PaXini’s TORA DoubleOne Conquers Obstacles at CES 2026

CES Spotlight Steals

PaXini Technology turned heads at CES 2026 in Las Vegas by unveiling the TORA DoubleOne, a wheeled humanoid robot that deftly navigated obstacles, responded to physical touches, and executed handshakes with attendees. This dynamic demo highlighted the robot's ability to handle real-time interactions in crowded spaces, pushing boundaries for embodied AI in everyday environments. Drawing from its Waseda University roots, PaXini positioned TORA DoubleOne as a bridge from lab prototypes to practical deployment, emphasizing tactile smarts over mere mobility.

Touch Mastery Revealed

TORA DoubleOne stands out with its multimodal perception suite, blending 3D LiDAR, fisheye cameras, depth sensors, and multidimensional tactile units for human-like awareness. These enable seamless obstacle avoidance and adaptive responses, like gentle handshakes or evading bumps during motion. Fully foldable design adjusts operating heights for tight spaces, while modular hands swap for task-specific dexterity, redefining how humanoids engage dynamically without rigid programming.

Sensing Tech Leap

Engineers at PaXini integrated high-precision tactile sensors with 0.01 N resolution across 15 dimensions, including six-axis force, texture, and resilience, powering nuanced manipulation. NVIDIA Jetson AGX Orin drives onboard AI for low-latency decisions, paired with a Linux-based OS and ROS2 compatibility. This stack supports VTLA models for generalization, allowing the robot to learn from touches and visuals, conquering unpredictable terrains shown in CES obstacle courses.

Deployment Horizons Expand

Beyond demos, TORA DoubleOne targets logistics warehouses, automotive factories, elderly care, and food service, where upper-body precision handles packages or assists humans. In real trials, it sorts items, inspects goods, and supports medical tasks via safe, collaborative modes. Prioritizing seated or fixed-station work cuts costs on legged designs, focusing resources on interaction skills that boost efficiency in shared human-robot zones.

Dexterity Enablers Unpacked

TORA DoubleOne's skill architecture shines through sensors like RGB-D cameras, 3D LiDAR, IMU, gyroscope, force, and tactile arrays, enabling precise people interactions and dexterous handling. Wheeled high-speed AGV-hybrid mobility hits 1.8 m/s for swift navigation via Laser-SLAM and visual mapping, while 3-5 year battery life sustains long assistance sessions. Model P (686 x 528 x 1671 mm, 70 kg) and X (774 x 502 x 1687 mm, 73 kg) frames carry tools, packages, or instruments, with force-limiting safety for collaborative fluidity.

Rivals Face-Off