of dynamical systems without any geometric feature extraction are implementable on the familiar control hardwares. Roughly speaking, our approach is a choice of the domain and co-domain (a choice of the input and output spaces) instead of thinking new mappings (new controllers). The proposed method is demonstrated by a liquid sloshing application.