Vision-based Control of a Quadrotor in User Proximity: Mediated vs\n End-to-End Learning Approaches

2018/09/24 by Dario Mantegazza, Jérôme Guzzi, Mantegazza, Dario +5
Computer Science · Neuroscience · #Computer Vision and Pattern Recognition (cs.CV) #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #Machine Learning (cs.LG) #Neural Networks and Reservoir Computing #Robotics (cs.RO) #Target Tracking and Data Fusion in Sensor Networks

paper · pdf · doi:10.48550/arxiv.1809.08881

openalex publication_date 2018/09/24 · openalex created_date 2022/08/03 · openalex updated_date 2026/07/28

Abstract

We consider the task of controlling a quadrotor to hover in front of a freely\nmoving user, using input data from an onboard camera. On this specific task we\ncompare two widespread learning paradigms: a mediated approach, which learns an\nhigh-level state from the input and then uses it for deriving control signals;\nand an end-to-end approach, which skips high-level state estimation altogether.\nWe show that despite their fundamental difference, both approaches yield\nequivalent performance on this task. We finally qualitatively analyze the\nbehavior of a quadrotor implementing such approaches.\n

Related

• Efficient Self-Supervised Neuro-Analytic Visual Servoing for Real-time Quadrotor Control
• Vision Transformers for End-to-End Vision-Based Quadrotor Obstacle Avoidance
• All Eyes, no IMU: Learning Flight Attitude from Vision Alone
• Unifying Foundation Models with Quadrotor Control for Visual Tracking Beyond Object Categories
• FalconGym: A Photorealistic Simulation Framework for Zero-Shot Sim-to-Real Vision-Based Quadrotor Navigation