Situation-aware assistance relies on computational models and sensor observations in order to reason about the user actions and needs. The computational models represent the possible human behaviour in a given situation. We use computational causal behaviour models (CCBM) to represent the human behaviour model. CCBM relies on planning operators to generate all possible variations in behaviour.

As manual modelling is time consuming and error prone process, our tool automatically generates the planning operators from textual instructions.

 

For more information on the procedure, have a look at the paper “Learning Models of Human Behaviour from Textual Instructions”, which can be downloaded here and the paper “TextToHBM: A Generalised Approach to Learning Models of Human Behaviour for Activity Recognition from Textual Instructions”, which can be downloaded here.