1 / 9

Practical Planning for Robots and Other Autonomous Agents

Practical Planning for Robots and Other Autonomous Agents. Michael Brenner & Patrick Eyerich. Planning for Autonomous Agents. Autonomous agents do not only make plans, but also... execute and monitor them in dynamic environments with limited knowledge concurrently with others

kylee
Télécharger la présentation

Practical Planning for Robots and Other Autonomous Agents

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Practical Planning for Robots and Other Autonomous Agents Michael Brenner & Patrick Eyerich

  2. Planning for Autonomous Agents Autonomous agents do not only make plans, but also... • execute and monitor them • in dynamic environments • with limited knowledge • concurrently with others • in interaction with others

  3. Robots The prototypical autonomous agents! • real-time planning and acting • limited sensing • human-robot interaction We study planning for high-level robot control in several projects • Cosy, Desire, CogX

  4. Dora the Explorer

  5. What we're working on • Active Continual Planning: Intelligently interleaving planning, execution, and monitoring • Multiagent Planning: Reasoning about others • Dialogue Planning: Planning how to interact with others (in particular: humans) • Internal process planning: coordination of subarchitectures, demand-driven generation of planning states, planning for internal information gathering • Extended expressivity: action durations, cost measures, probabilistic outcomes, ...

  6. Approach Research framework: • Base planners: FF, TFD • Continual Collaborative Planning • MAPSIM Evaluation: • Simulation • Simulation + human participation • Real Human-Robot Interaction

  7. A MAPSIM Run MAPSIM run starts. There are 2 agents: Lilli and R2D2. • Lilli: ”Please give me the cookie, R2D2.” • R2D2: ”Okay, Lilli.” • R2D2: ”Where is the cookie, Lilli?” • Lilli: ”The cookie is in the kitchen, R2D2.” • R2D2: ”Thanks.” • R2D2: ”Please open the kitchen door, Lilli.” • Lilli opens the kitchen door. • R2D2: ”Thanks.” • R2D2 moves to the kitchen. • R2D2 grasps the cookie. • R2D2 moves to the living room. • R2D2 gives Lilli the cookie. • Lilli: ”Thanks for giving me the cookie, R2D2.” MAPSIM terminates successfully.

  8. And a fun side project... This is a story about Prince Valiant, King Arthur and a dragon. King Arthur traveled to the castle. Prince Valiant saw King Arthur. King Arthur said: "Please bring me the treasure, Prince Valiant!". Prince Valiant said: "As you wish, King Arthur." Prince Valiant asked: "Where is the treasure, King Arthur?" King Arthur said: "The treasure is in the dragon's lair." Prince Valiant said: "Thank you." Prince Valiant traveled to the dragon's lair. Prince Valiant saw the dragon. The dragon tried to kill Prince Valiant, but failed. Prince Valiant killed the dragon. Prince Valiant took the treasure. Prince Valiant traveled to the castle. Prince Valiant gave the treasure to King Arthur. King Arthur said: "Thank you for bringing me the treasure, Prince Valiant."

  9. Interested? Many projects possible: • practical extensions to the TFD planner • interaction planning • active plan monitoring • story generation • ... Contact us: {brenner,eyerich}@informatik.uni-freiburg.de

More Related