Guaranteed Performance of Dynamic Behaviour of Multiple Unmanned Aerial Vehicles

The use of multiple unmanned aerial vehicles (MUAVs) can provide significant reductions in manpower and risk to humans for critical security and defence roles. In particular, MUAVs potentially offer: 1) enhancement of the coverage of large areas by improvement in latency and accuracy of information gathering; 2) increase in the mission success rate; 3) enabling new tasks and operations by increasing autonomy; 4) robustness and benign degradation in performance. A key utility feature of a single UAV is that it is a mobile sensor platform. MUAVs offer a magnification of the sensingcapability by creating an airborne, relocatable multi-sensor. This capability offers new opportunities for surveillance and reconnaissance missions which provide the practical context of the proposed research.A key issue which must be addressed in order for the potential benefits to become real is guaranteed performance of the dynamic behaviour of MUAVs. In particular, coordinating flight dynamics of MUAVs in a predictable and verifiable way is crucial for their certification in the airspace and acceptance for safety-critical missions. Therefore, the main focus of this proposal is to develop and validate a rigorous, analytical framework for deriving and accessing guaranteed dynamic and kinematic performance of MUAVs. The dynamic behaviour of MUAVs is fundamentally different to that of a single vehicle. This pronounced difference occurs, because a group of MUAVs constitutes an interconnected dynamical system depending on the individual dynamics and on the nature of the vehicle-vehicle and environment-vehicle interactions. The representation of this dynamical system as a distinct entity and the optimisation of its performance is achallenging study. Despite the recent advances, an actual operation of a group of MUAVS has not been fully realised yet.The main thrust of the proposed project is to develop a rigorous, analytical framework that determines in detail the number and characteristics of the vehicles, the interactions between the vehicles and their embedded guidance and control schemes. This framework will be used to predict MUAVs behaviour in two multitask mission scenarios, and assess the resulting performance through simulations and ground vehicle experiments. The framework is expressed in terms of a hierarchical character of co-operative controller of multiple UAVs; the hierarchy comprises three layers. Cooperative decision-making in Layer 1 is done by UAVs acting as a group co-operating according to common mission/tasks. Layer 1 decision-making results in generation of co-operative trajectories in Layer 2, realising coordinated guidance for the UAVs acting as a group. The coordinated guidance produces co-operating trajectories. Each reference trajectory (guidance demand) generated in Layer 2 is then followed by the individual controller of the corresponding UAV in Layer 3. Thus, the overall controller is obtained by co-operation of all individual controllers of the UAVs, with the cooperationdecided on Level 1, and defined by the trajectory tracking requirements in Level 2. The central design issue is how to design each layer and their interaction, so that the individual and group performance can be guaranteed.The programme of this research work will involve both the study of each of three layers of cooperative control and the integration of them for two example scenarios. Two PhD students and two post doctoral research associates (RA) will develop new solutions for both scenarios. Cranfield's PhD student will develop new algorithms for coordinated guidance, while Imperial's PhD student will focus on the use of Mixed-Integer Optimization and Parametric Programming methods to express dynamics of the vehicles. Cranfield's post-doc RA will focus on the integration of planning (Layer 1) andguidance (Layer 2). Imperial's RA will focus on the integration of guidance (Layer 2) and tracking control (Layer 3).