Prescient and Robust Autonomy for Operations on Non-Cooperative Targets

Barron Associates, Inc. (Barron) will develop and commercialize the Automated Prescient and Enhanced Robust Control (AUTOPERCH) system to autonomously land and launch from non-cooperative targets in austere conditions. This program addresses the challenges posed by launch and recovery of rotary-wing vehicles from non-cooperative targets, i.e., operating from unknown ships in potentially high sea-state and windy conditions. The initial focus will be on Tier 1-3 UAS with extensions to piloted rotorcraft in latter phases. Past research in this area typically employs landing aids, communication between the ship and aircraft, or known visual markers to assist in navigation and landing. The current program assumes none of these aids and uses deep learning artificial intelligence (AI) methods to develop navigation aids and target (landing site) relative pose estimates to overcome the lack of landing aids on the ship deck. The Phase I Base program will demonstrate feasibility of the AUTOPERCH approach through simulation and experimental evaluation of the key components, including ship detection and tracking algorithms, landing site tracking and relative pose estimation, ship motion prediction, optimal trajectory generation, and robust guidance, navigation, and control (GNC). The inherent disturbance rejection properties of the flight controller (based on nonlinear dynamic inversion (NLDI)) are augmented with a disturbance estimator that accounts for varying turbulence, ship airwakes, and ground effect from a moving deck. Flight experiments will be conducted using available assets from the Barron and ESA team. The Phase I Option will continue algorithm maturation and integration efforts, implement feedback from the Base period, and finalize Phase II development plans. The Phase II program will culminate with experimental testing in the Navy maneuvering and seakeeping basin (MASK) and on surface vessels at sea under diverse operating conditions. A successful Phase II program will realize the AUTOPERCH algorithms implemented on edge-devices onboard the UAS aircraft. The final product employs state-of-the-art AI, autonomy, GNC, and estimation theory implemented on edge devices running in real-time onboard the aircraft.