Short Bio
Mission
Expertise
Research Interests
Teaching
NSERC Type Bio
Professional Links
Personal Links
Home
Short Bio
Mission
Expertise
Research Interests
Teaching
NSERC Type Bio
Professional Links
Personal Links
Short Bio
Mission
Expertise
Research Interests
Teaching
NSERC Type Bio
Professional Links
Personal Links
Towards a new architecture and simulation environment for
Air Traffic Management Systems (ATM)
An insight into the problem
In an environment where:
the traffic is growing with a rate of 3% up to 5% per annum due to a soaring demand in air travel
the present architecture and management are unable to handle this increase after a certain limit due to inefficient airspace utilisation
the ATC workload is increasing and old dated technologies are still operating in some areas
we consider that an effort is necessary to meet the next century challenges.
Two main streams were followed up to this moment in the automation of ATM systems:
The development of traffic alert and collision avoidance system (TCAS) which is in fact an on board algorithm which provides resolution advisories to the pilots when detecting conflicts between aircraft.
The introduction of the concept of free flight developed for the first time by the Radio Technical Commission for Aeronautics, concept which allows the pilots to give the aircraft possibility to self optimise by choosing its own routes, altitude and speed in order to ensure a prescribed separation with respect to the surrounding aircraft and prevent a violation of the special use airspace.
According to several of the authors involved up to now in the research pursued in this field [MIT, NASA Ames, UC at Berkely, SeaGull, Honeywell] there are few enabling technologies that make the free flight concept feasible:
Global Positioning System (GPS)
Data-link type of communications
Automatic Dependence Surveillance-Broadcast (ADSB)
Traffic Alert and Collision Avoidance System (TCAS)
Powerful on-board computers
Over the past few years the progress in solving this problem was slow due to the complex migration from a centralised system (the present ATC which assumes most of the workload based on ground controllers) towards a decentralised architecture (based on on-board equipment that will share some of the ground workload, navigation, weather prediction and last but not least the aircraft en route separation). The novel approach towards which we are heading has fault tolerant features which enable us to consider that it will improve the current standards of safety.
Therefore, research has to be done in two major directions:
the conflict detection and resolution
the simulation of the ATM systems
the traffic is growing with a rate of 3% up to 5% per annum due to a soaring demand in air travel
the present architecture and management are unable to handle this increase after a certain limit due to inefficient airspace utilisation
the ATC workload is increasing and old dated technologies are still operating in some areas
we consider that an effort is necessary to meet the next century challenges.
Two main streams were followed up to this moment in the automation of ATM systems:
The development of traffic alert and collision avoidance system (TCAS) which is in fact an on board algorithm which provides resolution advisories to the pilots when detecting conflicts between aircraft.
The introduction of the concept of free flight developed for the first time by the Radio Technical Commission for Aeronautics, concept which allows the pilots to give the aircraft possibility to self optimise by choosing its own routes, altitude and speed in order to ensure a prescribed separation with respect to the surrounding aircraft and prevent a violation of the special use airspace.
According to several of the authors involved up to now in the research pursued in this field [MIT, NASA Ames, UC at Berkely, SeaGull, Honeywell] there are few enabling technologies that make the free flight concept feasible:
Global Positioning System (GPS)
Data-link type of communications
Automatic Dependence Surveillance-Broadcast (ADSB)
Traffic Alert and Collision Avoidance System (TCAS)
Powerful on-board computers
Over the past few years the progress in solving this problem was slow due to the complex migration from a centralised system (the present ATC which assumes most of the workload based on ground controllers) towards a decentralised architecture (based on on-board equipment that will share some of the ground workload, navigation, weather prediction and last but not least the aircraft en route separation). The novel approach towards which we are heading has fault tolerant features which enable us to consider that it will improve the current standards of safety.
Therefore, research has to be done in two major directions:
the conflict detection and resolution
the simulation of the ATM systems
Steps forward
Conflict detection and resolution
Dealing with conflict detection and resolution in the free flight concept we are going to consider the following approaches:
1. The conflict probability estimation This is based on the trajectory prediction followed by an algorithm that computes the conflict probability [NASA Ames]
2. The non-cooperative way This means that the problem is posed in a game approach fashion for a model obtained based on material point dynamics for two or many aircraft when no information is exchanged between them. The optimisation is carried out for the worst possible actions of all the other aircraft. This algorithm provides the safest possible strategy to each aircraft regardless of the actions of others by imposing a trajectory that guarantees the minimum separation. Of course this approach is somehow conservative.
3. The cooperative way In this case the aircraft exchange data and perform coordinated manoeuvres in order to solve the potential conflict. An added advantage of this method is the easy understanding by the ATC controllers (the future supervisors of this automated system) and the possibility to implement it on the present available on-board computers which currently perform flight management and guidance tasks [UC at Berkeley, EUROCONTROL].
Considering the deterministic way of approaching both non-cooperative and cooperative strategies we have imagined two original strategies to deal with the foregoing issues:
In the non-cooperative strategy case we suggest a nonlinear predictive control approach to be carried out when performing the minimisation of the Hamiltonian involved when dealing with the problem of conflict resolution in the case of no communication between aircraft.
In the cooperative strategy situation we suggest that the flight management system developed in [ACC'97] as a combination of the model based predictive control and robust H-inf loop-shaping can be extended to embed this novel concept and provide both the pilot and the ground air traffic controllers with a set of manoeuvres that is proven safe and covers all possible conflict scenarios. A hybrid system approach will be taken when modelling the behaviour of the ATM system. This leaves enough space for a theoretical analysis of the stability of this complex set-up.
Dealing with conflict detection and resolution in the free flight concept we are going to consider the following approaches:
1. The conflict probability estimation This is based on the trajectory prediction followed by an algorithm that computes the conflict probability [NASA Ames]
2. The non-cooperative way This means that the problem is posed in a game approach fashion for a model obtained based on material point dynamics for two or many aircraft when no information is exchanged between them. The optimisation is carried out for the worst possible actions of all the other aircraft. This algorithm provides the safest possible strategy to each aircraft regardless of the actions of others by imposing a trajectory that guarantees the minimum separation. Of course this approach is somehow conservative.
3. The cooperative way In this case the aircraft exchange data and perform coordinated manoeuvres in order to solve the potential conflict. An added advantage of this method is the easy understanding by the ATC controllers (the future supervisors of this automated system) and the possibility to implement it on the present available on-board computers which currently perform flight management and guidance tasks [UC at Berkeley, EUROCONTROL].
Considering the deterministic way of approaching both non-cooperative and cooperative strategies we have imagined two original strategies to deal with the foregoing issues:
In the non-cooperative strategy case we suggest a nonlinear predictive control approach to be carried out when performing the minimisation of the Hamiltonian involved when dealing with the problem of conflict resolution in the case of no communication between aircraft.
In the cooperative strategy situation we suggest that the flight management system developed in [ACC'97] as a combination of the model based predictive control and robust H-inf loop-shaping can be extended to embed this novel concept and provide both the pilot and the ground air traffic controllers with a set of manoeuvres that is proven safe and covers all possible conflict scenarios. A hybrid system approach will be taken when modelling the behaviour of the ATM system. This leaves enough space for a theoretical analysis of the stability of this complex set-up.
The architecture
A possible structure for such a system to be developed can be:
A strategic planner designed to provide a coarse trajectory for the aircraft in the form of a sequence of control points (obtained by interpolating the way points given by ATC).
A tactical and trajectory planner placed on-board having multiple functions from the prediction of possible conflicts with other aircrafts to the optimisation of the output trajectory or notification of the strategic planner about the conflicting trajectories.
The flight control system able to follow the references provided by the tactical and trajectory planner. So, the hierarchical system suggested is regarded as a discrete event controller acting at a higher level and providing time stamped references to the flight management component of the automatic pilot.
A strategic planner designed to provide a coarse trajectory for the aircraft in the form of a sequence of control points (obtained by interpolating the way points given by ATC).
A tactical and trajectory planner placed on-board having multiple functions from the prediction of possible conflicts with other aircrafts to the optimisation of the output trajectory or notification of the strategic planner about the conflicting trajectories.
The flight control system able to follow the references provided by the tactical and trajectory planner. So, the hierarchical system suggested is regarded as a discrete event controller acting at a higher level and providing time stamped references to the flight management component of the automatic pilot.
The ATM simulation environment
In evaluating the abilities of the newly developed ATM strategy we consider vital the development of an object oriented modelling and simulation facility. Such an environment should cope with modelling and simulation of airspace operations, conflict resolution, airport and runway capacity and operations, respectively.
A certain trend in developing simulation software for ATM systems (i.e. hybrid systems based on a hierarchical and parallel combination of multiple hybrid automata created, interconnected and dissolved as the system evolves) is materialised in the attempts to use SHIFT - an object oriented language employed for hybrid systems design and simulation - to deal with such issues. While evaluating this language we have concluded that, due to its features which address non-continuous dependence on the initial conditions, event detection and multiple time scales, it can be at the base of a much complex system enabling, the development of a library of functions oriented towards ATM.
As an extension of the simulation system imagined above we consider as necessary the development of visualisation tool. This software should enable the user to view the controlled airspace, the trajectory followed by a particular aircraft and to reflect in an animated manner the evolution of the "targets" such as to enhance the evaluation of the overall system performance as well as coordination protocols. Classic ways to present the trajectory of the aircraft situated in the control zone such as air-traffic control centre, control tower, fixed point and cockpit will be required. Moreover the ability of the software to configure for a particular ATC area or airport is necessary in order to ensure enough flexibility which will enable the designer, user or certification authority to compare various procedures adopted. A compatible interface has to be adopted in order to enable a comparison with other software available in the future on the market. Therefore this visualisation tool will be used extensively in the design process as well in the certification procedures.
A certain trend in developing simulation software for ATM systems (i.e. hybrid systems based on a hierarchical and parallel combination of multiple hybrid automata created, interconnected and dissolved as the system evolves) is materialised in the attempts to use SHIFT - an object oriented language employed for hybrid systems design and simulation - to deal with such issues. While evaluating this language we have concluded that, due to its features which address non-continuous dependence on the initial conditions, event detection and multiple time scales, it can be at the base of a much complex system enabling, the development of a library of functions oriented towards ATM.
As an extension of the simulation system imagined above we consider as necessary the development of visualisation tool. This software should enable the user to view the controlled airspace, the trajectory followed by a particular aircraft and to reflect in an animated manner the evolution of the "targets" such as to enhance the evaluation of the overall system performance as well as coordination protocols. Classic ways to present the trajectory of the aircraft situated in the control zone such as air-traffic control centre, control tower, fixed point and cockpit will be required. Moreover the ability of the software to configure for a particular ATC area or airport is necessary in order to ensure enough flexibility which will enable the designer, user or certification authority to compare various procedures adopted. A compatible interface has to be adopted in order to enable a comparison with other software available in the future on the market. Therefore this visualisation tool will be used extensively in the design process as well in the certification procedures.
Project milestones
Advantages and drawbacks of such a decentralised decision making system for Air Traffic Management (ATM) will be investigated having as a result a complete working environment for air traffic personnel. We regard this system as to be implemented in the immediate future, therefore we need to consider mainly its advisory features, both for pilots and air traffic controllers. Such an approach will provide enough funds to continue the development by financing it from the profit made on its first release. Only after a through testing period we can think about certification and a possible implementation at the ground and on-board levels.
The outcome of the project after two years of research, should contain an ATC management simulator -- system which will involve the above described control methods implemented on a hardware platform -- and a flight simulator -- reflecting a virtual set of aircrafts -- fitted with a new flight management system able to interpret commands provided by the modified ATC.
This research aims at an integration with other state of the art techniques developed in USA and Europe. Previous contacts with the European Group for Aerospace Research and Technology in Europe will enable a good technology transfer in order to provide in time the first results of the proposed research.
Supported from both the academic and industrial side we consider that the research will generate the necessary funds required for the application of the solution suggested. This will represent a very good starting point for generalisation of the control methodology on a larger scale.
The outcome of the project after two years of research, should contain an ATC management simulator -- system which will involve the above described control methods implemented on a hardware platform -- and a flight simulator -- reflecting a virtual set of aircrafts -- fitted with a new flight management system able to interpret commands provided by the modified ATC.
This research aims at an integration with other state of the art techniques developed in USA and Europe. Previous contacts with the European Group for Aerospace Research and Technology in Europe will enable a good technology transfer in order to provide in time the first results of the proposed research.
Supported from both the academic and industrial side we consider that the research will generate the necessary funds required for the application of the solution suggested. This will represent a very good starting point for generalisation of the control methodology on a larger scale.
