AI-IO: An Aerodynamics-Inspired Real-Time Inertial Odometry for Quadrotors
Jiahao Cui, Feng Yu, Linzuo Zhang, Yu Hu, Danping Zou
AI summary
Problem
Current learning-based inertial odometry for quadrotors ignores aerodynamic drag and rotor speed, leading to poor observability, limited accuracy, and weak generalization on agile flights.
Approach
The authors derive a physics-based aerodynamic model proving rotor speed is essential for velocity observability, then design a lightweight transformer that fuses IMU and rotor speed data to predict velocity and uncertainty in real time.
Key results
- 36.9% velocity prediction error reduction from rotor speed inputs
- 22.4% accuracy gain over prior methods using a transformer backbone
- Release of a new high-speed quadrotor flight dataset with IMU and rotor speed telemetry
- Real-time deployment on physical drones demonstrating superior accuracy and robustness
Why it matters
Provides a physically grounded, real-time state estimation framework that improves autonomous drone navigation in GPS-denied or visually degraded environments.
Abstract
Inertial Odometry (IO) has gained attention in quadrotor applications due to its sole reliance on inertial measurement units (IMUs), attributed to its lightweight design, low cost, and robust performance across diverse environments. However, most existing learning-based inertial odometry sys- tems for quadrotors either use only IMU data or include additional dynamics-related inputs such as thrust, but still lack a principled formulation of the underlying physical model to be learned. This lack of interpretability hampers the model’s ability to generalize and often limits its accuracy. In this work, we approach the inertial odometry learning problem from a different perspective. Inspired by the aerodynamics model and IMU measurement model, we identify the key phys- ical quantity—rotor speed measurements required for inertial odometry and design a transformer-based inertial odometry. By incorporating rotor speed measurements, the proposed model improves velocity prediction accuracy by 36.9%. Furthermore, the transformer architecture more effectively exploits temporal dependencies for denoising and aerodynamic modeling, yielding an additional 22.4% accuracy gain over previous results. To support evaluation, we also provide a real-world quadrotor flight dataset capturing IMU measurements and rotor speed for high-speed motion. Finally, combined with an uncertainty- aware extended Kalman filter (EKF), our framework is vali- dated across multiple datasets and real-time systems, demon- strating superior accuracy, generalization, and real-time perfor- mance. We share the code and data to promote further research (https://github.com/SJTU-ViSYS-team/AI-IO).