Volume 3 Issue 3
Authors: Cheyun Xia; Yuan Li; Wai-Choong Wong; Lei Wang
Abstract: With the proliferation of location based services (LBS), various indoor localization systems have been proposed based on received signal strength (RSS). Many existing infrastructures of wireless local area networks (WLANs) have been deployed for widespread communication coverage. Hence, one mobile device may receive signals from only one or two official access points (APs), which renders the conventional localization systems impractical. However, many unknown wireless APs are often perceivable and can be utilized by the RSS fingerprint approach, which suffers from tremendous training costs and device diversity. With this motivation, this paper proposes a robust and cost-effective localization system to mitigate the effects of device diversity as well as reduce the training costs by employing two algorithms: a power-gap elimination algorithm and an unsupervised training algorithm. Simulation and experimental results demonstrate that the mean error of the proposed localization system is approximately five meters under various conditions, and the mean error of using supervised RSS fingerprints is 3.2 meters.
Keywords: Robust Localization; Unsupervised; Fingerprinting; Calibration-free; Device Diversity
Authors: Sergey Sakulin; Alexander Alfimtsev
Abstract: Information retrieval based on weighted zone scoring means the assignment weight for each zone or each field in the document metadata. All these weights are obtained using machine learning methods. The paper presents a method of determining the weights using the fuzzy Choquet integral. This allows taking into possible account interdependence between the zone parameters when calculating the relevance and allows to obtain higher scoring accuracy.
Keywords: Information Retrieval; Aggregation Operator; Fuzzy Measure; Choquet Fuzzy Integral
Authors: Ana Lilia Laureano-Cruces; Javier Ramírez-Rodríguez; Lourdes Sánchez-Guerrero; Martha Mora-Torres
Abstract: One of the most significant problems in artificial intelligence is knowledge representation linked to the decision-making process in order to simultaneously consider a set of events to achieve the combination that allows the trigger of an action. This work uses a parallel and distributed design approach to represent knowledge, taking a decision-making process during a risk event in an applied engineering process as a case study, which also includes the uncertainty that underlies the process. This allows us to consider the advantages of this kind of knowledge representation. The design is based on innovative fuzzy cognitive maps and their ability to simultaneously consider the causality of all elements that comprise the behavior to be modeled. The approach used by the cognitive model includes: 1) event process; and 2) behavior of the expert in the case study. The analysis utilizes mental models, genetic graphs, and behavioral analysis of the process to identify elements, their causal relationships, and their relative weights.
Keywords: Knowledge Representation; Initial Scenario; Future Scenario; Fuzzy Cognitive Maps; Decision-Making Process; Reactive Behaviors