➀ Northeastern University developed SCCRUB - a soft robotic arm capable of removing over 99% of hardened dirt through scrubbing, marking a breakthrough in household cleaning robots;

➁ Features planetary gear-driven bristle brush and machine learning-based controllers, enabling safe human-robot collaboration while maintaining cleaning effectiveness;

➂ Leverages modular TRUNC cell design to balance flexibility and scrubbing force, with future plans to expand applications in industrial and public scenarios

➀ Researchers developed an AI system enabling four-legged robots to adapt walking gaits using deep reinforcement learning, mimicking animal strategies like gait switching and real-time adjustment;

➁ The robot autonomously navigates diverse terrains relying solely on internal sensors, learning 9x faster than biological evolution without visual input;

➂ This breakthrough enhances robotic deployment in hazardous environments and provides new methods for studying animal locomotion through physics-based simulations.

➀ ETH Zurich researchers developed a four-legged robot named ANYmal-D that plays badminton with humans using reinforcement learning and a perception noise model;

➁ The robot's quadruped design enhances stability, while its stereo camera and dynamic arm enable precise shuttlecock tracking and hitting;

➂ Successful tests demonstrated continuous rallies of up to 10 shots, showcasing applications for legged robots in dynamic sports.

➀ Researchers have developed a shape-shifting robot called GOAT that can adapt to different terrains like animals.

➁ The robot can change its form between a flat rover shape and a sphere, enabling it to drive, roll, and even swim while using less energy.

➂ Future research will focus on refining algorithms and exploring size variations for different payloads, with potential applications in environmental monitoring, disaster response, and space exploration.

➀ Researchers at Northeastern University in China have developed an H-shaped bionic robot that mimics a cheetah’s running motion using piezoelectric materials.

➁ The H-shaped bionic piezoelectric robot (H-BPR) features four legs connected by three piezoelectric beams, allowing it to mimic a cheetah’s gait through periodic leg movements.

➂ Experiments showed the robot could achieve a maximum velocity of 66.79 mm/s and a load capacity of 55 g, demonstrating potential for industrial applications and search-and-rescue missions.

➀ Engineers and roboticists from the University of Hong Kong have developed a high-speed aerial robot named SUPER that can navigate unknown environments while avoiding obstacles.

➁ The robot uses 3D LiDAR technology instead of traditional cameras and sensors to provide real-time feedback and generate a flight path strategy in real-time.

➂ During testing, the robot demonstrated the ability to navigate obstacle courses at 20 meters per second and track targets in complex environments.