Three-Dimensional Path Planning and Collision-Free Flight Control for Drone-Assisted Autonomous Pollination Systems

Abstract

In this paper, the authors’ previous work regarding a conceptual drone-assisted Autonomous Pollination System (APS) is extended with regards to path planning and flight control. The APS Path Planning module is extended to optimize the path for missions requiring three-dimensional (3D) path planning, such as the pollination of almond trees. A new method of simplifying the 3D path planning problem by selecting cells or groups of flowers to visit is shown in this paper. This method is numerically demonstrated based on a simulated almond tree. The Flight Control module is extended to incorporate drag into a novel convex-optimization-based flight controller and a new method of collision avoidance called Control Sequence Stitching (CSS). A linear drag model is integrated into the flight control formulation, which is validated through a simulated test flight. The concept of CSS is developed and explained as a method to generate seamless flight trajectories while still reaping the benefits of convex optimization. This method can be used to generate collision-free trajectories and control commands rapidly for potential real-world APS missions.