{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T18:47:08Z","timestamp":1792003628354,"version":"4.4.0"},"reference-count":69,"publisher":"American Association for the Advancement of Science (AAAS)","issue":"119","funder":[{"DOI":"10.13039\/100010661","name":"Horizon 2020 Framework Programme","doi-asserted-by":"publisher","award":["101119774"],"award-info":[{"award-number":["101119774"]}],"id":[{"id":"10.13039\/100010661","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.science.org"],"crossmark-restriction":true},"short-container-title":["Sci. Robot."],"published-print":{"date-parts":[[2026,10,14]]},"abstract":"<jats:p>Autonomous drone racing represents a major frontier in robotics research. It requires an artificial intelligence (AI) that can run onboard lightweight flying robots under tight resource and time constraints while pushing the physical system to its limits. The state of the art in this area consists of a system with a stereo camera and an inertial measurement unit (IMU) that beat human drone-racing champions in a controlled indoor environment. Here, we present MonoRace, an onboard drone-racing approach that generalizes to a competition environment without any external motion tracking system, using only a monocular rolling-shutter camera and an IMU. The approach features robust state estimation that combines neural network\u2013based gate segmentation with a drone model. Moreover, it includes an offline optimization procedure that leverages the known geometry of gates to refine any state estimation parameter. This offline optimization is based purely on onboard flight data and is important for fine-tuning the extrinsic camera calibration parameters. Furthermore, the guidance and control are performed by a neural network that forgoes inner-loop controllers by directly sending motor commands. This small network runs on the flight controller at 500 hertz. The proposed approach won the Abu Dhabi Autonomous Drone Racing Competition 2025, outperforming all competing autonomous drone race teams and three world champion human pilots in a direct knockout tournament. It reached speeds up to 100 kilometers per hour on the competition track and successfully coped with problems such as camera interference and IMU saturation.<\/jats:p>","DOI":"10.1126\/scirobotics.aee9250","type":"journal-article","created":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T17:59:57Z","timestamp":1792000797000},"update-policy":"https:\/\/doi.org\/10.34133\/aaas_crossmark","source":"Crossref","is-referenced-by-count":0,"title":["MonoRace: Robust monocular AI for champion-level drone racing"],"prefix":"10.1126","volume":"11","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5207-4771","authenticated-orcid":true,"given":"Stavrow A.","family":"Bahnam","sequence":"first","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5158-565X","authenticated-orcid":true,"given":"Robin","family":"Ferede","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0006-0881-1002","authenticated-orcid":true,"given":"Till M.","family":"Blaha","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-9002-5710","authenticated-orcid":true,"given":"Anton E.","family":"Lang","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Erin","family":"Lucassen","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-3318-6977","authenticated-orcid":true,"given":"Quentin","family":"Missinne","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-9295-0998","authenticated-orcid":true,"given":"Aderik E. C.","family":"Verraest","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6795-8454","authenticated-orcid":true,"given":"Christophe","family":"De Wagter","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8265-1496","authenticated-orcid":true,"given":"Guido C. H. E.","family":"de Croon","sequence":"additional","affiliation":[{"name":"Control &amp; Operation, Delft University of Technology, Delft 2629 HS, Netherlands."}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"221","reference":[{"key":"e_1_3_2_2_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11370-018-00271-6"},{"key":"e_1_3_2_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2024.3400838"},{"key":"e_1_3_2_4_2","doi-asserted-by":"crossref","unstructured":"S. Sun R. Schilder C. C. de Visser \u201cIdentification of quadrotor aerodynamic model from high speed flight data\u201d in 2018 AIAA Atmospheric Flight Mechanics Conference (AIAA 2018) p. 0523.","DOI":"10.2514\/6.2018-0523"},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1002\/(SICI)1099-1239(199809)8:11<995::AID-RNC373>3.0.CO;2-W"},{"key":"e_1_3_2_6_2","doi-asserted-by":"crossref","unstructured":"D. Mellinger V. Kumar \u201cMinimum snap trajectory generation and control for quadrotors\u201d in 2011 IEEE International Conference on Robotics and Automation (IEEE 2011) pp. 2520\u20132525.","DOI":"10.1109\/ICRA.2011.5980409"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2017.2776353"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1109\/TCST.2020.3001117"},{"key":"e_1_3_2_9_2","doi-asserted-by":"publisher","DOI":"10.3390\/aerospace4020031"},{"key":"e_1_3_2_10_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10846-020-01250-9"},{"key":"e_1_3_2_11_2","doi-asserted-by":"crossref","unstructured":"C. Liu H. Lu W.-H. Chen \u201cAn explicit MPC for quadrotor trajectory tracking\u201d in 2015 34th Chinese Control Conference (CCC) (IEEE 2015) pp. 4055\u20134060.","DOI":"10.1109\/ChiCC.2015.7260264"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2021.3061307"},{"key":"e_1_3_2_13_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2022.3173711"},{"key":"e_1_3_2_14_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2022.3185772"},{"key":"e_1_3_2_15_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2022.3177279"},{"key":"e_1_3_2_16_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2021.3131690"},{"key":"e_1_3_2_17_2","doi-asserted-by":"crossref","unstructured":"K. Nguyen S. Schoedel A. Alavilli B. Plancher Z. Manchester \u201cTinyMPC: Model-predictive control on resource-constrained microcontrollers\u201d in 2024 IEEE International Conference on Robotics and Automation (ICRA) (IEEE 2024) pp. 1\u20137.","DOI":"10.1109\/ICRA57147.2024.10610987"},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1126\/scirobotics.adg1462"},{"key":"e_1_3_2_19_2","doi-asserted-by":"crossref","unstructured":"Y. Song M. Steinweg E. Kaufmann D. Scaramuzza \u201cAutonomous drone racing with deep reinforcement learning\u201d in 2021 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE 2021) pp. 1205\u20131212.","DOI":"10.1109\/IROS51168.2021.9636053"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2022.3181755"},{"key":"e_1_3_2_21_2","doi-asserted-by":"crossref","unstructured":"E. Kaufmann L. Bauersfeld D. Scaramuzza \u201cA benchmark comparison of learned control policies for agile quadrotor flight\u201d in 2022 International Conference on Robotics and Automation (ICRA) (IEEE 2022) pp. 10504\u201310510.","DOI":"10.1109\/ICRA46639.2022.9811564"},{"key":"e_1_3_2_22_2","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-023-06419-4"},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.robot.2023.104588"},{"key":"e_1_3_2_24_2","doi-asserted-by":"publisher","DOI":"10.1126\/scirobotics.adi6421"},{"key":"e_1_3_2_25_2","doi-asserted-by":"crossref","unstructured":"R. Ferede C. De Wagter D. Izzo G. C. de Croon \u201cEnd-to-end reinforcement learning for time-optimal quadcopter flight\u201d in 2024 IEEE International Conference on Robotics and Automation (ICRA) (IEEE 2024) pp. 6172\u20136177.","DOI":"10.1109\/ICRA57147.2024.10611665"},{"key":"e_1_3_2_26_2","doi-asserted-by":"crossref","unstructured":"R. Ferede T. Blaha E. Lucassen C. De Wagter G. C. de Croon \u201cOne net to rule them all: Domain randomization in quadcopter racing across different platforms\u201d in 2025 IEEE International Conference on Robotics and Automation (ICRA) (IEEE 2025) pp. 6357\u20136363.","DOI":"10.1109\/ICRA55743.2025.11128790"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2024.3396025"},{"key":"e_1_3_2_28_2","doi-asserted-by":"publisher","DOI":"10.1126\/scirobotics.aec1481"},{"key":"e_1_3_2_29_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2025.3577037"},{"key":"e_1_3_2_30_2","unstructured":"S. M. Hegre W. Rehberg M. Kulkarni K. Alexis A neural network mode for PX4 on embedded flight controllers. arXiv: 2505.00432 [cs.RO] (2025)."},{"key":"e_1_3_2_31_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2019.2942989"},{"key":"e_1_3_2_32_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2023.3339543"},{"key":"e_1_3_2_33_2","doi-asserted-by":"publisher","DOI":"10.1177\/02783649251401223"},{"key":"e_1_3_2_34_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2024.3422836"},{"key":"e_1_3_2_35_2","doi-asserted-by":"publisher","DOI":"10.1109\/TRO.2025.3644945"},{"key":"e_1_3_2_36_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2026.3653405"},{"key":"e_1_3_2_37_2","doi-asserted-by":"crossref","unstructured":"Y. Xu G. C. de Croon \u201cCNN-based ego-motion estimation for fast MAV maneuvers\u201d in 2021 IEEE International Conference on Robotics and Automation (ICRA) (IEEE 2021) pp. 7606\u20137612.","DOI":"10.1109\/ICRA48506.2021.9561714"},{"key":"e_1_3_2_38_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.robot.2024.104866"},{"key":"e_1_3_2_39_2","doi-asserted-by":"crossref","unstructured":"S. A. Bahnam C. De Wagter G. C. de Croon \u201cSelf-supervised monocular visual drone model identification through improved occlusion handling\u201d in 2025 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS) (IEEE 2025) pp. 18977\u201318984.","DOI":"10.1109\/IROS60139.2025.11247627"},{"key":"e_1_3_2_40_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.robot.2020.103621"},{"key":"e_1_3_2_41_2","doi-asserted-by":"publisher","DOI":"10.55417\/fr.2022042"},{"key":"e_1_3_2_42_2","doi-asserted-by":"crossref","unstructured":"I. Geles L. Bauersfeld A. Romero J. Xing D. Scaramuzza \u201cDemonstrating agile flight from pixels without state estimation\u201d in Proceedings of Robotics: Science and Systems D. Kulic G. Venture K. Bekris E. Coronado Eds. (RSS Foundation 2024); 10.15607\/RSS.2024.XX.082.","DOI":"10.15607\/RSS.2024.XX.082"},{"key":"e_1_3_2_43_2","unstructured":"J. Xing A. Romero L. Bauersfeld D. Scaramuzza Bootstrapping reinforcement learning with imitation for vision-based agile flight. arXiv: 2403.12203 [cs.RO] (2024)."},{"key":"e_1_3_2_44_2","doi-asserted-by":"crossref","unstructured":"A. Romero A. Shenai I. Geles E. Aljalbout D. Scaramuzza Dream to fly: Model-based reinforcement learning for vision-based drone flight. arXiv: 2501.14377 [cs.RO] (2025).","DOI":"10.1109\/ICRA57385.2026.11695931"},{"key":"e_1_3_2_45_2","unstructured":"M. Krinner E. Aljalbout A. Romero D. Scaramuzza Accelerating model-based reinforcement learning with state-space world models. arXiv: 2502.20168 [cs.RO] (2025)."},{"key":"e_1_3_2_46_2","doi-asserted-by":"crossref","unstructured":"Z. Liu H. Hu Y. Liu Z. Yao Z. Xie Y. Wei J. Ning Y. Cao Z. Zhang L. Dong F. Wei B. Guo \u201cSwin Transformer V2: Scaling up capacity and resolution\u201d in 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (IEEE 2022) pp. 11999\u201312009.","DOI":"10.1109\/CVPR52688.2022.01170"},{"key":"e_1_3_2_47_2","doi-asserted-by":"publisher","DOI":"10.1109\/MRA.2016.2646090"},{"key":"e_1_3_2_48_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10514-021-10011-y"},{"key":"e_1_3_2_49_2","doi-asserted-by":"publisher","DOI":"10.1038\/d41586-023-02506-8"},{"key":"e_1_3_2_50_2","unstructured":"M. B. Azhari D. Han J. I. You S. Park D. H. Shim Drift-corrected monocular VIO and perception-aware planning for autonomous drone racing. arXiv: 2512.20475 [cs.RO] (2025)."},{"key":"e_1_3_2_51_2","unstructured":"F. Novak M. Petrl\u00edk M. Novosad P. M. Gupta R. P\u011bni\u010dka M. Saska Vision-only UAV state estimation for fast flights without external localization systems: A2RL Drone Racing finalist approach. arXiv: 2602.01860 [cs.RO] (2026)."},{"key":"e_1_3_2_52_2","doi-asserted-by":"publisher","DOI":"10.2514\/1.G002357"},{"key":"e_1_3_2_53_2","unstructured":"LiveTime LLC LiveTime Scoring System (2026); https:\/\/livetimescoring.com\/."},{"key":"e_1_3_2_54_2","doi-asserted-by":"crossref","unstructured":"D. Simon Optimal State Estimation: Kalman H Infinity and Nonlinear Approaches (John Wiley & Sons 2006).","DOI":"10.1002\/0470045345"},{"key":"e_1_3_2_55_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10339-012-0519-z"},{"key":"e_1_3_2_56_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2019.2930480"},{"key":"e_1_3_2_57_2","unstructured":"J. Snoek H. Larochelle R. P. Adams \u201cPractical Bayesian optimization of machine learning algorithms\u201d in Advances in Neural Information Processing Systems vol. 25 F. Pereira C. J. Burges L. Bottou K. Weinberger Eds. (Curran Associates 2012) pp. 2951\u20132959."},{"key":"e_1_3_2_58_2","doi-asserted-by":"crossref","unstructured":"P. Furgale J. Rehder R. Siegwart \u201cUnified temporal and spatial calibration for multi-sensor systems\u201d in 2013 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IEEE 2013) pp. 1280\u20131286.","DOI":"10.1109\/IROS.2013.6696514"},{"key":"e_1_3_2_59_2","doi-asserted-by":"publisher","DOI":"10.1145\/358669.358692"},{"key":"e_1_3_2_60_2","doi-asserted-by":"crossref","unstructured":"H. Wang J. Xing N. Messikommer D. Scaramuzza \u201cEnvironment as policy: Learning to race in unseen tracks\u201d in 2025 IEEE International Conference on Robotics and Automation (ICRA) (IEEE 2025) pp. 11333\u201311339.","DOI":"10.1109\/ICRA55743.2025.11127376"},{"key":"e_1_3_2_61_2","doi-asserted-by":"publisher","DOI":"10.1109\/LRA.2022.3145063"},{"key":"e_1_3_2_62_2","unstructured":"J. Schulman F. Wolski P. Dhariwal A. Radford O. Klimov Proximal policy optimization algorithms. arXiv: 1707.06347 [cs.LG] (2017)."},{"key":"e_1_3_2_63_2","doi-asserted-by":"crossref","unstructured":"O. Ronneberger P. Fischer T. Brox \u201cU-Net: Convolutional networks for biomedical image segmentation\u201d in Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2015 vol. 9351 of Lecture Notes in Computer Science (Springer 2015) pp. 234\u2013241.","DOI":"10.1007\/978-3-319-24574-4_28"},{"key":"e_1_3_2_64_2","unstructured":"X. Glorot Y. Bengio \u201cUnderstanding the difficulty of training deep feedforward neural networks\u201d in Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics vol. 9 of Proceedings of Machine Learning Research Y. W. Teh M. Titterington Eds. (MLResearchPress 2010) pp. 249\u2013256."},{"key":"e_1_3_2_65_2","doi-asserted-by":"crossref","unstructured":"I. Brouk A. Nemirovsky Y. Nemirovsky \u201cAnalysis of noise in CMOS image sensor\u201d in 2008 IEEE International Conference on Microwaves Communications Antennas and Electronic Systems (IEEE 2008) pp. 1\u20138.","DOI":"10.1109\/COMCAS.2008.4562800"},{"key":"e_1_3_2_66_2","doi-asserted-by":"publisher","DOI":"10.5201\/ipol.2012.gjmr-lsd"},{"key":"e_1_3_2_67_2","doi-asserted-by":"crossref","unstructured":"M. Muja D. Lowe \u201cFast approximate nearest neighbors with automatic algorithm configuration\u201d in VISAPP 2009 - Proceedings of the 4th International Conference on Computer Vision Theory and Applications vol. 1 (SciTePress 2009) pp. 331\u2013340.","DOI":"10.5220\/0001787803310340"},{"key":"e_1_3_2_68_2","doi-asserted-by":"crossref","unstructured":"J. Svacha K. Mohta V. Kumar \u201cImproving quadrotor trajectory tracking by compensating for aerodynamic effects\u201d in 2017 International Conference on Unmanned Aircraft Systems (ICUAS) (IEEE 2017) pp. 860\u2013866.","DOI":"10.1109\/ICUAS.2017.7991501"},{"key":"e_1_3_2_69_2","first-page":"1","article-title":"Stable-Baselines3: Reliable reinforcement learning implementations","volume":"22","author":"Raffin A.","year":"2021","unstructured":"A. Raffin, A. Hill, A. Gleave, A. Kanervisto, M. Ernestus, N. Dormann, Stable-Baselines3: Reliable reinforcement learning implementations. J. Mach. Learn. Res. 22, 1\u20138 (2021).","journal-title":"J. Mach. Learn. Res."},{"key":"e_1_3_2_70_2","unstructured":"S. Bahnam Data and code underlying the publication MonoRace: Robust monocular AI for champion-level drone racing 4TU.ResearchData (2026); https:\/\/doi.org\/10.4121\/039abd82-094c-4c93-baa1-e7446f562e0e."}],"container-title":["Science Robotics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.science.org\/doi\/pdf\/10.1126\/scirobotics.aee9250","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T18:00:41Z","timestamp":1792000841000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.science.org\/doi\/10.1126\/scirobotics.aee9250"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,14]]},"references-count":69,"journal-issue":{"issue":"119","published-print":{"date-parts":[[2026,10,14]]}},"alternative-id":["10.1126\/scirobotics.aee9250"],"URL":"https:\/\/doi.org\/10.1126\/scirobotics.aee9250","relation":{},"ISSN":["2470-9476"],"issn-type":[{"value":"2470-9476","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,14]]},"assertion":[{"value":"2025-12-19","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-09-15","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-10-14","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"eaee9250"}}