PAARL: An Interactive System for Policy-Aware Planning in Autonomous Agents(…)
summary
PAARL implements an interactive AAA loop using Answer Set Programming with Clingo and a Python controller plus GUI to plan under authorization/obligation policies encoded in AOPL. The prototype plans a self‑driving car from location 5 to 10, quantifies violations (speeding incurred 3‑point penalties per event, total 18 penalty points), and replans when a school bus observation at time step 2 forces a stop, raising completion time 35 -> 37.
AI-written summary · about
chapters (12)
- 0:00PAARL system intro
- 1:03Autonomous agents basics
- 2:10Planning and ASP basics
- 4:59Policies and compliance
- 6:47Penalties in violations
- 9:01PAARL design and contributions
- 12:09System architecture
- 13:14Agent behavior differences
- 15:42Dynamic domain injection
- 18:23Live PAARL demo
- 21:36Key takeaways and future
- 23:01Q&A
related talks
14mTowards Unobtrusive Physical AI: Augmenting Everyday Objects with Intelligence and Robotic Moveme...
29mMulti-Agent Pick and Delivery with Capacities: Action Planning vs Path Finding
59mOn Human-robot Interactions via Multimodal Intent Modeling and Reachability-based Safety Assurance
01h 24mBuilding Multiagent Systems: An Ontology-Driven Approach with SPADE
18mRobotics: why now?
29mEngineering Intelligence through Agent Interaction
29mEfficient Path Planning for Search and Rescue Missions Using Neural Networks
21mDistributed Reinforcement Learning for Robotic Assembly