Aerial Core

We are thrilled to share our breakthrough research on "Agile Flight from Pixels without State Estimation," to be presented and live-demonstrated at #RSS2024 next week! You heard well: no state estimation means no explicit visual localization, no SLAM, no VIO, and no IMU! Paper: arxiv.org/pdf/2406.12505 Video (Narrated): youtu.be/a1MSkTD-Tl8 Last year, we demonstrated that #ReinforcementLearning (RL) policies could outperform world-champion drone-racing pilots using the same quadrotor hardware; however, unlike human pilots, these policies continuously estimated an explicit state from known gate positions, the camera feed, and inertial measurements (IMU). In this new work, we tackle the challenge of learning vision-based drone racing using an end-to-end reinforcement learning approach that eliminates the need for IMU data or explicit state estimation. Like professional pilots, we go directly from images to control commands. The training is facilitated by an asymmetric actor-critic with access to privileged information. To overcome the computational complexity during image-based RL training, we use an appropriate sensor representation, which can be efficiently simulated during training without rendering images. We achieve agile flight at speeds up to 40 km/h with accelerations up to 2 g's. Although our demonstration focuses on drone racing, we believe that our method has an impact beyond drone racing and can serve as a foundation for future research into real-world applications in structured environments. Besides the paper presentation, we will also give a live demo next Tuesday and Wednesday between and hrs at TU Delft: roboticsconference.org/progr… Reference: Ismail Geles*, Leonard Bauersfeld*, Angel Romero, Jiaxu Xing, Davide Scaramuzza "Demonstrating Agile Flight from Pixels without State Estimation" Robotics: Science and Systems (RSS), 2024. Kudos to @isgeles @l_bauersfeld @roaguiangel @jixing24! @UZH_en @UZH_Science @UZHspacehub @aerialcore @AUTOASSESS_EU @ERC_Research

Check out our #ICRA2024 paper, “Contrastive Learning for Enhancing Robust Scene Transfer in Vision-based Agile Flight,” where we present an approach to train agile vision-based flight policies that can adapt to and transfer across multiple unseen environments. We propose an adaptive contrastive learning approach to train a vision encoder that outputs task-related but environment-agnostic embedding and generalizes it to unseen environments without finetuning. Our contrastive learning approach aligns the visual embedding to ensure that similar task progress results in similar embeddings, while far task progress will yield distinct embeddings.  Our vision encoder is paired with a small action net, trained via privileged imitation learning purely in simulation. Our approach transfers to unseen environments where the action net has never been trained. Reference: Contrastive Learning for Enhancing Robust Scene Transfer in Vision-based Agile Flight IEEE International Conference on Robotics and Automation (ICRA), 2024 PDF: arxiv.org/pdf/2309.09865.pdf Video: youtu.be/4A4YyPgEWD8 Kudos to @jixing24, @l_bauersfeld, @realyunlong, @ChunweiXing! @aerialcore @ERC_Research @UZH_en @UZHspacehub @ieee_ras_icra @UZH_Science

Check out our #ICRA2024 paper "Contrastive Initial State Buffer for Reinforcement Learning," which tackles the sample inefficiency in #ReinforcementLearning head-on. Code released! We introduce an approach agnostic to the underlying RL algorithm: the Contrastive Initial State Buffer. This tool strategically selects states from past experiences and uses them to initialize the agent in the environment to guide it toward more informative states. Our experiments on drone racing and legged locomotion show that our method achieves higher task performance while also speeding up training convergence. Reference: Nico Messikommer, Yunlong Song, Davide Scaramuzza Contrastive Initial State Buffer for Reinforcement Learning IEEE International Conference on Robotics and Automation (ICRA), 2024. PDF: arxiv.org/pdf/2309.09752 Code: github.com/uzh-rpg/cl_initia… Video: youtu.be/RB7mDq2fhho Kudos to @Messikommer1 @realyunlong @aerialcore @ERC_Research @UZH_en @UZHspacehub @ieee_ras_icra @UZH_Science