Outstanding Research Organisation Finalists

AMO – UC San Diego

Recognised as one of the top 20 research universities worldwide, UC San Diego offers a wealth of AI facilities including the UCSD DER Connect Research Facility and the Materials Research Science and Engineering Center Materials Characterization Facility. Researchers at UC San Diego developed a framework called AMO that empowers hyper-dexterous humanoid robots to perform complex tasks in real-world scenarios. AMO uses reinforcement learning and trajectory optimisation to enable the humanoid to achieve autonomous task execution through imitation learning. 

DexWrist – Massachusetts Institute of Technology

The Massachusetts Institute of Technology (MIT) is a private research university located in Cambridge, Massachusetts. MIT’s robotics research spans multiple world-class labs, including CSAIL, LIDS, and the Media Lab, where interdisciplinary teams explore humanoid control, autonomous systems, and human-robot interaction. Researchers from MIT developed DexWrist, a robotic wrist designed for human-like dexterous manipulation, enabling faster data collection and more reliable policies. By mimicking the kinematics of the human wrist, DexWrist allows robots to operate effectively in challenging environments. 

HITTER – University of California Berkeley

The University of California is ranked as one of the US’ top public universities. There are many facilities at UC Berkeley including the Autonomy, Robotics, and Controls (ARC) Lab and the Human Engineering Laboratory. Researchers developed HITTER, a hierarchical framework for humanoid table tennis that integrates a model-based planner for ball trajectory prediction and racket target planning. Trained via a reinforcement learning-based whole-body controller, it allows the humanoid to mimic human strikes and maintain stability and agility across consecutive rallies. 

MoRE – ShanghaiTech University

Founded in 2013, ShanghaiTech is a university with a modern campus in the heart of Shanghai Pudong’s Zhangjiang Hi-Tech Park. ShanghaiTech, in collaboration with the University of Science and Technology of China, the Institute of Artificial Intelligence and China Telecom (TeleAI), and Harbin Engineering University have unveiled MoRE, a novel framework that enables humanoid robots to transition seamlessly between diverse human-like gait patterns and navigate complex terrains with agility. 

NADIA – IHMC

The Institute for Human & Machine Cognition (IHMC) is a not-for-profit research institute of the Florida University System. At IHMC, they have developed Nadia, a hydraulic humanoid robot designed to serve as a new platform for research in locomotion and control. Researchers at IHMC developed Anticipatory and Adaptive Footstep Streaming for Teleoperated Bipedal Robots. It retargets user steps to robot footstep locations, allowing the robot to utilise its own dynamics for locomotion, ensuring better balance and stability. The method anticipates user footsteps to minimise delays between when the user initiates and completes a step and when the robot does it. Additionally, the system autonomously adjusts the robot’s steps to account for its surrounding terrain, overcoming challenges posed by environmental mismatches between the user’s flat-ground setup and the robot’s uneven terrain. 

Telemanipulation of Multi-Arm Systems – Istituto Italiano di Tecnologia

The Istituto Italiano di Tecnologia (IIT) is a foundation established jointly by the Italian Ministry of Education, Universities and Research and the Ministry of Economy and Finance. IIT is headquartered in Genoa and includes 14 centers and over 80 research units across Italy. IIT has authored a research paper on the telemanipulation of multi-robotic arm systems. The paper discusses the development of a human-in-the-loop framework that aims to control multiple robotic arms with different levels of shared autonomy. This is achieved by combining EMG and Xsens Health & Sports Movella motion capture data. 

ULC – Harbin Institute of Technology

Harbin Institute of Technology (HIT) has established 31 national-level research platforms, including 17 key laboratories and nine national engineering laboratories and research centers. HIT’s Robotics Research Institute is a core hub for humanoid robotics, autonomous systems, and intelligent control. Researchers at Harbin Institute of Technology collaborated with Westlake University to create ULC, a framework that enables precise, robust whole-body control in humanoid robots using a single policy for movement and manipulation. The unified dual-arm tracking enables precise manipulation under external loads while maintaining coordinated whole-body control for complex loco-manipulation tasks. 

VisualMimic – Stanford University

Stanford University’s robotics center hosts a range of specialised labs, including the Vision and Learning Lab, the Intelligence through Robotic Interaction at Scale Lab, and the Autonomous Systems Lab. Researchers at Stanford have developed VisualMimic, a visual sim-to-real framework that unifies egocentric vision with hierarchical whole-body control for humanoid robots. VisualMimic features a task-agnostic, low-level keypoint tracker trained on human motion data through a teacher-student learning approach. The framework enables zero-shot transfer of visuomotor policies trained in simulation to real humanoid robots, allowing them to perform a wide range of loco-manipulation tasks. 

ViTacFormer – Peking University

Founded in 1898, Peking University (PKU) has established four research institutes and 18 AI+ interdisciplinary research centers in collaboration with 28 schools and departments on campus. Some of these include the Institute of General Artificial Intelligence and the AI+Medical Research Institute. Researchers at PKU, in collaboration with UC Berkeley, developed ViTacFormer, which integrates active vision and high-resolution tactile sensing to enable dexterous robotic manipulation. ViTacFormer demonstrates high effectiveness in long-horizon tasks, maintaining continuous, high-precision control with an anthropomorphic hand for approximately 2.5 minutes. 

Whole-Body Throwing with High-Frequency Residual Policy – ETH Zürich

ETH Zurich consistently ranks among the top universities worldwide for engineering and computer science. One of its key labs is the Robotic Systems Lab, which investigates the development of machines and their intelligence to operate in rough and challenging environments. Researchers at the Robotic Systems Lab developed a unified reinforcement learning approach that enables legged mobile manipulators to play badminton and quadrupeds and humanoids to traverse challenging terrain.