5G-DIL: Domain Incremental Learning with Similarity-Aware Sampling for Dynamic 5G Indoor Localization

TL;DR

This paper presents 5G-DIL, a domain incremental learning method for indoor localization that uses similarity-aware sampling to adapt quickly to environmental changes with minimal data, maintaining high accuracy.

Contribution

Introduces a novel similarity-aware sampling technique for efficient domain adaptation in 5G indoor localization, reducing training data and time.

Findings

Achieves MAE of 0.261 meters in dynamic environments

Requires as few as 50 exemplars for adaptation

Outperforms existing DIL techniques in real-world tests

Abstract

Indoor positioning based on 5G data has achieved high accuracy through the adoption of recent machine learning (ML) techniques. However, the performance of learning-based methods degrades significantly when environmental conditions change, thereby hindering their applicability to new scenarios. Acquiring new training data for each environmental change and fine-tuning ML models is both time-consuming and resource-intensive. This paper introduces a domain incremental learning (DIL) approach for dynamic 5G indoor localization, called 5G-DIL, enabling rapid adaptation to environmental changes. We present a novel similarity-aware sampling technique based on the Chebyshev distance, designed to efficiently select specific exemplars from the previous environment while training only on the modified regions of the new environment. This avoids the need to train on the entire region, significantly reducing the time and resources required for adaptation without compromising localization accuracy. This approach requires as few as 50 exemplars from adaptation domains, significantly reducing training time while maintaining high positioning accuracy in previous environments. Comparative evaluations against state-of-the-art DIL techniques on a challenging real-world indoor dataset demonstrate the effectiveness of the proposed sample selection method. Our approach is adaptable to real-world non-line-of-sight propagation scenarios and achieves an MAE positioning error of 0.261 meters, even under dynamic environmental conditions. Code: https://gitlab.cc-asp.fraunhofer.de/5g-pos/5g-dil