China
Africa
Explore chapters and articles related to this topic
Pervasive Computing for Home Automation and Telecare
Published in Syed Ijlal Ali Shah, Mohammad Ilyas, Hussein T. Mouftah, Pervasive Communications Handbook, 2017
Claire Maternaghan, Kenneth J. Turner
Goals are high-level objectives for how a system should behave. They are often broken down into sub-goals and ultimately into concrete actions to achieve goals. Kaos (“Knowledge Acquisition in Automated Specification” [14]) is an example of goal refinement in requirements engineering. Other approaches to goal refinement such as [15] are typically based on logic. In the context of policy-based management, the idea is to refine goals into policies that can realize them (or alternatively to combine policies that achieve goals). As an example, Bandara et al. [16] use event calculus for formal refinement of goals into policies. Ache [17] is specifically for goals in a home context. However, the kinds of goals supported by Ache are restricted to comfort and cost, and the approach emphasizes how the system can learn to meet these goals.
Using Constraint Technology for Predictive Control of Urban Traffic Based on Qualitative and Temporal Reasoning
Published in Takushi Tanaka, Setsuo Ohsuga, Moonis Ali, Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, 2022
F. Toledo, S. Moreno, E. Bonet, G. Martin
The reasoning about traffic problems can be performed by means of a spatio-temporal analysis of the database generated by the qualitative model. We use for this task the General Representation Formalism (GRF) [Equ91] developed in the EQUATOR project. It is based on the Event Calculus of Kowalski [KS86], a formalism for reasoning about time using the Horn Clauses subset of the first order logic augmented with negation as failure. The GRF provides several extensions with respect to the Event Calculus (i.e. temporal granularity and continuous change). Due to the detected insufficiencies of Prolog, the GRF implementation has consisted on developing a proof procedure in the Constraint Logic Programming (CLP) paradigm [Ba+94].
Data Fusion for Telemonitoring Application to Health and Autonomy
Published in Hassen Fourati, Krzysztof Iniewski, Multisensor Data Fusion, 2016
Céline Franco, Nicolas Vuillerme, Bruno Diot, Jacques Demongeot, Anthony Fleury
To estimate new features, two approaches may be distinguished: knowledge-driven and data-driven techniques. In the first case, an expert defines logical rules between potential events from a priori knowledge on which the decision is based. This approach includes event calculus, temporal relation among events, fuzzy logic, and so forth. Although a logical approach enjoys the comfort and the rigor of a formal framework that is efficient on simulated data, it encounters difficulties while treating noisy and uncertain data such as real ones. Therefore the most well-spread techniques come from machine learning and rely on probabilities (Table 30.2).
Designing a system to extract and interpret timed causal sentences in medical reports
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2019
C. Puente, A. Sobrino, J. A. Olivas, A. Villa-Monte
The first one historically includes approximations such as the situation calculus, the fluent calculus or the event calculus, and aims to study and formalise the changes in the world due to actions or events that occur one after another. The second one deals with reasoning under temporal constraints and approaches the relationships between temporal entities (points or intervals) for planning, scheduling or NLP. Typical oncoming are those of qualitative constraint representation, representation based on dating schemas or duration-based approach. Allen’s interval relationship is a good example of this last taxon.