Machine Learning for Localization and Navigation

Description of the special session:

This special session aims to explore the cutting-edge applications of machine learning (ML) and deep learning (DL) techniques in indoor localization and navigation, emphasizing the unique challenges and opportunities in this domain. We invite research works that employ novel data-driven methodologies for localization, navigation, and tracking using various sensors and hybrid systems and that integrate ML/DL with traditional tracking filters. Additionally, the session welcomes studies along the processing pipeline of AI, e.g., adapting pre-trained ML models to new environments, movement types, and sensor systems. 

Novelty and motivation:

The integration of ML/DL into indoor localization systems represents a significant advancement, promising to overcome longstanding challenges such as non-line-of-sight issues, device heterogeneity, and environmental variability. For the IPIN community, this topic is of high relevance as it leverages the latest in AI to enhance localization accuracy and reliability in complex indoor environments. The proposed special session is crucial as it aims to highlight innovative AI-assisted approaches and practical solutions that push the boundaries of current indoor positioning technologies, fostering new research directions and applications that are pivotal for both academic and industry advancements. 

This special session differentiates itself from regular sessions by focusing exclusively on the innovative applications of AI in indoor localization, targeting specific challenges and opportunities within this niche. Unlike broader sessions, it emphasizes novel data-driven methodologies, hybrid systems integrating ML/DL with conventional tracking filters, and the adaptation of pre-trained models to diverse and dynamic indoor environments. This targeted approach fosters deeper, more focused discussions and attracts specialized contributions that push the frontier of indoor positioning technology, providing the IPIN community with cutting-edge insights and fostering collaboration among experts in these advanced fields.

Scope of the session:

In this Special Session, we invite authors to submit papers related (but not limited) to:

  • Advanced ML/DL techniques for indoor positioning, e.g., fingerprinting, self-supervised learning, state-space-models, large language models
  • Sensor integration and fusion, e.g., RF, IMUs, cameras, and hybrid systems that combine ML/DL with conventional tracking filters such as KF
  • Identification and mitigating Non-Line-of-Sight (NLOS) issues, e.g., out-of-distribution detection, TOA / AOA channel parameter estimation
  • Addressing device heterogeneity and environmental variations, e.g., discrepancies across different devices and changing environments and adaptation of pre-trained models to new devices and varying indoor settings.
  • Accounting for diverse movement types, e.g., movement patterns and velocities
  • Adapting pre-trained models, e.g., to new and unexplored conditions and transfer learning approaches for improving localization performance across different scenarios


  • Supervised, semi-supervised, unsupervised, self-supervised, and reinforcement learning schemes
  • From SVM and K-NN over CNN and RNN to Attention, Transformer, and GPT
  • UWB, 5G, GNSS, Wi-Fi, BLE, Light, LIDAR, Sonic, IMU, Barometer, …
  • Indoor and outdoor localization, navigation, and tracking; (pedestrian) dead reckoning (PDR), parameter estimation, SLAM, signal processing, filters, anomaly detection / mitigation … 

Session chairs:

  • Dr. Tobias Feigl (Fraunhofer IIS)
  • Dr. Christopher Mutschler (Fraunhofer IIS)

Organizers (Technical Program Committee):

  • Dr. Felix Ott (Fraunhofer IIS)
  • Maximilian Stahlke (Fraunhofer IIS)
  • Jonathan Ott (Fraunhofer IIS)