Towards a Commonsense Estimator for Activity Tracking

Catherine Tessier

Our purpose in this work is to track agents’ activities and to issue a report in terms of plans or procedures that are likely to be in progress, for applications such as surveillance, situation assessment, hazard detection. Starting from both Kalman filtering and Boutilier’s generalized update, a symbolic "commonsense" estimator has been designed and implemented. The relevance of prediction is shown for activity tracking and the symbolic estimator -- KALMANSYMBO -- is designed as a two-step (prediction and correction) iterative process that calculates a preference relation on a set of activities, given the successive observations.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.