@phdthesis{discovery10197711,
          school = {UCL (University College London)},
            note = {Copyright {\copyright} The Author 2024. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/).  Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms.  Access may initially be restricted at the author's request.},
           pages = {1--312},
          editor = {Benjamin Heydecker},
            year = {2024},
           title = {Location and speed estimation for telematic signalling in
railways},
           month = {September},
        keywords = {Railway, sensor fusion, ERTMS, CBTC, Kalman filter},
          author = {Sengupta, Monish},
             url = {https://discovery.ucl.ac.uk/id/eprint/10197711/},
        abstract = {This study considers the relationship between railway capacity and signalling
systems. To increase railway capacity, signalling has been increased from two
to four aspects, which can accommodate a maximum of 35 trains per hour.
Although aspect signalling relies on fixed block control, the introduction of
a moving block signalling system could achieve a capacity of 44 trains per
hour. Achieving this would involve the introduction of telematic signalling
systems (ERTMS and CBTC), which depend on the continuous and accurate
measurement of the train location and speed. The current stopping accuracy
required for a train is {$\pm$} 1m on Britain's mainline railway and {$\pm$} 0.5m for London
Underground. This is attributed mostly to the sensor fusion framework
that is currently in place to obtain accurate train location and speed. In this
thesis, a framework is developed based on the Kalman filter. This research
shows that a linear Kalman filter fusing data from railway sensors such as
Doppler radar, tachometers and wayside balise can provide good estimates of
train location and speed. However, an unscented Kalman filter is capable of
achieving more than four times better accuracy. The performance of these
filters is investigated in the presence of wheel slip and slide, and instances of
missing balise, leading to reduced capacity. In order to counter this, two adaptive
sensor fusion frameworks are developed based on Magill's filter bank and
innovation tracking. Both are applied to linear and unscented Kalman filters
which shows improved state estimation and stopping accuracy, hence capacity.
Analysis shows that balise placement on the approach to the stopping point is
important. Numerical results show that interception of a single balise reduces
the location error well within the required range. Therefore, a maximum of
two balises will serve the purpose with one as a backup.}
}