Intelligent Control Architectures and Testbed Development f

Autonomous control of multiple UAVs and UCAVS offers special advantages for reconnaissance,surveillance and precision strike. Such missions call for unprecedented levels of autonomy, reliability, precision and fault tolerance over long periods of time, especially for missions with communication blackouts. The complexity of the design process for autonomous fault tolerant control of such multi-levels requires hierarchical supervisory controllers of a hybrid type where low level systems involve continuous dynamics and the higher levels involve discrete actions. In this STTR effort, Scientific Systems Company (SSC) in collaboration with Univ. of California, Berkeley, addresses the problem of systematic design, testing and verification of Intelligent Supervisory Control Architectures for Health Monitoring and Realtime Reconfigurable Control for multiple UCAV missions. We will investigate theoretical and design issues involved in the choice of system architecture and methods for interfacing elements of the resulting hybrid system. To incorporate fault tolerance, a novel failure detection, identification and reconfiguration (FDIR) approach using the combined techniques of Interacting Multiple Model-Extended Kalman Filters(IMM-EKF) and adaptive Nonlinear Model Predictive Control (NMPC) are proposed. This approach autonomously detects and reconfigures the control law for suden or gradual failures of various sensors, actuators and system components. The architectures and controller design developed in this STTR effort will be implemented into a software testbed, by integrating them with SHIFT and CASTLE. SHIFT is a realtime simulation environment for hybrid systems developed by UC Berkeley and CASTLE is a nonlinear 6DOF simulation package developed by NAWC. Specific Phase I tasks are (i) Data Acquisition and Problem Formulation, (ii) Design of Flexible Architectures, Verification and Validation Methodologies for Realtime Supervisory Control for Multiple UCAVs, (iii) Development of a Fault Tolerant Controller and Failure Detection Algorithms (iv) Development of a Simulation Testbed using SHIFT and CASTLE to Evaluate Various Supervisory Control Architectures (v) Reports, Meetings and Phase II Recommendations. Phase II will involve further development of the software testbed and protototype development . Issues related to robustness, optimality and realtime FDIR will be futher investigated and concepts will be tested on a practical system.