Deep Learning Methods for Fingerprint-Based Indoor Positioning: A Review

Fahad Alhomayani,Mohammad H. Mahoor

DOI: https://doi.org/10.1080/17489725.2020.1817582

2022-05-30

Abstract:Outdoor positioning systems based on the Global Navigation Satellite System have several shortcomings that have deemed their use for indoor positioning impractical. Location fingerprinting, which utilizes machine learning, has emerged as a viable method and solution for indoor positioning due to its simple concept and accurate performance. In the past, shallow learning algorithms were traditionally used in location fingerprinting. Recently, the research community started utilizing deep learning methods for fingerprinting after witnessing the great success and superiority these methods have over traditional/shallow machine learning algorithms. This paper provides a comprehensive review of deep learning methods in indoor positioning. First, the advantages and disadvantages of various fingerprint types for indoor positioning are discussed. The solutions proposed in the literature are then analyzed, categorized, and compared against various performance evaluation metrics. Since data is key in fingerprinting, a detailed review of publicly available indoor positioning datasets is presented. While incorporating deep learning into fingerprinting has resulted in significant improvements, doing so, has also introduced new challenges. These challenges along with the common implementation pitfalls are discussed. Finally, the paper is concluded with some remarks as well as future research trends.

Machine Learning,Signal Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the accuracy and reliability of indoor positioning systems in complex environments. Specifically, the paper focuses on indoor positioning technology based on location fingerprints, especially using deep - learning methods to improve this technology. Traditionally, indoor positioning relies on satellite navigation systems (such as GPS), but these systems perform poorly in indoor environments because signals can be blocked or weakened by buildings. Therefore, researchers have begun to explore other solutions, among which location fingerprint technology has become a viable method due to its simplicity and accurate performance. Location fingerprint technology builds a database by collecting signal characteristics (such as WiFi signal strength, magnetic field strength, etc.) at specific locations, and then uses machine - learning algorithms to match the data collected in real - time with the data in the database, thereby determining the user's location.

The paper points out that although shallow - learning algorithms have been widely used for location fingerprint recognition in the past, in recent years, with the success of deep - learning methods in multiple fields, the research community has begun to turn to using deep - learning methods for fingerprint recognition. Deep - learning methods can automatically extract features, process large amounts of raw data, and perform well in handling high - dimensional and non - linear data. These characteristics make them particularly suitable for fingerprint recognition tasks in indoor positioning. However, applying deep - learning to fingerprint recognition also brings new challenges, such as the need for a large amount of training data and long model training time.

In summary, the main objective of this paper is to conduct a comprehensive review of the application of deep - learning methods in indoor positioning based on location fingerprints, discuss their advantages, challenges, and future research directions. This includes analyzing the advantages and disadvantages of different types of fingerprints (such as radio - frequency fingerprints, magnetic - field fingerprints, etc.), evaluating existing solutions, and proposing possible future research trends.