Πλοήγηση ανά Συγγραφέας "Tsamis, Georgios"

Τώρα δείχνει 1 - 1 of 1
  • Μικρογραφία εικόνας
    Τεκμήριο
    The design, architecture, implementation and artificial intelligence support of a smart bio and cultural guiding system
    (ΕΛΜΕΠΑ, Πολυτεχνική Σχολή, Τμήμα Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών, 2026-04-22) Tsamis, Georgios; Τσάμης, Γεώργιος; Papadakis, Nikolaos; Παπαδάκης, Νικόλαος
    This study presents the design, architecture, implementation, and evaluation of a smart bio- and cultural-guiding system that integrates artificial intelligence–driven personalization with augmented reality (AR) guidance on commodity Android devices. Unlike conventional mobile guiding applications that rely on static proximity-based listings or rule-based user profiles, the proposed platform introduces a neural-network-based recommender and an AR-assisted navigation interface, aiming to measurably improve recommendation relevance, spatial understanding, and overall user experience without requiring human guide intervention. The novelty of the proposed approach is demonstrated through a rigorously separated dual-method evaluation framework combining synthetic-data-based system analysis with controlled observational usability studies. Quantitative results show that the AI recommender significantly outperforms proximity-based and rule-based baselines, achieving improvements of up to +19 percentage points in Precision@5 (0.41 vs. 0.22) and +21 points in NDCG@5 (0.45 vs. 0.24), confirming superior ranking quality for personalized points of interest. User classification accuracy is also enhanced through a multi-layer neural network, reaching 85% accuracy compared to 78% for a K-Nearest-Neighbors baseline. Furthermore, controlled usability experiments demonstrate that AR-based guidance yields substantial performance gains in real-world navigation tasks. In spatial POI location tasks, the proposed system increases task success rates from 78% to 83% while reducing median completion time by 38% (11 s to 6.8 s), indicating a marked reduction in cognitive load when translating digital guidance into physical navigation. Overall usability is rated as excellent, with a System Usability Scale (SUS) score of 82.5, compared to 68.0 for a non-AR baseline application. Despite incorporating computationally intensive AR and AI components, system profiling confirms practical deploy ability on mid-range Android devices, with average CPU utilization remaining below 25% in standard interaction scenarios and peak usage limited to short-lived AR and vision-related operations. Network latency for core services consistently remains within sub-second thresholds, supporting responsive user interaction even in bandwidth-constrained environments. In practical terms, the proposed system provides a scalable technological foundation for sustainable bio- and cultural-tourism applications, enabling data-driven tourism management, enhanced visitor education, and improved engagement with natural and cultural heritage sites. By empirically demonstrating that AI personalization and AR guidance deliver measurable benefits in recommendation quality, navigation efficiency, and usability, this work contributes validated evidence that intelligent, augmented mobile guiding systems can effectively replace traditional human-guided experiences while preserving ecological and cultural sensitivity.