Real-time Resilient Trajectory Planning for Rocket PropelledProjectiles
Abstract
The objective of the proposed research is to develop reliable, resilient, and real-time implementable computational methods for automated trajectory optimization and control of Rocket Propelled Projectiles (RPPs). We will demonstrate the proposed algorithms with flight experiments by using in-house developed quad-rotors as surrogate flight platforms, via both indoor and outdoor flights. We propose an optimization-based framework to formulate resilienttrajectory planning (guidance) problems for both single and multiple RPPs. The trajectory planning problems will be formulated as optimization problems, which are typically non-convex and hence difficult to solve. We will use a Successive Convexification (SCvx) algorithm to convert these difficult, non-convex problems into sequences of convex optimization problems.These convex problems will be numerically tractable both for pre-flight and in-flight computations. We have already developed the initial SCvx algorithm for trajectory optimization of a single RPP as well as the theory supporting its convergence properties. We have also performed preliminary testing of this algorithm through both indoor and outdoor quad-rotor flight experiments. These tests have demonstrated the core trajectory planning and feedback control capabilities in the presence of aerodynamic disturbance forces. Building on these developments, we propose to extend the SCvx method to handle resilient trajectory planning problems for single and multi RPP flight scenarios. The key difference in resilient planning is that each planning solution will generate a family of trajectories reaching to multiple objectives.These objectives include various targets and safe escape paths. This will maximize the probability of mission success while ensuring safety in the presence of operational and environmental uncertainties.In summary, we aim to develop resilient planning and control capabilities for RPPs via a rigorous optimization-based computational framework by building: A mathematical formulation and algorithms utilizing SCvx to solve resilient guidance problems for a single RPP. Mathematical and computational methods extending the resilient guidance from a single RPP to multiple RPPs. A pseudo-haptic human-autonomy interface for non-expert users to interface with the resilient trajectory planner. Indoor and outdoor flight experiments to demonstrate the developed algorithmic capabilities.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 29, 2020
- Source ID
- N000142012288
Entities
People
- Behçet Açıkmeşe
Organizations
- Office of Naval Research
- United States Navy
- University of Washington