Izvorni znanstveni članak
https://doi.org/10.7906/indecs.24.6.1
Explainable Machine Learning for Patient Satisfaction Prediction in Healthcare: A Regression and Classification Study
Jelena Pisarov
; University of Novi Sad – Faculty of Sciences, Department of Physics, Novi Sad, Serbia
*
Gyula Mester
; University of Szeged – Faculty of Engineering, Szeged, Hungary
* Dopisni autor.
Sažetak
Autonomous vehicles represent a transformative paradigm shift in global mobility, aiming to drastically reduce human-error-related fatalities while enhancing transport efficiency. The primary aim of this research is to evaluate the “Perception-to-Decision” pipeline, specifically focusing on how Artificial Intelligence (AI) integrates multi-modal sensory data to create a resilient cognitive framework for Level 5 autonomy. The methodology employed involves a comprehensive analysis of sensor fusion architectures, juxtaposing high-resolution visual data from Cameras with precise 3D point clouds from LiDAR and weather-resilient telemetry from Radar systems. Furthermore, the study examines the implementation of Deep Learning models for real-time object detection and Rein-forcement Learning for dynamic trajectory planning within non-deterministic urban environments. The results of this study demonstrate that while individual sensors possess inherent physical limita-tions – such as camera sensitivity to glare or LiDAR’s degradation in heavy precipitation – AI-driven sensor fusion significantly mitigates these vulnerabilities by cross-referencing data streams to reduce environmental uncertainty. Data synthesis indicates that hybrid AI architectures, combined with edge computing, enhance the reliability of object recognition and behaviour prediction. In conclusion, the research finds that the synergy between robust multi-modal perception and explainable AI (XAI) is the fundamental cornerstone for safe autonomous deployment. The study suggests that future scala-bility depends on the integration of V2X (Vehicle-to-Everything) cooperative autonomy, which pro-vides the “extended eye” necessary for navigating complex edge cases, thereby establishing a roadmap for the next generation of intelligent, trustworthy transportation systems.
Ključne riječi
autonomous vehicles; multi-sensor fusion; deep learning algorithms; explainable AI; V2X communication
Hrčak ID:
350254
URI
Datum izdavanja:
30.12.2026.
Posjeta: 0 *