Location estimation is a critical capability in modern life; it is used for navigation by both people and autonomous vehicles, for asset tracking, and for numerous other applications. GNSS systems like GPS and GLONASS provide this capability outdoors. However, in many settings, such as indoors, location estimation remains challenging and is still the subject of intensive research.
The tutorial will provide a broad technical overview to location estimation techniques. The tutorial will start from the basics, including the principles of parameter estimation, maximum likelihood, and least-squares minimization, and will show how to apply these principles to location estimation. We will also briefly describe many advanced techniques, like orthonormal and differencing elimination techniques for separable problems, uncertainty quantification, information-theoretic bounds, space-state (Kalman) filtering, and mixed integer estimation (e.g., RTK).
Most of these techniques are also useful in other estimation problems, but they are particularly easy to understand in the context of location estimation, because the constraints that relate the unknown parameters to the observations are geometric, so understanding them requires no specialized knowledge.
The presentation is designed particularly for computer scientists, who often lack relevant background in statistical signal processing, but electrical engineers will also find it a useful recap, especially if they have not done much work in parameter estimation.
In a half-day tutorial, there is not enough time to cover all the details. The intent is to give a big-picture overview of the statistical principles and results, the modeling techniques, and the algorithms, but without going into proofs or algorithmic details. Pointers for further study will be provided, for those planning to pursue the topic in more depth.
Sivan Toledo is a professor of Computer Science and is currently
serving as the head of Tel Aviv University’s Blavatnik School of Computer Science.
He received his PhD from MIT in 1995 and has been at Tel Aviv University since 1998.
From 2007 to 2009, he was associate visiting professor at MIT.
He has been working for the past decade on wildlife tracking and sensing systems,
and in particular on the ATLAS wildlife tracking system.
This work has led to high-impact publications,
including two papers in