Computational Perceptions of uninterpretable data. A case study on the linguistic modeling of human gait as a quasi-periodic phenomenon

As a contribution to the field of the Zadehʹs Computational Theory of Perceptions, this paper deals with generating linguistic models of those phenomena that the designer cannot completely recognize at running time. We explore the possibility of generating linguistic descriptions including in their meaning the fact the complete interpretation of input data is not feasible. With this purpose, we extend our previous research in this field by introducing the concept of Computational Perception of uninterpretable data. This extension provides more expressiveness and flexibility to the currently available resources. We define a linguistic variable that explains the degree with which phenomena suit the model, i.e., if input signals can be interpreted. In order to demonstrate the advantages of this contribution, we apply it to create a definition of fuzzy set of quasi-periodic phenomena. We use these new linguistic models to analyze and linguistically describe some characteristics of the human gait, as a case study of quasi-periodic phenomenon.