TEeVTOL: Balancing Energy and Time Efficiency in eVTOL Aircraft Path Planning Across City-Scale Wind Fields

Abstract

Electric vertical-takeoff and landing (eVTOL) aircraft, recognized for their maneuverability and flexibility, offer a promising alternative to our transportation system. However, the operational effectiveness of these aircraft faces many challenges, such as the delicate balance between energy and time efficiency, stemming from unpredictable environmental factors, including wind fields. Mathematical modeling-based approaches have been adopted to plan aircraft flight path in urban wind fields with the goal to save energy and time costs. While effective, they are limited in adapting to dynamic and complex environments. To optimize energy and time efficiency in eVTOL's flight through dynamic wind fields, we introduce a novel path planning method leveraging deep reinforcement learning. We assess our method with extensive experiments, comparing it to Dijkstra's algorithm -- the theoretically optimal approach for determining shortest paths in a weighted graph, where weights represent either energy or time cost. The results show that our method achieves a graceful balance between energy and time efficiency, closely resembling the theoretically optimal values for both objectives.

Figures (5)

• Figure 1: Path planning of eVTOL aircraft through city-scale wind fields enabled by deep reinforcement learning. For illustration purposes, the wind field is shown in 2D.
• Figure 2: For stage training, we divide the urban environment into 18 areas. Origin-destination pairs are sampled within these areas and are categorized into three classes: near-distance, mid-distance, and far-distance. During training, the eVTOL aircraft will learn to master near distances first, then move on to further distances.
• Figure 3: City-scale wind field simulation. 1. We first create a detailed 3D model of an urban area in SideFX: Houdini Houdini. 2. The 3D model is then imported into Unreal Engine Unreal to create an immersive simulation environment. 3. Next, the simulation environment is imported into OpenFOAM OpenFOAM via scripting. This allows for the visualization of the environment in Paraview Paraview. 4. The result of wind field simulation is visualized as volumetric rendering of velocity field magnitude.
• Figure 4: Training performance comparisons. LEFT: The addition of stage training (ST) enables significant improvement over its counterpart. MIDDLE: Our method starts to outperform vanilla PPO Schulman2017 starting around the 6000$th$ episode. RIGHT: Our approach allows the eVTOL aircraft to reach its destination far earlier than vanilla PPO. These results demonstrate the effectiveness of our algorithm design.
• Figure 5: Example eVTOL aircraft flight paths during wind field D0-4 (first column), D90-4 (second column), and D180-4 (third column). The subfigures in the first row show the comparisons of the paths from Ours-energy and Ours-all. The two sets of paths are highly similar indicating that Ours-all is effective in minimizing the energy consumption. The subfigures in the second row compare the paths from Ours-time and Ours-all. In these examples, Ours-all tends to opt for a different path, leveraging wind fields to achieve reduced energy consumption without significantly increasing travel times, as opposed to the paths chosen by Ours-time.