The AgentFly ATM simulation suite is a complex tool for modeling and simulation of air traffic and air traffic management. The ATM AgentFly is fast-time gate-to-gate simulation used in NextGen and SESAR programs. The simulation consists of IFR, VFR and unmanned air traffic and main actors - air traffic controllers, pilots, and airline operation centers. The AgentFly platform is agent-based simulation framework designed to be used as Fast-Time Simulation (FTS) as well as Real-Time Simulation (RTS).
AgentFly ATM Simulation Suite
The FTS mode is suitable for what-if studies and analysis, validation of new concepts or interconnection with other FTS systems. The system allows running quickly various options, settings and parameters to evaluate changes. The RTS mode is suitable for connection with other real-time systems and can include humans-in-the-loop, e.g., pseudo-pilots or air traffic controllers.
Agent-based simulation allows precise control of simulation time, large-scale scenarios with various actors (thousands of air traffic controllers and tens of thousands of aircraft) and controlled uncertainty and randomization. The architecture of the system is highly modular, widely configurable and flexible, and it allows easy creation of scenarios.
Key benefits of the AgentFly ATM platform:
- Complex and high fidelity simulation including behavior models
- Integration with other systems
- Wide range of reported metrics
- Easy configuration of various scenarios
- Fast and flexible development of custom features
The AgentFly ATM platform has wide range of use:
- What-if simulation for validation, optimization or improvement of current ATM
- Evaluation of different sectorization
- Effect of Datalink (CPDLC) or SYSCO use
- Validation of improved tools (MTCD, TCT, conflict resolution advisory)
- Study of UAS integration into shared airspace
Model of Actors and Environment
One of the major components of the simulation is model of human cognitive behavior. The model is designed to be generic and it can model various human actors. The AgentFly currently supports executive and planning air traffic controllers in different types of sectors, traffic manager, pilot, remote pilot for UAS, airline operations center operator and others under development. New actors for future concepts (e.g., Flight Centric Environment) are also supported, e.g., incoming traffic allocator, extended planner, etc.. Each actor (air traffic, controller, pilot, etc.) can have specific configuration or same configuration can be used for group of actors.
The cognitive behavior model is based on Multiple Resource Theory using visual, cognitive, auditory and psychomotor resources. Human behavior is defined as set of tasks that represents each actor's interaction with systems, environment, and other actors. The AgentFly emulates inputs (e.g., controller’s screen), communications (radio, telephone, datalink, sysco, etc.), outputs (keyboards, mouse) or environment (view out of windows). The model measures total cognitive workload, execution delays, composition of tasks and other metrics related to human behavior.
Environment of AgentFly simulation can be defined by:
- Maps and satellite images
- Surface height maps
- Wind, convective weather, and other meteorological conditions
AgentFly supports IFR and VFR flights that can be created based on:
- Flight plans
- Actual records of positions
- or can be generated (e.g. emulate increased future traffic)
Airspace definition in AgentFly ATM simulation supports:
- Standard navaids, fixes, routes, SIDs, STARs, etc.
- Sectors and centers are defined as volumic airspaces including border coordination conditions, letters of agreement, standard operating procedures, etc.
- Additional information can define restricted areas, military zones, transition altitudes
AgentFly platform supports easy creation of customer specific data loaders.
Simulated Earth Globe
BADA Performed Trajectory
Trajectory Influenced by Wind
Airspace Divided into Sectors
Aircraft simulation is based on BADA performance model family 3 and 4 which allows precise computation of vertical profile and measure fuel consumption, distance flown, duration the flight, etc. Other models can be integrated into system, e.g., simpler models representing smaller or unmanned aircraft or more detailed model for specific aircraft type. The trajectory can be planned using standard navaids, allowing partial or complete free-routing or full 4D trajectory. The trajectory can be optimized to achieve increased efficiency or other metrics.
What AgentFly Offers
The ATM AgentFly has wide range of use. It can be used as fast-time large-scale what-if simulation to evaluate, validate, optimize or improve current ATM, e.g., evaluation of different sectorizations, use of datalink or sysco coordination, validation of improved tools (MTCD, TCT, conflict resolution advisory), use of directs and free routing, continuous descend evaluation, etc. Thanks to architecture design and modularity, the AgentFly is extremely fast and efficient in design and development of user-defined new tools, strategies and complex concepts and ideas, e.g., validation of flight centric / sectorless environment concept and its components.
- The AgentFly can be interconnected with external systems, both fast-time and real-time.
- The AgentFly can provide interaction and increased realism by simulation of sectors surrounding selected sectors controlled by human-in-the-loop air traffic controllers.
- Simulated ATCs can be used to execute instructions of human traffic managers to provide feedback and study impact of these instructions to traffic flows.
- The AgentFly actors are able to communicate with human actors using automation (datalink, sysco, etc.).
The AgentFly supports wide range of logs, statistics, reports, and analytics. Basic data loggers can capture detailed raw data, e.g., all aircraft states logged each simulation step, detailed performance data of simulated actors. Statistics aggregate raw data and provide more complex data, e.g., dynamic density inside sectors, flight levels occupancy and total aircraft fuel consumption. Reports provide high-level information, e.g., flow characteristics, low and peak periods of workload. Analytics can provide a more complex view of the simulation, categorize data based on conditions, and indicate optimization suggestions. All output can be provided in user-defined format.
ATM AgentFly | Video