Complex systems such as aircraft present major challenges to control system designers. Both military and civilian sectors are now demanding new and sophisticated capabilities from aircraft systems. Vertical take-off and landing vehicles of tilt rotor or helicopter type exhibit the advantages in the maneuverability features required by Unmanned Air Vehicles (UAV). [1] Such vehicles will require little human intervention from take-off to landing. While flying, the terrain is scanned in many wavelengths at the same time when basic stability augmentation measurements are acquired or estimated through inertial and GPS type sensors. Such vehicles will have to communicate with each other while in flight to coordinate their missions. Also, they will report to each other about special conditions, weather and the location of the “target”. Since the missions are to be pre-programmed the operators will intercede only if mission parameters change. Operators will send the mission data and other commands to UAV via satellite or manned aircraft on multiple redundant paths, such that is one communication path fails or is jammed others ensure that the UAV stays in control. If communications should become cut off the UAV will fly to a preprogrammed area or return to the base.

Such systems, in spite of the wide range of applications that they can be subjected to, are not traditionally offered by manufacturers since the present architectures are customized for specific missions. The UAV has to have implemented on the same hardware platform, algorithms providing adaptability, reconfigurability and interoperability. As a consequence the degree of autonomy varies and, besides common factors, like power or energy stored on-board, the limited bandwidth of current flight control systems represents a limiting factor. Such extra design constraints are to be overcome through research.

Our proposal is mainly concentrated at the reconfigurable and high-level supervision of more common tasks for the UAV such as the attitude stability augmentation controller, obstacle avoidance and operator interface. We are planning to prove such advanced control concepts using a build prototype that has a helicopter type structure that blends in concepts from tilt rotor flying (i.e. a redundant number of ducted fans capable of tilt surrounding a central fixed pitch propeller actuated by pairs of electric motors  powered by a battery or a generator driven by a gas engine). This vehicle configuration has been imagined such as to embed enough redundancy in the system to allow for control reconfiguration, which advocate an increase safety and hence the ability to fulfill the mission.

Significant performance improvement is sought through the UAV’s flight controller ability to switch algorithm components at the same time as the mission priorities are adapted to the new limitations encountered by the damaged system. Under such circumstances the flexibility for the control reconfiguration problem creates the requirement for a high fidelity model ready to embed all necessary plant characteristics. Relevant expertise in the area of flight dynamics combined with achievements in the area of reconfigurable control are to be focused with the aim of providing a reliable solution that impacts this niche in the aerospace industry.

One of the major objectives of this project is to improve safety through reconfigurable control. This creates the opportunity for the operator to ensure completing the mission in a reduced workload context. The main challenge of ensuring reconfiguration is located at the fidelity with which we encompass faults ready to develop in fatal failures. Aerospace systems are generally characterized by a model structure, which requires preservation. Being more specific we are planning to employ for the practical rig a modeling technique, which leaves open for identification a number of aerodynamic and coupling parameters. Hence, grey box identification techniques are suitable in the process of fault detection.

The nonlinear model, which is created to represent at all times a good approximation of the true aircraft, can be transformed into a Linear Parameter Varying (LPV) system through the use of function substitution method or quasi-LPV techniques, especially   suitable when we deal with state/output nonlinearities. Beyond the modeling part, this special form of an indirect adaptive controller will benefit from several new improvements in standard Model Based Predictive Control (MBPC), which are the key to a real time implementation with a sampling frequency as high as 50Hz.

A hybrid control system, acting as a supervisor, has as main feature the ability to follow a mode selected by the operator through a comparison with the vehicle state in terms of its capabilities and an alteration of the original mission if required.

We are aware that situation awareness, reactive control, mode selection and transition require fast response. Yet, the computation necessary for the communication and production of the control variables severely taxes the limits of the available hardware/software platforms. Since the practicalities of a rig demonstration are faced, it is essential to use a rapid development and prototyping tool, which will permit the controller implementation. The methodology to be investigated as practical solution requires an open control platform design implemented on parallel processors. Such a solution prioritizes the dynamic reconfiguration of the controller resulting in a scheme called adaptive resource management. The question still left open at the end of this research pursuit is “How can we certify such complex flight controllers?”