January 4th, 2022 Ruibin Zhang, Yuze Wu, Lixian Zhang, Chao Xu, Fei Gao’ s work on “Autonomous and Adaptive Navigation for Terrestrial-Aerial Bimodal Vehicles” is accepted by IEEE Robotics and Automation Letters (RA-L) for publication.
On July 28, the “UAV Competing” Intelligent UAV clustering system challenge round in 2021 came to an end in Laishui, Hebei Province. Fastlab of Zhejiang University participated in this competition for the first time and won the third place in the subject 3. group photo The competition began on July 18 and lasted for 10 days which consisted of two stages: Open Competition and Invitational Competition. 51 teams and more than 500 candidates from all walks of life. Five difficult subjects were set up in the preliminary competition : subject 1 (speed crossing), subject 2 (witnessed by the whole people), subject 3 (inseparable), subject 4 (air handshake) and subject 5 (peak duel – virtual and real flight). The teams of the fixed wing group and the non- fixed wing group, separately, competed more than 60 games, according to the rules. Fastlab of Zhejiang University, relying on Xisaishan (Huzhou) base, together with Hangzhou Institute of Beijing University of Aeronautics and Astronautics, jointly established the Sea Hawk CSC team. In order to prepare for this competition, the students of Fastlab worked hard for two months. Under the hot weather, sun exposure and mosquito bites, they did outdoor experiments again and again. the outdoor experiments After several rounds, via excellent technical strength and outstanding performance, Sea Hawk CSC team ranked the second place in semi-finals with 148 points and won the third place in the final. This is the first time Fastlab participated in this competition, and the result is impressive.
July 1st, 2021 Four conference papers are accepted by IROS 2021: FAST-Dynamic-Vision: Detection and Tracking Dynamic Objects with Event and Depth Sensing, Botao He, Haojia Li, Siyuan Wu, Dong Wang, Zhiwei Zhang, Qianli Dong, Chao Xu, Fei Gao. Autonomous Flights in Dynamic Environments with Onboard Vision, Yingjian Wang, Jialin Ji, Qianhao Wang, Chao Xu, Fei Gao. Visibility-aware Trajectory Optimization with Application to Aerial Tracking, Qianhao Wang, Yuman Gao, Jialin Ji, Chao Xu, Fei Gao. Learning-based 3D Occupancy Prediction for Autonomous Navigation in Occluded Environments, Lizi Wang, Hongkai Ye, Qianhao Wang, Yuman Gao, Chao Xu, Fei Gao.
Aerial swarm navigation has always been challenging in unknown, complex environments, especially without any external positioning and computing system. Recently, the FAST Lab of Zhejiang University has made further breakthroughs in this field. A distributed swarm trajectory planning algorithm was proposed in the latest technical report. It mainly features spatial-temporal optimization, which significantly improves the quality of generated trajectory and saves computation. Specifically, it only takes several milliseconds to calculate a local spatial-temporal optimal trajectory on embedded devices such as onboard computers. Furthermore, its complexity only increases linearly to the number of agents. Indoor and outdoor experiments verify the algorithm’s effectiveness, and the relevant results are reported by the authoritative scientific and technological media IEEE spectrum.
Fast-Tracker is a systematic solution that uses an unmanned aerial vehicle (UAV) to aggressively and safely track an agile target. The open-source project releases the tracking algorithm with a simulation of a car-tracking scenario as a demo. Author: Zhichao Han*, Ruibin Zhang*, Neng Pan*, Chao Xu and Fei Gao from the ZJU Fast Lab Related Paper: Fast-Tracker: A Robust Aerial System for Tracking AgileTarget in Cluttered Environments, Zhichao Han*, Ruibin Zhang*, Neng Pan*, Chao Xu, Fei Gao, International Conference on Robotics and Automation (ICRA 2021) Video Links: youtube or bilibili Code Link: https://github.com/ZJU-FAST-Lab/Fast-tracker External Forces Resilient Planner is a systematic motion planning framework that can resiliently generate safe trajectories under volatile external forces. This work integrates our previous work, VID-Fusion as the external force estimator. Author: Yuwei WU and Fei GAO from the ZJU Fast Lab Related Paper: External Forces Resilient Safe Motion Planning for Quadrotor, Yuwei Wu, Ziming Ding, Chao Xu, Fei Gao, submitted to IEEE Robotics and Automation Letter (RA-L). Video Links: youtube or bilibili Code Link: https://github.com/ZJU-FAST-Lab/forces_resilient_planner VID-Fusion is a work to estimate odometry and external force simultaneously by a tightly coupled Visual-Inertial-Dynamics state estimator. The open source project releases the code of VID-Fusion, along with an experimental dataset in the real word. Author: Ziming Ding, Tiankai Yang, Kunyi Zhang, Chao Xu, and Fei Gao from the ZJU FAST Lab. Related Paper: VID-Fusion: Robust Visual-Inertial-Dynamics Odometry for Accurate External Force Estimation, Ziming Ding, Tiankai Yang, Kunyi Zhang, Chao Xu, and Fei Gao, International Conference on Robotics and Automation (ICRA 2021) Video Links: Youtube or bilibili Code Link: https://github.com/ZJU-FAST-Lab/VID-Fusion
Recently, Dr. Fei Gao, the deputy director and Co-PI of the Field Autonomous System & Computing Lab (Fast Lab) of College of Control Science and Engineering, received Honorable Mention in the 2020 IEEE Transactions on Robotics (IEEE-TRO) King-Sun Fu Memorial Best Paper Award for the paper “Teach-Repeat-Replan: A Complete and Robust System for Aggressive Flight in Complex Environments” as the first and corresponding author. This is the first time that the colleges and universities of the Chinese Mainland have won the honor as the leading institution. It is also the first time that IEEE-TRO has awarded this honor to the papers published in IEEE-TRO from the UAV community. There is a massive market for consumer drones nowadays. However, most of the operators of consumer drones are not professional pilots and would struggle to generate their ideal trajectory for a long time. In some scenarios, such as drone racing or aerial filming, it is impossible for a beginner-level pilot to control the drone to finish the race safely or take an aerial video smoothly without months of training. In a drone racing competition, each quadrotor is controlled by a human pilot to fly through several gates towards the terminal as quickly as possible. In the racing flight, collisions must be avoided to ensure safety, while the flight aggressiveness is expected to be extremely high. However, it is hard for a human pilot to master the skill of balancing speed and safety. As opposed to drone racing, aerial filming/videography does not prefer […]
Dr. Fei Gao, the co-director of ZJU Field Autonomous System and Computing Laboratory (ZJU FAST Lab), received Honorable Mention in the 2020 IEEE Transactions on Robotics King-Sun Fu Memorial Best Paper Award for the paper “Teach-Repeat-Replan: A Complete and Robust System for Aggressive Flight in Complex Environments” as the first and corresponding author. This is the first time that a college or university of mainland China has won this honor as the first completion unit, and it is also the first time that IEEE-TRO has awarded this honor to a paper published in the field of unmanned aerial vehicles on IEEE-TRO. This paper presents a complete solution towards robust aerial autonomy, enabling a drone to accomplish a complicated task with professional performance under merely rough human operations. The human operator may provide an arbitrarily slow or jerky trajectory with an expected topological structure. The system then autonomously converts this poor teaching trajectory to a topologically equivalent and local optimal one. The aggressiveness of the generated repeating motions is tunable, which can meet speed requirements ranging from drone racing to aerial filming. Moreover, during the repeating flight, the system locally observes environmental changes and replans safe trajectories to avoid moving obstacles. The proposed system extends the classical robotics teach-and-repeat framework and is named as Teach-Repeat-Replan. By the time the news was released, this work had ranked among the top 50 popular articles in the IEEE-TRO database (ranked 14th) and received extensive attention and high praise from well-known scholars in the field […]
转载自：https://mp.weixin.qq.com/s/iR1BpSMP9aPPWbDfVw5eIA 近日，我院无人系统与自主计算实验室（FAST Lab）高飞老师以第一兼通讯作者身份所发表工作“Teach-Repeat-Replan: A Complete and Robust System for Aggressive Flight in Complex Environments”（以下简称TRR）获IEEE Transactions on Robotics (IEEE-TRO) 2020年度“傅京孙最佳论文荣誉奖”（King-Sun Fu Best Paper Award Honorable Mention）。这是中国大陆高校院所首次以第一完成单位获此殊荣，也是IEEE-TRO首次将此荣誉授予发表在IEEE-TRO上无人机领域的论文。 基于广泛应用于基础设施检查、空中运输、自主搜救等领域的经典“示教-复演”框架，该工作建立了一个完备且鲁棒的四旋翼无人机自主导航飞行系统。在经典“示教-复演”框架中，人类的意图对于决定无人机飞行轨迹的拓扑结构至关重要，然而用户输入的低效示教轨迹和不断变化的环境导致了“示教-复演”系统无法被灵活和稳健地应用。针对此问题，该研究首次提出无人机“示教-复演-重规划”概念，通过生成贴合环境障碍物的飞行走廊方式，生成平滑、安全、动力学可行并与示教路径拓扑等效的高质量飞行轨迹。同时，为了避免飞行过程中出现动态障碍物或定位飘移情况，该工作集成了快速局部感知和滑动窗口重规划方法，用于在线生成无人机局部避障轨迹。大量的室内外复杂场景实验结果，展现了该系统具备远超人类操作无人机水平的全自主无人机竞速性能。 截至消息发布前，该工作位列IEEE-TRO全数据库50大热门文章（排第14），并受到了国内外知名学者的广泛关注和高度评价。IEEE Transactions on Robotics总主编、美国西北大学教授Kevin Lynch（IEEE Fellow、Modern Robotics等著名机器人学教材作者）在授予该论文最佳论文荣誉奖时评价：“该工作在理论和实验上的贡献给IEEE-TRO编委会留下了深刻的印象（原文：‘The T-RO editorial board was impressed by the theoretical and experimental contributions of your work.’）。” IEEE-TRO是IEEE旗下的机器人领域TOP期刊，包含了机器人领域各方面的研究。IEEE-TRO对论文质量有着极高的要求，需要工作对机器人领域有着显著卓越的理论、实践贡献并能推动领域发展，而“傅京孙最佳论文荣誉奖”用以表彰每年IEEE-TRO发表论文中的最佳工作。因此，能在IEEE-TRO期刊上发表论文并同时获得“傅京孙最佳论文荣誉奖”是所有机器人相关领域工作者的极高荣誉。 相关链接 ● 论文官网链接 https://ieeexplore.ieee.org/document/9102390 ● 工作开源链接 https://github.com/HKUST-Aerial-Robotics/Teach-Repeat-Replan ● 工作视频链接 https://www.bilibili.com/video/BV1Fx411o78w?from=search&seid=4539658003138370390 ● FAST实验室链接 http://zju-fast.com