Context

The adoption of positioning systems to locate people, animals, and objects is beneficial to the development of many applications, including food tracking, geofencing, targeted advertising, traffic congestion management, and disease containment. In that, a significant interest has been triggered within the research community for the need of a navigation infrastructure able to trade-off between energy efficient communications (radio transceivers are fed by batteries) and precise positioning. In this general context, the most known and exploited method for geolocation lets a device collect the position and timing sent by some dedicated satellites in direct radio visibility, and then infer its relative position on ground. Roughly speaking, this is how GNSS-enabled devices can be tracked. However, continuously idle listening for GNSS signals would lead to a complete discharge of batteries in short time. This technological limitation sets up a constraint on the application scope. For instance, it would not be safe to track Alzheimer patients or pets that can get lost: once the battery is discharged, localizing them would not be possible anymore. In addition, having a ground-based localization system would increase the reliability of position-based applications when GNSS signals are temporary unavailable. Therein, Low Power Wide Area Networks (LPWANs) enable energy-efficient communications over long distances, thus resulting as a credible technological candidate for low power tracking systems. Based on cellular-type architectures, LPWANs aim to provide long range connectivity on the top of unlicensed spectra, thus easing their deployment at a global scope. Among LPWANs, Long Range (LoRa) networks targets several application needs, thus resulting as the most flexible in terms of offered communication modes. Nonetheless, to achieve a fair use of the unlicensed spectrum among concurrent operators and/or LPWAN technologies, the radio access is policed through duty cycle limitation to be hard-coded on end-devices. As matter of fact, LoRa is a novel technology, that has triggered research studies about network performance, and whose architectural features make it suitable for low power positioning systems. The objective of the internship is to setup a real deployment of LoRaWAN that is able to perform geolocalization.

Organization

The internship is organized as follows:

1. study of LoRa and related state of the art on performance evaluation (1 month);
2. study of geo-localization techniques for LoRaWAN (1 month);
3. getting familiar with real LoRa devices (1 month);
4. implementation of novel radiolocalization techniques (1 month);
5. performance evaluation through experiments (2 months).

Required skills

Excellent knowledge of communication networks (PHY and MAC layers), embedded systems, and programming languages (Python and C at least).

Location and contact

The internship will take place at LAAS-CNRS (Toulouse, France) within the SARA team (Services et Architectures pour les Réseaux Avancés) under the supervision of Nicola Accettura (nicola.accettura@laas.fr) and Pascal Berthou (pascal.berthou@laas.fr).