Analysis of human activities and identification of uncertain situations in context-aware systems

• Autor(a) principal: Freitas, Leandro O.
• Data de Publicação: 2020
• Outros Autores: Henriques, Pedro Rangel, Novais, Paulo
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/1822/69206
• topic: Attribute grammar; Context-Aware Systems; Quality of context; Uncertainty identification; Ciências Naturais::Ciências da Computação e da Informação

One of the main obstacles faced by context-aware systems developed for human activities recognition is related to dealing with incomplete data for decision making. Problems from several sources contribute to this phenomenon. Imprecise acquisition of data from sensors and the system’s design issues are among them. This paper presents an approach to tackle this trouble by dividing it into three parts. The first one refers to the validation of context data through an Attribute Grammar. The formalism and expressiveness provided a grammar, enriched with attribute evaluation rules and contextual constraints, ensures that the system will use only valid data for reasoning. Anomalous data will be detected, and the situation will be signalized. Also, the analysis of Quality of Context is provided, considering a set of characteristics, vouching that only useful information will be considered. At last, the identification of the sources of uncertain situations followed by a sequence of actions aiming to minimize the negative impacts of it helps the system to work with more complete sets of data. The formalization of the approach is provided together with an algorithm to validate it.