180 likes | 363 Views
EUCLID RTP 6.1. The GRACE Consortium. Objective: To accelerate the application of AI techniques and advanced HCI and software engineering in C 3 I workstations. Size of EUCLID RTP 6.1. 21 Mecus of which company investment 24% 125 man-years 5 years duration - to September 1998
E N D
EUCLID RTP 6.1 The GRACE Consortium Objective: To accelerate the application of AI techniques and advanced HCI and software engineering in C3I workstations
Size of EUCLID RTP 6.1 • 21 Mecus of which company investment 24% • 125 man-years • 5 years duration - to September 1998 • 7 nations, 18 companies
EUCLID RTP 6.1 Participants EUCLID RTP 6.1 Management Group UK DERA Denmark France Italy Netherlands Norway Spain Logica TERMA CSC Data- centralen CRI Matra Steria Datamat Marconi Origin Signaal NLR TNO-FEL KOG SINTEF NDRE INDRA CESAT The GRACE Consortium Grouping for Research into Advanced C3I for Europe
Achieving RTP 6.1 objectives • demonstrate key AI and HCI methods and tools applied to C3I systems • define and implement a workstation architecture able to accommodate existing and emerging methods and tools and to forge new standards • evaluate using domain simulations of landtactical and naval scenarios • encourage European collaboration in C3I R&D
Wide Area Picture EUCLID RTP 6.1 demonstrator Automated Report Analysis Naval Decision Support, Planning and Tasking Army Decision Support, Planning and Tasking Human-Computer Interface Framework CABLE C3I Application Building Environment - An environment for developing multi-agent systems
Reports Reports Scenarios, simulation & testing Naval Scenario Naval simulator ARA Naval DSPT WAP Army DSPT Army Scenario Army simulator HCI Framework CABLE Evaluation metrics Test Harness EUCLID RTP 6.1 Demonstrator
User facilities • Uses Artificial Intelligence • Give a military benefit • Has a user interface
Command tasks supported • Automated Report Analysis • Message handling and report analysis • Wide Area Picture compilation • Army Decision Support, Planning and Tasking • Terrain analysis • Course Of Action construction • Manoeuvre and Fire Support planning • Naval Decision Support, Planning and Tasking • Tactical Threat evaluation • Anti-Surface Warfare planning • Engagement Coordination
Wide Area Picture Automated Report Analysis Reports from Simulators Wide Area Picture Compiler Message Handler To DSPT Reduce the volume of data and ensure important information is extracted and prioritised. Parse structured messages using Yacc++. Knowledge-based analysis of keywords and content. Common Object Model based on ATCCIS for WAP. Clustering techniques to identify groupings.Fuzzy logic inferencing to identify behaviours.
Wide Area Picture Army Decision Support, Planning & Tasking Terrain Analysis Course of Action Construction Manoeuvre Planning Fire Support Planning Plan quicker than the enemy. Cope with complex constraints. Investigate more alternatives. Analyse terrain to identify Mobility Corridors. Compare own and enemy Courses Of Action Combine knowledge- and constraint-based planning. Simulated annealing for fire support planning. Tools for ORBAT browsing and working with time.
Wide Area Picture Naval Decision Support, Planning & Tasking Engagement Coordination Planning Anti-Surface Warfare Planning Tactical Threat Evaluation Time-critical planning.Time, space, resource constraints.Investigate more alternatives. Knowledge-based recognition of enemy’s plan. Constraint-based manoeuvre planning. Anytime analysis of routes and intervisibility. Combined uncertainty- and constraint-based planning Case-based retrieval of relevant existing plans.
Multi-agent architecture • Intelligent agents • logical next step combining OO and KBS • autonomous agents • inter-agent communication • Benefits of a multi-agent system • cope with complexity • intuitive: analogy with human organisations • heterogeneous, plug and play, supports integration • transparent distribution across a network • flexibly partition tasks between humans and machines • emergent behaviour
CABLE • CABLE • C3I Application Building Environment • software to develop and run multi-agent applications • CORBA-compliant • based on ORBIX • inter-agent communication using IDL • Specification, design and implementation of multi-agent systems • analogy with human organisations
Terrain features Mobility corridors Course of action Own forces situation C3I Application Building Environment CABLE HCI Framework Agent: an autonomous intelligent software entity Agent Agents communicate Cooperative working Agent Agent Computer 1 Agent Computer 2 Agents can be distributed Based on ILOG Views Based on CORBA, ORBIX, IDL, STL
Continuous, timely planning Deadline Cost of change Re-plan Act Plan Commit New info. Communicate Request & Deadline Resource request AI planning Cost/quality trade-off Resource agent Anytime algorithms Resource Other processes Other processes Other processes
Groupworking of humans and agents Group working controller User interface Computer Agents
Conclusions • Support heterogeneous AI techniques • Integrate using a multi-agent architecture • Common representation of military concepts • Architecture & techniques used in 10 projects, £2M • More information • paper, Web site: http://public.logica.com/~grace • for UK: Final Report, CD-ROM, FTP server