Abstract
In this paper we couple a deterministic planner with
an ontology, in order to adapt to new discoveries
during plan execution and to reason about the affordances
that are available to the planner as the set
of known objects is updated. This allows us to extend
the planning agent’s functionality during execution.
We use as an example planning for persistent
autonomous behaviour in underwater vehicles.
Planning in this scenario takes place in a symbolic
model of the environment, simulating sequences
of possible decisions. Ensuring that the simulation
remains robust requires careful matching of
the model to the real world, including dynamically
updating the model from continuous sensing actions.
We describe how our system constructs an
initial state for planning, using the ontology; how
the ontology is also used to determine the results
of each action performed by the planner; and fi-
nally demonstrate the performance of the system
in a simulation, in which two AUVs are required
to cooperate in an unknown environment, demonstrating
that with additional reasoning the planning
system is able to make new efficient choices, taking
advantage of the environment in new ways.
an ontology, in order to adapt to new discoveries
during plan execution and to reason about the affordances
that are available to the planner as the set
of known objects is updated. This allows us to extend
the planning agent’s functionality during execution.
We use as an example planning for persistent
autonomous behaviour in underwater vehicles.
Planning in this scenario takes place in a symbolic
model of the environment, simulating sequences
of possible decisions. Ensuring that the simulation
remains robust requires careful matching of
the model to the real world, including dynamically
updating the model from continuous sensing actions.
We describe how our system constructs an
initial state for planning, using the ontology; how
the ontology is also used to determine the results
of each action performed by the planner; and fi-
nally demonstrate the performance of the system
in a simulation, in which two AUVs are required
to cooperate in an unknown environment, demonstrating
that with additional reasoning the planning
system is able to make new efficient choices, taking
advantage of the environment in new ways.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 2nd ICAPS Workshop on Planning and Robotics (PlanRob-15) |
| Pages | 79-85 |
| Publication status | Published - 2015 |