FraseWare: Mecanismos de razonamiento evolutivo y personalizable con gestión de ambigüedad para entornos asistidos


The FRASEware subproject complements the FRASE coordinated project in the following aspects. Firstly, it provides a knowledge-driven approach for the diagnosis of frailty and senility. It proposes the creation of a knowledge base based on an ontology and a ruleset which operates over it, both with support to ambiguity. Such ambiguity comes, on one side, from the uncertainty associated to the values of the gait analysis brought forward by the sensing systems of the other two subprojects (based in accelerometry and vision, respectively) and, on the other hand, from the vagueness with which geriatricians can express the a priori correlation among the identified variables. The use of a knowledge-driven approach, as the one proposed, has clear advantages. The diagnosis of frailty and senility and their correlation can be determined from the very first moment, without training. Moreover, such diagnosis can be explained, i.e., the expert can revise the conditions under which the diagnosis has been produced. Secondly, FRASEware will personalize the diagnosis mechanism, moving it from a generic diagnosis system to another which is personalized to the subject. For that, it will permit to incorporate additional sensing measures coming from devices already available in the market (heartbeat monitor), together with adding new conditions generated by the correlation of gaits variable values associated to a subject. It will integrate this additional knowledge in the knowledge base to give place to a more precise and adjusted diagnosis to the individual particularities. Thirdly, the reasoning mechanism will evolve in time. It will adjust to the common aspects observed in the diagnosis of several individuals which may be generalized. As derived contribution, a knowledge base will be generated that gathers the historic measures and diagnosis bound to a group of individuals in time which can be exploited by third parties. Finally, FRASEware will link its customizable and evolving reasoning mechanisms with the multi-sensorial and classification mechanisms contributed by the other two subprojects through an architecture which follows the Mobile Cloud Computing paradigm. Such architecture will enable the integration of multi-sensing systems, both from this project and external ones, with intelligence techniques to facilitate the diagnosis of syndromes associated to ageing and react upon them. This intelligence layer will be executed, preferably, in the mobile (SmartPhones) or low-cost embedded (Rapsberry Pi) devices, deployed in the elderly home. However, it will use external resources, Cloud Computing, to accomplish more complex processes and thus avoid installing complex to maintain and costly computing infrastructure at elderly homes. FRASEware contributions may be extrapolated to other application scenarios within AAL.