Abstract

Onboard aircraft systems must frequently plan dynamically feasible paths or trajectories to safely and quickly reach their goal position. Optimized B-splines are promising because they can guarantee kinematic feasibility in a computationally efficient manner and can be applied to a variety of vehicle models. However, current B-spline solutions do not adequately constrain trajectories to the physical limitation of the vehicles they are planning for, or are often too conservative. Most are also limited to path smoothing or require pre-optimized timing parameters. This thesis develops a real-time B-spline optimization-based trajectory generator capable of minimizing time while respecting the velocity, acceleration, and turn limitations of various vehicles. This thesis also develops a minimum-distance B-spline optimization-based path planner with obstacle, curvature, and slope constraints. This work takes advantage of the convex bounding property of B-splines to construct constraints that do not require discrete sampling of the path. This allows the path to scale in size without added computational costs. By converting B-spline control points to Minvo control points, the constraints are also non-conservative in relation to the vehicle limitations. Novel contributions include the development of B-spline optimization constraints that guarantee feasibility over curvature, slope, angular rate, centripetal acceleration, and tangential acceleration. This research also develops a direct obstacle avoidance constraint that does not require pre-planning steps. It additionally implements linear safe flight-corridor constraints into the path optimization. These techniques are validated through successful path and trajectory tracking by a fixed-wing aircraft in simulation.

Degree

College and Department

Ira A. Fulton College of Engineering; Mechanical Engineering

Rights

https://lib.byu.edu/about/copyright/