Terrain and Visual Perception for Navigation Without Satellite Signals: Performance Evaluation of Environment-Referenced Techniques
Terrain and Visual Perception for Navigation Without Satellite Signals: Performance Evaluation of Environment-Referenced Techniques
Amir Moghtadaei Rad1
1) Department of Computer and Electrical Engineering, Has.C., Islamic Azad University, Hashtgerd, Iran.
IOI Code
:
AAHD-ABKAZH
Publication :
National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
The growing vulnerability of Global Navigation Satellite Systems (GNSS) to electronic warfare has intensified the need for alternative positioning methods that exploit environmental features [1]. This paper presents a comprehensive performance evaluation of two prominent environment-referenced navigation techniques: Terrain-Aided Navigation (TAN) and Vision-Based Navigation (VBN) coupled with Simultaneous Localization and Mapping (SLAM). TAN estimates vehicle position by correlating real-time terrain measurements obtained from radar altimeters, lidar, or sonar with pre-existing digital elevation models (DEMs), providing bounded drift correction independent of radio-frequency signals [3,4]. Conversely, VBN and SLAM extract visual features from optical sensors to simultaneously localize the vehicle and construct a map of unknown environments, offering exceptional flexibility without requiring prior maps [5,6]. This study derives the mathematical foundations of both approaches, including correlation metrics (MAD, MSD, NCC) for TAN and nonlinear optimization techniques (Bundle Adjustment, iSAM2) for SLAM [13,15]. The analysis demonstrates that TAN provides reliable absolute fixes in structured, feature-rich terrains but fails in homogeneous regions, whereas VBN/SLAM excels in visually rich unknown environments but degrades under poor lighting, adverse weather, and low-texture scenes [9,10,18]. A detailed comparative assessment based on computational complexity, environmental dependency, and operational robustness is presented [14,17]. The paper introduces a covariance-based hybridization strategy with a novel temporal synchronization framework to dynamically switch between TAN and SLAM, ensuring resilient navigation across diverse operational scenarios. Results indicate that their synergistic integration, particularly when combined with inertial navigation, offers a robust solution for next-generation autonomous defense platforms operating in GNSS-denied environments [19,20]. This performance evaluation provides practical guidelines for selecting, integrating, and deploying terrain- and vision-referenced navigation technologies according to mission requirements, environmental conditions, and computational constraints, thereby supporting the development of resilient autonomous positioning systems for future aerospace and defense applications.
Keywords : terrain aided navigation (tan) vision based navigation (vbn) slam digital elevation model gnss denied