The growing complexity of future military systems demands the development of high-performance nonlinear control algorithms. The goal of this proposal is to develop a specially designed software/hardware architecture to enable real-time autonomous closed-loop control of nonlinear high-dimensional dynamical systems. Following the recent success of pseudospectral (PS) computational control methods in solving highly nonlinear real-world control problems, we propose and investigate feasibility of a reconfigurable real-time nonlinear optimal control architecture that combines the recent advances and high efficiency of PS methods with the striking growth in the capabilities of modern computing hardware. The proposed closed-loop control algorithm has the potential to yield significantly improved system performance, meet the stringent physical and engineering requirements of various Air Force systems, and easily adapt to changes in the application, environment, and control objectives. The main objectives of this proposal include: (i ) developing a reconfigurable real-time computational optimal control architecture, (ii ) designing an algorithmic structure that is hardware implementable, (iii ) analyzing the computational feasibility of the algorithm for high-dimensional systems, (iv ) designing the appropriate software/hardware architecture, and, (v ) testing the performance of the proposed algorithm on an example problem of autonomous aerial vehicle.
Benefit: The real-time reconfigurable nonlinear control algorithm paired with an efficient custom designed hardware configuration is an enabling technology for managing the complexity and achieving superior performance of many Air Force warfare platforms. The results of the proposed research will particularly play a pivotal role in unmanned vehicle platforms used in both military and civilian applications. Unmanned drones and unmanned ground robots are delivering game-changing capabilities in the battlefield and are used by law enforcement agencies around the world. The realtime software/hardware partnership proposed in this project will be of great value to the unmanned vehicle industry and will be pursued as part of the future Air Force advanced technology. Elissars commercialization plans are focused on commercializing hardware-specific real-time reconfigurable optimal control technology by facilitating the technology transfer to Air Force platforms. Upon successful completion of Phases 1 and 2 of this STTR proposal, Elissar will pursue this endeavor by deployment of the software/harware package on small-scale unmanned ground vehicles and mid-scale unmanned aerial vehicles. In that regard, Elissar has started investigating potential industry partners who are providers of various hardware architectures. As such, two companies SRC Computers and Versa Logic have expressed their interest to partner with Elissar towards commercialization of the outcomes of Phases 1 and 2 of this STTR project, provided the result of phases 1 and 2 indicate the potential suitability of such hardware structure. Please see the attached letter from the president of SRC computers, one of the leading providers of high-performance reconfigurable computing hardware.
Keywords: Closed-Loop Nonlinear Control, Pseudospectral Methods, Real-Time Computational Optimal Control, Reconfigurable Hardware Architecture, Aerospace Guidance And Control Algorithms
