Files
Raviart_52492100_2026.pdf
Open access - Adobe PDF
- 2.63 MB
Details
- Supervisors
- Faculty
- Degree label
- Abstract
- During their interventions, firefighters are forced to evolve in dangerous and unpredictable conditions created by the environment of buildings in fire. Due to the threats of fire, thick smoke, and building collapses, there is a risk that they become disoriented, injured, or trapped inside the burning building. Being able to quickly and accurately localised the firefighters in trouble is essential to efficiently bring them assistance, which can make the difference between life and death in conditions of extreme heat and low oxygen. The objective of this master thesis is to design an indoor positioning system that is suited for the specific use case of firefighter interventions and simulate it to evaluate its performance. The propagation of electromagnetic waves in typical firefighter scenarios is modelled and simulated through ray-tracing to evaluate localisation impairments such as multipath that indoor environments are known to cause. To provide a realistic evaluation of the performance that can be achieved in those scenarios, a complete simulation of the system is implemented to consider the physical propagation and communication of wireless signals which are then used by positioning algorithms. A Multiple Signal Classification (MUSIC) algorithm is used to jointly estimate the time of arrival (ToA) and angle of arrival (AoA) of propagation paths, and the performance of traditional localisation algorithms in the firefighter scenario simulated are evaluated. A Random Sample Consensus (RANSAC) algorithm for a hybrid localisation scheme using TDoAs and AoAs is proposed to discard multipath components that do not correspond to a line-of-sight (Los), and the use of reflected paths to improve localisation accuracy is explored. Their performances are evaluated by complete simulations in the modelled firefighter environment.