Relational object tracking and learning

We propose a relational model for online object tracking during human activities using the Distributional Clauses Particle Filter framework, which allows to encode commonsense world knowledge such as qualitative physical laws, object properties as well as relations between them. We tested the framework during a packaging activity where many objects are invisible for longer periods of time. In addition, we extended the framework to learn the parameters online and tested it in a tracking scenario involving objects connected by strings.

Authors

• Davide Nitti 0001
• Tinne De Laet
• Luc De Raedt

Keywords

• Random variables
• Robots
• Probabilistic logic
• Packaging
• Computational modeling
• Object tracking
• Human Activities
• Object Properties
• Particle Filter
• Physical Laws
• Track Model
• Learning Algorithms
• Nonlinear Model
• Measurement Model
• Body Condition
• Online Learning
• Continuous Distribution
• Transition Model
• Inference Procedure
• Object Distance
• Sufficient Statistics
• String Length
• Discrete Random Variable
• Related Languages
• State Transition Model
• Physics Engine
• First-order Logic
• Program Logic
• Probabilistic Programming
• Belief State
• Linear Model
• Object Boxes
• State Space
• Learning Strategies