Susanne Biundo, Karen Myers, Kanna Rajan How is Planning & Scheduling Changing the World?

Susanne Biundo, Karen Myers, Kanna Rajan

How is Planning & Scheduling Changing the World?

Susanne Biundo, Karen Myers, Kanna Rajan

A brief word on the State of the PracticeAutominder: An Intelligent Cognitive Orthotic System

Client Client ModelerModeler

Plan Plan ManagerManager

Intelligent Intelligent Reminder Reminder GeneratorGenerator

ClientPlan

Activity Info

Inferred Activity

Sensor Data

Reminders

Client Model Info

Activity Info

Preferences

Plan Updates

ClientModel

PI: Martha Pollack

Pléiades:Earth Observation Satellite Management

Daily automatic selection and scheduling of complex observations: mono or stereo, ground targets or areas.

Fair sharing of the use of the satellite

between several owner entities

Jean Michel Lachiver, CNESMichel Lemaître, Gérard Verfaillie, ONERA

Toulouse, France

The AMC Allocator:Advanced Scheduling for US Airforce

•Problem: Day-to-day allocation of aircraft & crews to airlift/tanker missions•Characteristics– Large scale: 1,000s of missions; 100s of assets– Continuous, dynamic stream of mission

requirements

•Core Technology: Incremental, constraint-based search– Rapid gen. of airlift/tanker schedules– Localized revision in response to changing circumstances– Flexible, what-if option generation

•Status: Embedded in AMC’s operational planning system & transitioning into use

DARPADARPACarnegie MellonPI: Steve Smith

Susanne Biundo, Karen Myers, Kanna Rajan

ICAPS 2005

http://siadex.ugr.eshttp://siadex.ugr.es

Andalusian Regional Ministry Of The

Enviornment(SPAIN)

University of Granada

SEPIAPlanning Group

Distributed execution over the internet

PI: Luis Castillo

The EO-1 Autonomous Sciencecraft

• Onboard planning part of autonomy software flying onboard Earth Observing One Spacecraft Fall 2003 – present

• Planning software onboard enables spacecraft to autonomously monitor and retarget volcanoes, flooding, cryosphere

PI: Steve Chien

• ASTEP LITA Atacama Field Campaign (Sep-Oct 2004)– Zöe rover with life detecting instruments– On-board planning and autonomous navigation over long

distances

• Rover executive results (preliminary, telemetry still being analyzed)– Total hours of operations (cumulative over several runs):

17 hours– Total distance covered: 16 km– Longest autonomous traverse: 3.3Km 2h 29m– “Roughest traverse”: 1h 2m with 19 faults recovered– Faults addressed:

• Navigator “confused”• Internal processes failed• Early and late arrival at waypoint

Robust Task Execution for Rovers: LITA

Courtesy: Nicola Muscettola

IDEA PI: Nicola Muscettola

Susanne Biundo, Karen Myers, Kanna Rajan

Mexar

The problem of spacecraft memory dumping• Domain

– ESA Mars Express mission

• Problem components– Finite memory banks

– Limited downlink windows

– Limited data rate

Planning and Scheduling Team, CNR - Italy – http://pst.istc.cnr.it

Input-solverInterface

Solver-outputInterface

Solver

PI: Amedeo Cesta

Commanding Spirit & Opportunity with MAPGEN

• Mixed-Initiative ground-based Activity Planning Decision Support system– Generative planning– Plan editing– Constraint formulation and

moves – Deals with time and resources

• First AI based system to command a vehicle on the surface of another planet

• ROI for NASA > 20% for science return in comparision to a manual planning process

PI: Kanna Rajan