Time-Optimal Gate-Traversing Planner for Autonomous Drone Racing

TL;DR

The paper tackles the time-optimal gate-traversing problem for autonomous drone racing, arguing that previous methods neglect gate geometry by treating gates as points or simple balls. It presents a fast, polynomial-based TOGT planner that models full quadrotor dynamics with single-rotor thrust limits, and uses a MINCO flat-output trajectory framework to parameterize feasible paths. By transforming the problem into a waypoint-time segmentation and applying smooth surjections to eliminate gate and time constraints, the method solves an unconstrained optimization with L-BFGS and then reconstructs full state and control trajectories, achieving seconds-level computation even for tracks with dozens of gates. Extensive simulations and real-world experiments demonstrate improved time efficiency by exploiting gate configurations, generalizability to varied race tracks, and robustness to challenging gate geometries such as dives and tunnels.

Abstract

In drone racing, the time-minimum trajectory is affected by the drone's capabilities, the layout of the race track, and the configurations of the gates (e.g., their shapes and sizes). However, previous studies neglect the configuration of the gates, simply rendering drone racing a waypoint-passing task. This formulation often leads to a conservative choice of paths through the gates, as the spatial potential of the gates is not fully utilized. To address this issue, we present a time-optimal planner that can faithfully model gate constraints with various configurations and thereby generate a more time-efficient trajectory while considering the single-rotor-thrust limits. Our approach excels in computational efficiency which only takes a few seconds to compute the full state and control trajectories of the drone through tracks with dozens of different gates. Extensive simulations and experiments confirm the effectiveness of the proposed methodology, showing that the lap time can be further reduced by taking into account the gate's configuration. We validate our planner in real-world flights and demonstrate super-extreme flight trajectory through race tracks.

Figures (8)

• Figure 1: (a) A time-optimal flight path generated by the proposed TOGT planner and executed in a motion capture room. (b) Our framework supports a wide range of gate shapes. (c-d) Comparison of trajectories obtained from our method with gate constraints (left figure) and with waypoint constraints (right figure). We can clearly see that the trajectory on the left is more time-efficient as it traverses the gates along their boundaries.
• Figure 2: Illustration of the segmented trajectories based on the gate order.
• Figure 3: Our planner presents a nearly linear increase in computation time as the number of gates increases.
• Figure 4: Comparison of single-rotor thrust trajectories from our method (the TOGT and TOGT-WP) and CPC. Although the TOGT suffers from inherent smoothness, leveraging the spatial potential of gates allows it to produce almost the same flight time as in the CPC.
• Figure 5: Comparison of trajectories obtained from the TOGT and CPC in the task of traversing 19 square gates located at 7 separate locations. Note that the gate in the lower-left corner comprises two vertically stacked gates. The trajectories are colored based on their speed profiles.