Continual Planning in Golog

Abstract

To solve ever more complex and longer tasks, mobile robots need to generate more elaborate plans and must handle dynamic environments and incomplete knowledge. We address this challenge by integrating two seemingly different approaches — PDDL-based planning for efficient plan generation and Golog for highly expressive behavior specification — in a coherent framework that supports continual planning. The latter allows to interleave plan generation and execution through assertions, which are placeholder actions that are dynamically expanded into conditional sub-plans (using classical planners) once a replanning condition is satisfied. We formalize and implement continual planning in Golog which was so far only supported in PDDL-based systems. This enables combining the execution of generated plans with regular Golog programs and execution monitoring. Experiments on autonomous mobile robots show that the approach supports expressive behavior specification combined with efficient sub-plan generation to handle dynamic environments and incomplete knowledge in a unified way.

Situation calculus for controller synthesis in manufacturing systems with first-order state representation

“…Our framework uses Situation Calculus action theories for capturing actions in manufacturing processes, and high-level ConGolog programs over such action theories for capturing the processes defined over these actions. This makes a whole body of related work readily available to address a number of problems arising in the context of manufacturing systems [46,47,48,49,50,51]. In addition, we can leverage the first-order state representations of action formalisms and the second-order/fixpoint characterization of state-change provided by programs, giving a formal and declarative representation of the MaaS manufacturing setting.…”

Goal Reasoning in the CLIPS Executive for Integrated Planning and Execution

“…GOLOG has been extended for interleaved concurrency (De Giacomo, Lespérance, and Levesque 2000), on-line execution (De Giacomo, Lespérance, and Levesque 2000), and execution monitoring (De Giacomo, Reiter, and Soutchanski 1998). GOLOG can also use PDDL for continual planning (Hofmann et al 2016), which interleaves PDDL-based planning with GOLOG plan execution, plans for acquiring missing knowledge, and monitors the environment for unexpected events and exogenous actions.…”

Initial Results on Generating Macro Actions from a Plan Database for Planning on Autonomous Mobile Robots

“…Long planning times (at run-time) are a major issue for many real-time and on-line scenarios. Especially continual planning (Brenner and Nebel 2009;Hofmann et al 2016), where re-planning occurs frequently when new observations become available or sub-plans are expanded, can be sensitive to this issue -but it provides important features such as robustness and overall efficiency by dividing the overall planning problems in smaller chunks. A key observation in many robotic domains is that robots repeatedly execute similar plans.…”