Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Nowadays we start noticing all the potential and amazing applications of virtual and augmented reality. The mass-market availability of cheap and powerful CPUs and GPUs has mostly solved the computational complexity issue of these applications. The displays and cameras, basic building blocks of many augmented reality applications, are also getting cheaper. But the environment sensing systems (radars, lidars, 3D scanners, head-trackers...) that are accurate and reliable enough to be used for critical applications are still rare and expensive. They might be affordable for the plane and helicopter manufacturers but it is a different story when it comes to small training simulators.
This Master Thesis tackles the issue of producing a cheap, accurate and robust head tracker with 6 degrees of freedom for a simulation and training setup. Our solution consists of an inertial and optical hybrid tracker using Inertial Measurement Units and cameras. Both are now cheaply available due to massive use in smartphones. The developed tracker will be operated in helicopters. So, in this report, we focus on the operational constraints of this environment. These constraints are mainly connected to latency, accuracy and robustness issues but also include compactness and discretion.
We successfully designed a cheap, reliable, highly modular and accurate head tracker with latency below 4ms. It is designed to work with helicopters pilots’ head-mounted displays in order to align the pilots eyes, a target and its corresponding “augmented” information. The trackers modularity allows a lot of flexibility on costs and accuracy so it will hopefully allow many other virtual reality applications to emerge.