Situation Awareness for Autonomous Agents Under Limited Sensing

Welcome to the half-time seminar for José Manuel Gaspar Sánchez!

Abstract:   Autonomous agents need to understand their environment and predict its future evolution based on the information collected by their sensors, i.e. they  need situation awareness. This awareness is used to make predictions about the future and plan their own actions accordingly.

Limited sensor range and occlusions constrain the portion of the environment that can be observed at once. Additionally, sensors suffer from noise and errors, which limits the confidence in their measurements.

This thesis focuses on improving the situation awareness of autonomous agents by explicitly considering sensor limitations and leveraging information collected over time using knowledge about the possible dynamics of the environment.

The main criteria used to evaluate the results from this work are the safety and performance of the autonomous agents, and the computational complexity of the algorithms.

Respondent: José Manuel Gaspar Sánchez
Discussion leader:  Professor Emilio Frazzoli, ETH and MIT
Supervisors: Martin Törngren and Jana Tumova
Time: 13:30 – 16.00, Tuesday 6 December 2022
Place: In real life at KTH (E32, Lindstedtsvägen 3) and via Zoom

Half-time seminar material:

• Half-time thesis overall summary
• Publication 1
• Publication 2