DARPA, managed by the U.S. Air Force Research Laboratory:
An Integrated, Multi-Layer Approach to Software-Enabled Control:
Mission Planning to Vehicle Control

Principal Investigator: Professor Gary Balas


Project Overview:

The proposed SEC program will develop a software environment to demonstrate integrated control system technologies which enable the use of multi-unmanned combat air vehicles in strategic situations. Test case scenarios for splitting and merging of Unmanned Combat Air Vehicles (UCAV's) will be selected based upon concepts of operations. System requirements will included adaptation to subsystem malfunctions, imperfect models, system uncertainty and unknown battlefield characteristics to achieve high all level mission objectives. We will develop a software architecture and framework for implementation of advanced integrated control algorithms. The software implementation provides and integrated multi-layer approach to provide autonomous reconfigurable vehicle control capability for UCAV's from top level multi-vehicle mission management to inner-loop vehicle control. This implementation will be demonstrated via real-life software simulation.

University of Minnesota has teamed with UC-Berkeley and Caltech in the development, enhancement, and transition of integrated control and software technologies. This team possesses the key ingredients needed to achieve program goals. The strength of the university team members is in the area of robust system algorithm and tool development. A three layer hierarchical problem structure is proposed for our integrated, multi-layer approach to strategic use of UCAV's.

The top layer multi-vehicle mission planner is implemented via genetic algorithms, the middle layer consists of a trajectory generation and conflict resolution optimization using dynamic programming with the bottom layer, the foundation, performing integrated on and off-line vehicle control. An indirect adaptive systems approach will form the basis of this control architecture. System configuration will use real-time parameter identification methods in conjunction with failure detection and identification approaches to identify unforeseen changes in the system dynamics. This information will be provided to all three layers of the control architecture.

The mission planning layer provides task assignment and outer-loop commands for the UCAV vehicle management system (VMS). The trajectory generation and conflict resolution layer algorithms make use of dynamic programming techniques combined with engineering insight. Mathematical models of individual UCAV's are included at the base level for use in the mission. The individual vehicle control combines on-line receding-horizon schemes, real-time identification and failure detection with the robust inner-loop control system. An off-line controller is designed using the most advanced, robust nonlinear control theory, linear-parameter varying control, to ensure a guaranteed level of stability and performance in the presence of specified uncertainties.

The on-line controller will provide adaptation to unforeseen changes in the aircraft dynamics. Our key technology innovation is integration of state-of-the-art approaches to every aspect of multi-vehicle UCAV planning and usage in a common design environment. This project provides a focal point for integrating all recent advances in robust, nonlinear control, trajectory optimization and conflict resolution with coordinated command, planning and control into a software design environment for multiple UCAV's. The team will work in close coordination to provide cost-effective, revolutionary advancements for UCAV systems.