Beijing Universities Forge Ahead in Humanoid Robotics with Historic Half-Marathon Triumph
The 2026 Beijing Yizhuang Half-Marathon marked a pivotal moment for China's higher education landscape, as students from premier Beijing universities showcased their ingenuity by fielding humanoid robots in the world's second annual Humanoid Robot Half-Marathon. Held on April 19 in the bustling Beijing Economic-Technological Development Area (E-Town), this event blended human endurance with cutting-edge artificial intelligence (AI) and robotics, drawing over 12,000 human runners and more than 100 robot teams. What stood out was the prominent role of university students, whose hands-on contributions highlighted the growing synergy between academia and industry in advancing embodied intelligence technologies.
Organized by local authorities and the Beijing Humanoid Robot Innovation Center, the 21.0975-kilometer race tested robots' speed, stability, autonomous navigation, and adaptability to real-world conditions like uneven terrain and obstacles. While commercial teams like Honor dominated the podium—claiming the top three spots with record-breaking times under 51 minutes—university squads from institutions such as Peking University, Beihang University, and Beijing Institute of Technology demonstrated remarkable progress, completing the course and earning accolades for gait and design.
Student Teams from Top Beijing Institutions Take Center Stage
Over a quarter of the competing teams hailed from universities and research institutes, underscoring Beijing's status as a robotics hub. Peking University (PKU) entered the Machine Synesthesia Team from its School of Electronics' Ubiquitous Connectivity and Networked Intelligence Lab (PCNI), led by Professor Cheng Xiang. Using the TianGong 3.0 platform from the Beijing Humanoid Robot Innovation Center, they iterated running algorithms incorporating machine synesthesia technology for enhanced sensory-motor integration, achieving a stable finish.
Beihang University (BUAA), renowned for aerospace engineering, fielded the 'Future Machine Domain - Intelligent Lingfeng Team' in collaboration with the Innovation Center. Guided by Associate Professor Tao Yong and Postdoc Ren Fan, the squad—including master's students Tao Yu and Chen Yixian, and undergraduate Wei Leyao from the Embodied Intelligent Robot Project—modified the TianGong ULTRA2026 robot. Employing the Tienkunglab framework for simulation-to-real (Sim2Real) adaptation, they navigated motor overheating and gait instability to complete the race unscathed. Wei Leyao noted, 'This was a test of endurance and teamwork, deepening my grasp of structural design and motion control.'
Beijing Institute of Technology (BIT) 'Beli Zhixingzhe' team from the School of Mechatronical Engineering's Cluster Game and Cooperative Control group secured second place in the best gait evaluation. Their TianGong-based robot, developed in just one month from simulation to deployment, exemplified rapid prototyping skills honed in project-based learning.
Innovative Preparations: From Training Camps to Real-World Deployment
Preparation was rigorous, centered at the Beijing Humanoid Robot Innovation Center's half-marathon training camp launched in late March. The center delivered 15 'Embodied TianGong' robots to teams from PKU, BUAA, BIT, Huazhong University of Science and Technology, and even international partners like Munich Technical University. Students engaged in high-intensity drills at venues like the 'Ice Ribbon' speed skating arena, focusing on long-distance stability, battery management, and environmental perception.
Beijing University of Posts and Telecommunications (BUPT) 'Youbing Kuai Pao' team, mentored by Professor Li Duanling, optimized a Unitree G1 platform ('Youbing No.1') for gait control and energy efficiency, tackling 10km+ challenges. China Agricultural University (CAU) teams—'Qinghe', 'Yunchuan', and 'Shanhai'—collaborated with Accelerate Evolution Co., blending mechanical, agricultural engineering, and AI expertise to hit 4m/s speeds with natural gaits. CAU's cross-disciplinary approach exemplifies how universities are fostering versatile robotics talent.
Vocational institutions like Beijing Information Science and Technology University (BISTU) 'Yuejin Zhe' (12 students under Liu Xiangquan) and Beijing Polytechnic University ('0306 Little Giant' veterans) brought agile, compact designs, proving that practical skills from applied programs rival elite research.
Technological Leaps: Student-Driven Advances in Embodiment and AI
These student efforts spotlighted breakthroughs in humanoid robotics: Sim2Real transfer for gait optimization, end-to-end visual-language-action (VLA) models, and adaptive control for dynamic environments. BUAA's micro-tuning enabled real-time stability, while BIT's framework pushed decision-making in unstructured settings. PKU's synesthesia tech improved sensory fusion, crucial for future applications like disaster response.
Compared to 2025's inaugural event—where only six of 21 robots finished, with champion TianGong Ultra at 2:40:42—2026 saw nearly half autonomous, podium times slashing human records (e.g., Honor Lightning at 50:26 vs. human WR 57:20). University teams' completion rates reflected curriculum emphasis on endurance testing and interdisciplinary integration.
Project-Based Learning: Cultivating Tomorrow's Robotics Pioneers
This race embodied China's National Education Plan's push for 'project bases' in STEM, where students apply theory via industry partnerships. BUAA integrated it into their Embodied Intelligence curriculum, BIT into cluster control research, CAU into agrotech-AI fusion. Mentors like Hu Biao at CAU praised how it built 'engineering literacy' through debugging marathons.
Challenges abounded: overheating, battery drain, obstacle avoidance. Yet, as Chen Yixian (BUAA) reflected, 'It solidified our resolve for embodied AI.' Such experiential learning addresses China's robotics talent gap, with universities producing graduates ready for firms like Unitree and UBTech.
Broader Impacts on Chinese Higher Education and Industry
The event aligns with Beijing's '15th Five-Year Plan' for 100,000 new tech degrees, positioning universities as innovation engines. Robotics enrollment surges 30% yearly at Tsinghua and PKU, with specialized labs like BUAA's. It fosters prod-acad-research ecosystems, as seen in Innovation Center collaborations.
Economically, humanoid market projected at $38B by 2030; student innovations accelerate commercialization in manufacturing, eldercare. Industry experts note tech transfers like Honor's cooling from student-tested endurance.
Student Voices: Inspiration Amid Technological Marvels
BUPT's Chu Tianqi, a spectator-turned-aspirant, marveled: 'AI's posture is impressive; resistance means obsolescence.' Young Guo Yukun aims for robotics uni post-event. Participants like Tao Yu envision 'prod-acad linkages yielding breakthroughs.'
Tsinghua's Zhao Minguo highlighted autonomy gains: 'Validates robots' 'diamond' capabilities.' These narratives reveal motivation fueling China's higher ed robotics boom.
Future Horizons: Scaling University-Led Robotics Innovation
Plans for 2027 include tougher terrains, multi-robot swarms. Universities eye VLA integration, ethical AI. With E-Town's subsidies, Beijing aims global leadership. For students, it's career launchpads: BIT grads to startups, BUAA to aerospace bots.
As humanoid tech permeates, Beijing universities exemplify how higher ed drives national strategy, blending youth passion with rigorous training for an AI-powered future.
For those eyeing robotics careers in China, explore opportunities at top institutions via specialized programs in embodied AI and mechatronics.
%20logo%20(1).png&w=128&q=75)
