MSc thesis project proposal

Localization using Time-of-Arrival Estimation in the LoRa Network

Project outside the university

KPN (The Hague)
LoRa (Long Range Low Power) is a radio protocol which is recently nationwide introduced by KPN for M2M (machine-to-machine) applications. LoRa uses a modulation based on spread spectrum technology that allows to demodulate signals below the noise level (i.e., 20 dB below the noise). The chip rate is 125 kchip/s and is equal to the bandwidth of the LoRa signal, namely 125kHz. LoRa offers different data rates ranging from 0.3 to 5.4 kbps using different spreading factors. The low power, low bit rate and intended use distinguish this type of network from a wireless WAN that is designed to connect users or businesses, and carry more data while using more power.

Positioning/localization is becoming more important and attractive, especially within wireless ad hoc networks or area dependent services. Having knowledge of the users' locations in a specific network allows for example the network designer and/or client to locate the nearest available link, greatly improving the Quality of Service (or optimizing logistics). The LoRa protocol gives the opportunity to localize devices using triangulation if the device is seen by at least three gateways (Base-stations). Different methods exist for positioning; however LoRa uses the Time Difference of Arrival (TDoA) technique where the synchronization between the gateways is done by GPS.

Assignment

The most important parameter for positioning is how to achieve a good (acceptable) time-of-arrival accuracy and especially in multipath environments. Many factors can cause errors in range/time estimation (e.g. hardware limitations, noise, interference, NLOS propagation, etc.). Therefore, the main objective of this MSc thesis project is to analyze/determine/improve the timestamping accuracy provided by LoRa since KPN is now looking into this feature to launch it commercially. To this end, the tasks of the student include:
  • Investigate possible concepts and algorithms to improve the timestamping accuracy taking into account complexity.
  • Build a simulation of the LoRa system as actually implemented for localization (focus only on localization part) including multipath effect.
  • Validate proposed concepts and algorithms by means of real measurement data. Please note, measurement data is available or will be provided.

Phases of the thesis project (approx 36 weeks)

  1. Problem formulation and literature study (±2 week).
  2. Investigate different concepts and algorithms in order to improve timestamping accuracy (±20 weeks).
  3. Build a simulation of the localization part: (±6 weeks).
  4. Apply proposed algorithm(s) to real data (±4 weeks).
  5. Thesis report writing and preparation presentation (±4 week).

The project is carried out partly at TU Delft, partly at KPN premises. Biweekly progress meetings with KPN are anticipated.

Requirements

Knowledge of ET4147 Signal Processing for Communication; experience with matlab.

Contact

prof.dr.ir. Alle-Jan van der Veen

Signal Processing Systems Group

Department of Microelectronics

Last modified: 2019-04-24