continual coordination of shared activities brad clement & tony barrett artificial intelligence...
TRANSCRIPT
Continual Coordination of Shared Activities
Brad Clement & Tony Barrett
Artificial Intelligence Group
Jet Propulsion Laboratory
California Institute of Technology{bclement,barrett}@aig.jpl.nasa.gov
http://www-aig.jpl.nasa.gov/
Overview
• Why decentralized planning?• Shared Activity Coordination (ShAC)
– multiagent modeling– framework for developing roles and
protocols– testbed for evaluating protocols– continual coordination algorithm
• Mars 2003 coordination scenario
Why Decentralized Planning?
• Why plan?– near-term actions can effect subsequent ones in
achieving longer-term goals
• Why decentralize?– competing objectives (self-interest)– physically distributed points of control– limited shared resources– communication constraints/costs– computation constraints (parallel processing)– greater capabilities/efficiency of distributed assets
Motivation for NASAOver 40 multi-spacecraft missions proposed!
– Autonomous single spacecraft missions have not yet reached maturity.
– How can we cost-effectively manage multiple spacecraft?
Earth Observing System Sun-Earth Connections
Origins Program
Structure & Evolution of the Universe
Mars Network
NMP
NMP
• Optimize a function of variable assignments with both local and non-local constraints.
Distributed Constrained Optimization
ControlControl
ExecutiveExecutive
PlannerPlanner
AnalystAnalyst
ExecutiveExecutive
Planner
ExecutiveExecutive
Planner
ExecutiveExecutive
Planner
Shared Activity Coordination
Shared activities implement team plans, joint actions, and shared states/resources
Shared Activity Model
• identifier• parameters (string, integer, etc.)
– constraints (e.g. agent4 allows start_time [0,20], [40,50])
• decompositions (shared subplans)
• permissions - to modify parameters, move, add, delete, choose decomposition, constrain
• roles - maps each agent to a local activity
• protocols - defined for each role– change constraints– change roles
• changes sharing agents and protocol assignments
– handle changes received from other agents
ShaC AlgorithmGiven: a plan with multiple activities, including a set of
shared_activities with constraints, and a projection of plan into the future.
1. Revise projection using the currently perceived state and any newly added goal activities.
2. Alter plan and projection while honoring constraints.3. Release relevant near-term activities of plan to the real-time execution
system.4. For each shared activity in shared_activities
– if outside consensus window,• apply each associated protocol to modify the activity
– else• apply simple consensus_protocol
5. Communicate changes in shared_activities and constraints.6. Update shared_activities and constraints based on received
communications.7. Go to 1.
Default Protocol Capabilities
• joint intention
• mutual belief
• resource sharing
• active/passive roles
• master/slave roles
Extending Protocol Classes
1. modify permissions
2. modify local parameter constraints
3. add/delete sharing agents
4. change roles of sharing agents
Example Protocol Classes1. modify permissions2. modify local parameter constraints3. add/delete sharing agents4. change roles of sharing agents
• Argumentation (1,2)• Delegation (3)• Asynchronous weak commitment (1,2)• Constraint-based conflict resolution (2,4)• Round robin (1)• Centralized conflict delegator (extends
delegation)
Asynchronous Weak CommitmentModify permissions:
– if have highest priority• remove self’s modification permissions (add, move,
delete)– else
• give self modification permissionsModify parameter constraints:
– if cannot resolve local conflicts and conflicts with constraints of higher ranking agents
• set own rank to highest rank plus one• generate parameter constraints (no-good) describing
locally consistent values
Constraint-Based Conflict Resolution
Modify parameter constraints:– if cannot resolve conflicts involving shared activity
• update parameter constraints describing locally consistent values
Modify roles:– if reached consensus on constraints or
time_elapsed > threshold• switch to role for solution convergence
(e.g. argumentation, voting, highest rank decides)
Coordinating the Commanding of Mars ‘03
• Distributed Odyssey, MER A, and MER B ASPEN planners
• Schedules generated independently
• MERs share bandwidth and memory for downlinks from Odyssey
Mars 2003
Coordinating the Commanding of Mars ‘03
• Distributed Odyssey, MER A, and MER B ASPEN planners
• Schedules generated independently
• MERs share bandwidth and memory for downlinks from Odyssey
Mars 2003
Coordinating the Commanding of Mars ‘03
• Distributed Odyssey, MER A, and MER B ASPEN planners
• Schedules generated independently
• MERs share bandwidth and memory for downlinks from Odyssey
Mars 2003
Motivation
• Considerable ground operations effort and cost involved in coordinating mission plans for interacting missions.
• Human collaboration can be error-prone and slow to react.
• Automating this coordination reduces operations costs and increases science return.
• Applies to autonomous missions
Mars 2003 Scenario• MERs send critical data to
Earth and need uplink to direct subsequent activities
• Odyssey can often get data to/from Earth faster than MERs
• MERs’ mission planners coordinate with Odyssey’s to determine how and when data is routed
• Bandwidth and memory is shared on Odyssey, and decisions affect other local resources, such as power which can restrict other activities
Mars 2003
Mars 2003 Scenario
no pendingrequest
requestwait foruplink
criticalpancam
commearth
commodyssey
MER activitiesOdyssey activities
no pendingrequest
commearth
through Odyssey direct
must-be wait
wait foruplink
no pendingrequest
requestwait foruplink
criticalpancam
commearth
must-be wait
odysseyreceived
no pendingrequest
commearth
commodyssey
wait foruplink
downlinkcritical
data
uplinkfromDSN
Mars 2003 Scenario
no pendingrequest
Odyssey
MER A
must wait
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
requestwait foruplink
science activities
science activities
Mars 2003 Scenario
Odyssey
MER A
must wait
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
requestwait foruplink
science activities
science activities
must waitcritical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
requestwait foruplink
Mars 2003 Scenario
Odyssey
MER A
must wait
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
requestwait foruplink
science activities
science activities
must waitcritical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
request
comm earth
must wait
wait foruplink
wait foruplink
Mars 2003 Scenario
Odyssey
MER A
must wait
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
requestwait foruplink
science activities
science activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
request
comm earth
must wait
wait foruplink
odysseyreceived
must wait
wait foruplink
no pendingrequest
Mars 2003 Scenario
Odyssey
MER A
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
requestno pending
request
science activities
science activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
request
comm earth
must wait
wait foruplink
odysseyreceived
must wait
wait foruplink
odysseyreceived
no pendingrequest
no pendingrequest
Mars 2003 Scenario
Odyssey
MER A
comm earth
MER activitiesOdyssey activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
request
science activities
science activities
critical pancam comm earth comm earth
comm odyssey
traversecomm earth
no pendingrequest
request
comm earth
must wait
wait foruplink
odysseyreceived
must wait
wait foruplink
odysseyreceived
no pendingrequest
no pendingrequest
Summary
• Shared Activity Coordination (ShAC)– multiagent modeling– framework for developing roles and
protocols– continual coordination algorithm
• Mars 2003 coordination scenario
Future Directions
• Evaluate protocols in testbed• Quick group response to
anomalies and discovered opportunities
• Use summary information for abstract reasoning
• Apply to– Ground operations for
Techsat-21– Antenna array automation for
DSN– Distributed, autonomous
missions
Mars 2003
Related Work(partial list)
• Collaborative planning (D-SIPE)• Team activity modeling (TEAMCORE)
– TOP (Team Oriented Programming)
• Protocol Evaluation (TÆMS/GPGP)• Plan merging/coordination
– Georgeff, Ephrati & Rosenchein, Lansky– DPOCL– Summary information
• Joint Intentions– STEAM– Shared Plans
Organizations and Data Flow Derived from Concepts
ESOC/LOC
MMO NAV
MMO MPSET(SEQ/ASP)
MMO MMCT LMA RTO
MSSS-MOC
MER IST/SCT/SCI
ESA Antenna
LMA SCT
MERA MERBODYBeagle MGSMEX
MMO MPSET(PLANNING)
Uplink and Downlink Data
Ground Planning and Uplink Products Essential to Relay and DTE Communications
MMO MDAPT
DSN
waggoner 6/14/01