4.30pm 11 June 2021
The idea of a smart, autonomous future is tantalisingly close to becoming reality. In recent years, the push towards fully autonomous transport has driven large advances in enabling techniques such as machine learning, sensor fusion and the development of lower cost, higher fidelity sensors. To achieve full autonomy however, vehicles must become self-aware of their occupants, their immediate surroundings, and lastly, their location in the world at large – autonomous navigation.
With the Global Positioning System (GPS), today we can determine our absolute location on the planet to within 10 metres, with an update rate of 1 Hz to 10 Hz - an impressive feat. However, the limited accuracy and update rate has a significant impact when, for example, autonomously operating in dense, urban areas or tunnels where GPS signals may be compromised. For full autonomous navigation, inertial navigation systems are therefore required to fill in the gaps. Inertial navigation refers to methods that integrate onboard measurements of acceleration and rotation rate to determine the change in position and direction of a device. This is a challenging task as inertial navigation systems are extremely sensitive to biases and drift, leading the measurements astray. For longer term GPS denied navigation this necessitates the use of highly sensitive, well-calibrated accelerometers and gyroscopes.
In this talk, I will discuss inertial navigation, its current applications and the key associated challenges, with a focus on rotation sensing using optical gyroscopes. I will describe the limits of the current state-of-the-art and then give an overview of the work undertaken at the Australian National University to develop the next-generation of high-precision, low-cost optical gyroscopes for an autonomous future.