An Intelligent Tutoring System (ITS) for Future Combat Systems (FCS) Robotic Vehicle Command I/ITSEC 2003 Presented by: Randy Jensen jensen@stottlerhenke. com Co-authors: Henry Marshall, US Army RDECOM Jeffrey Stahl, US Army RDECOM Richard Stottler, Stottler Henke
FCS Concept - Background Distributed robotic vehicles and sensors are networked to control vehicles, providing • heightened situational awareness • extended sensor capabilities • reduced human risk
FCS – Training Challenge • New paradigm requires scenario-based practice for FCS warfighters • Formal tactical doctrine for FCS operational concept has not been developed • Desirable to minimize costs of developing and administering training – reduce requirements for human instructors and simplify scenario definition • Intelligent Tutoring Systems are effective for simulating some of the benefits of a human instructor, especially for a domain with focused, task-based exercises
Simulation Testbed • Embedded Combined Arms Tactical Training and Mission Rehearsal (ECATT/MR) testbed developed at RDECOM • Multi-screen control interface, on OTB simulation • Software for controlling simulated entities is the same as that used for operating robotic vehicles
Testbed User Interface Closeup
Intelligent Tutoring System (ITS) Architecture Overview Simulation Interface provides two forms of data to Evaluation Machines: • Simulation states • Student actions Instructional Manager sends notifications back to the student in the OCU environment, based on conclusions from the Evaluation Machines
Principle-Based Evaluation Distinct evaluation mechanisms indexed to instructional principles
FCS ITS Instructional Principle Categories TACTICAL DECISION MAKING • Student’s ability to interpret the tactical situation and commander’s intent, and decide what should be done • Example: Use airborne sensor assets to complement knowledge from ground-based vehicles COMMAND FORMULATION • Student’s ability to translate tactical decisions into commands or orders that can be executed EXECUTION • Student’s application of correct buttonology in execution
FCS ITS Instructional Principle Examples TASK Use terrain concealment to detect enemy positions from Unmanned Ground Vehicles (UGVs) without being detected
FCS ITS Instructional Principle Examples: TACTICAL: Before cresting hills in terrain, halt UGV and use mast sensors to scan for enemy
FCS ITS Instructional Principle Examples: COMMAND FORMULATION: When UGV movement will include successive halt and resume, control the vehicle with draggable points in the OCU
FCS ITS Instructional Principle Examples: EXECUTION: The main HALT control halts all vehicles; the HALT control under “Assign Task” halts the current vehicle
Finite State Machine (FSM) Based Evaluations What are they? Transition networks executing in coordination with a simulation to gather data about instructionally significant events and states, and make evaluation conclusions in real time Why use them in an ITS? Several benefits: • Modularity – they can be used separately or in conjunction for a variety of scenarios • Instructional correspondence – individual instructional principles can be associated with independent evaluations • Integration – the FSM structure is easily integrated with free-play simulations and maps well to diagnosis of widely varied outcomes • Authoring ease – they can be represented visually, making them easy for non-programmers to create, maintain, and revise
Evaluation Machine Example TACTICAL: Before cresting hills in terrain, halt UGV and use mast sensors to scan for enemy
Lessons Learned • Automated evaluation is suited for the domain of training the employment of robotic vehicles under the FCS concept • Streamlining domain-specific requirements (simulation integration, scoping training objectives, etc. ) reduces ITS development time and cost • Preferable to avoid scenario-specific evaluation • Example: Identifying terrain where a UGV has an exposed hull. • Scenario-specific approach: Manually annotate areas on the map that represent hill crests where a UGV would be exposed • Scenario-independent approach: Use dynamic line of sight (LOS) calculations in the simulation to determine exposure
Future Work • Full system development with a rigorous collection of scenarios • Enhanced feedback mechanisms, potentially with controls to pause or rewind the simulation • Team training extensions • Similar architecture applies in the team setting • Scalable principle hierarchy supports reuse with scenarios involving a superset of instructional concepts • ITS capabilities proposed for Integration into the Tank and Automotive Research and Development Command (TARDEC) Crew instrumentation and Automation Testbed
Скачать презентацию An Intelligent Tutoring System ITS for Future Combat
2b2cf85c68e94a83ae6fb96249936b29.ppt