Answers are grounded in this paper and cite their sources.
Indoor localisation and asset tracking are central functions in industrial information systems, yet reliable positioning remains difficult in factories and warehouses where signal interference, non-line-of-sight propagation, dynamic layouts, and real-time constraints affect system performance. This systematic literature review examines multimodal sensor fusion architectures for indoor localisation and asset tracking, with emphasis on their relevance to industrial environments such as factories and warehouses. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, 61 peer-reviewed studies published between 2021 and 2026 were analysed, including 12 core industrial studies and 49 supporting studies from transferable indoor localisation contexts. Studies were identified through searches of ACM Digital Library, IEEE Xplore, Springer Link, Scopus, and ScienceDirect conducted in October 2026. The review synthesises evidence on radio-frequency identification (RFID), ultra-wideband (UWB), inertial measurement unit (IMU), light detection and ranging (LiDAR), visual, and hybrid sensing approaches, with particular attention to fusion architectures, information integration strategies, and algorithmic design. The findings show that fused systems generally outperform comparable standalone sensing approaches in representative studies under the evaluated experimental conditions, with reported accuracy improvements ranging from approximately 11.3% to 96% across the 61 included studies, and from 21.4% to 72.7% when restricted to the 12 core industrial studies, where baseline and fused-system results were directly comparable. This range is descriptive and condition-dependent rather than indicative of a central tendency. The strongest performance gains occur when complementary information sources are combined: radio-based systems provide absolute position cues, inertial and odometry-based methods provide continuous motion estimates, and vision or LiDAR provides spatial context. However, practical deployment remains limited by calibration effort, infrastructure cost, synchronisation complexity, map maintenance, scalability, privacy, and explainability. Future research should prioritise self-calibrating fusion architectures, edge-AI-based information processing, shared industrial benchmark datasets, explainable fusion models, and collaborative multi-asset tracking frameworks to support more reliable translation of indoor localisation methods into industrial asset-tracking applications.
No takes yet. Share an insight, caveat, or question.
Shoaib et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: