Model-Predictive Planning and Airspeed Regulation to Minimize Flight Energy Consumption

TL;DR

The paper addresses real-time energy optimization for electric multicopters along a corridor-constrained path by integrating a gradient-based PID airspeed regulator for cruise with an MPC framework for takeoff and landing. It formulates an energy-per-meter objective $EPM$ and analyzes downwash approximations to support accurate energy estimation, while modeling quad-rotor dynamics and employing trajectory optimization via IPOPT. The key contributions are a real-time $V_a$ control law that converges to $V_a^*$ without full system dynamics, an MPC-based energy-minimizing takeoff/landing planner with a landing incentive, and a comparative assessment of downwash models. The simulated results demonstrate significant energy efficiency gains, horizon-insensitive performance due to landing incentives, and practical implications for energy-aware urban drone operations under air-corridor regulations.

Abstract

Although battery technology has advanced tremendously over the past decade, it continues to be a bottleneck for the mass adoption of electric aircraft in long-haul cargo and passenger delivery. The onboard energy is expected to be utilized in an efficient manner. Energy concumption modeling research offers increasingly accurate mathematical models, but there is scant research pertaining to real-time energy optimization at an operational level. Additionally, few publications include landing and take-off energy demands in their governing models. This work presents fundamental energy equations and proposes a proportional-integral-derivative (PID) controller. The proposed method demonstrates a unique approach to an energy consumption model that tracks real-time energy optimization along a predetermined path. The proposed PID controller was tested in simulation, and the results show its effectiveness and accuracy in driving the actual airspeed to converge to the optimal velocity without knowing the system dynamics. We also propose a model-predictive method to minimize the energy usage in landing and take-off by optimizing the flight trajectory.

Figures (10)

• Figure 1: Top-down view of the sample route between the Three Crosses Regional Hospital (left) and the Memorial Medical Center (right) in Las Cruces, NM, USA.
• Figure 2: A sample route through a conceptualized air corridor between two hospitals in Las Cruces, NM, USA.
• Figure 3: EPM results using the Root ($w_R$), Hover ($W_H$), and Glauert ($W_G$) approximations for the downwash coefficient, respectively.
• Figure 4: Airspeed vs range (left) and Airspeed vs EPM (right) for a small drone with parameters in Table \ref{['tbl:drone_parameters']}.
• Figure 5: Block diagram for the proposed realtime airspeed controller.