Tech Xplore Robotics
Covers the latest news on robotics science and technology
Kinematic Intelligence Lets Three Different Robots Learn the Same Task Safely
Researchers at EPFL’s LASA laboratory introduced a control framework called kinematic intelligence that translates a single human‑demonstrated task into a generic movement strategy adaptable to multiple robots. The system mathematically maps demonstrations, classifies each robot’s joint limits, and automatically tailors motions, allowing three commercially distinct robots to push, place and throw a block safely on an assembly line. Published in Science Robotics, the work promises to cut reprogramming time when upgrading robot fleets. The authors envision extensions to natural‑language commands and collaborative settings.
AI-Guided Snakebot Unlocks Rolling Move that Doubles Speed per Unit Power
Researchers at Osaka Metropolitan University have created an AI‑guided snake robot that can switch to a rolling gait, achieving roughly twice the travel speed per unit of power on flat surfaces. The system uses deep reinforcement learning and an observation...
This Robot Sees Danger, Decides Its Route and Powers over Obstacles While Carrying Loads
KAIST researchers unveiled DreamWaQ++, a quadrupedal‑robot control system that fuses camera, LiDAR and proprioceptive data to anticipate terrain and adjust its gait in real time. The multimodal reinforcement‑learning architecture lets the robot climb 35° slopes, traverse stairs and clear obstacles...
Origami-Inspired Robot Built From Printable Polymers Uses Electric Current to Move
Engineers at Princeton have built a soft‑rigid hybrid robot using 3D‑printed liquid crystal elastomer (LCE) hinges and embedded flexible printed circuit boards. The robot moves by localized heating of the polymer, eliminating the need for motors or external pneumatic systems....
These AI-Powered Guide Dogs Don't Just Lead, They Talk
Binghamton University researchers have built a robot guide dog that uses GPT‑4 to converse with visually impaired users, offering route planning and real‑time verbal navigation cues. The system was demonstrated at AAAI 2026 and tested with seven legally blind participants who...
Magnetic Coil Setup Guides Microrobots without Seeing Them
SMU researchers have built a triaxial Helmholtz coil system that creates a uniform magnetic field gradient, enabling microrobots to be guided without continuous visual tracking. The six‑coil arrangement, calibrated with a triaxial magnetometer and refined by Tikhonov regularization, delivers consistent...
Too Many Cooks, or Too Many Robots? Finding a Goldilocks Level of Randomness to Keep Robot Swarms Moving
Harvard researchers led by L. Mahadevan and Ph.D. student Lucy Liu discovered that adding a calibrated amount of randomness to robot swarm movement can prevent gridlock and boost efficiency. Using simulations and physical wheeled robots, they identified a Goldilocks zone where...
Introducing MirrorBot, a Robot Designed to Foster Human Connection
Cornell University researchers unveiled MirrorBot, a four‑foot robot equipped with dual mirrors that creates eye contact between strangers. In a waiting‑room experiment with 32 participants, the robot sparked conversations and playful exchanges, with 12 of 16 pairs reporting that their...
Do You Trust Me? A Framework for Making Networks of Robots and Vehicles Safer
Harvard researchers introduced “cy‑trust,” a quantitative trust metric that lets autonomous robots and vehicles evaluate data from peers before acting. The framework assigns each data source a trust score between 0 and 1 using onboard sensors, lidar, radar and signal‑processing of...
Air-Powered Artificial Muscles Could Help Robots Lift 100 Times Their Weight
Arizona State University researchers have created air‑powered artificial muscles that mimic biological tissue. The pneumatic actuators can lift up to 100 times a robot’s own weight while remaining lightweight and untethered. Their design enables operation in extreme conditions such as...
Researchers Build a Robotic Swarm with No Electronics, No Batteries and No Brains
Georgia Tech researchers have created an electronic‑free robotic swarm whose behavior emerges solely from its mechanical design. Tiny particle robots latch, release and reconfigure when exposed to external vibrations, eliminating the need for sensors, processors or batteries. The system scales...
Combining the Robot Operating System with LLMs for Natural-Language Control
Researchers from Huawei Noah's Ark Lab, TU Darmstadt and ETH Zurich unveiled an open‑source framework that fuses large language models with the Robot Operating System (ROS). The system interprets natural‑language commands and translates them into robot actions via inline code...
Control Framework Lets Flexible Robots Move in Tight Spaces with Less Math
Researchers at the Indian Institute of Technology Gandhinagar unveiled a virtual actuation space (VAS) framework that simplifies control of tendon‑driven continuum robots (TDCR). By representing each robot section with just direction and magnitude, VAS eliminates the need for complex infinite‑degree‑of‑freedom...
Q&A: Robots Can't Feel, but Novel Sensors Could Change That
Researchers at Penn State have developed a flexible pressure‑sensor array using reduced graphene oxide aerogel, creating an electronic skin capable of ultrahigh sensitivity and a broad pressure range. Each 8 mm sensor supports roughly three ounces of force and endures over...
Robots with Different Bodies Can Now Share Skills: What Intention-Based Learning Changes
Researchers introduced Intention‑Aligned Imitation Learning (IAIL), a method that lets robots share skills by focusing on high‑level goals expressed in natural language rather than exact motions. Published in Science Robotics, the approach was validated on seven heterogeneous robots across 30...
