Unitree R1-A5 Smart C Intelligent Bionic Upper Body Humanoid Robot

Unitree R1-A5 Smart C Humanoid Robot

The R1-A5 Smart C takes the Dex3-1 platform further with integrated tactile sensors on each three-finger hand. Combined with wrist RGB cameras and the built-in Orin Nano 40 TOPS, this configuration enables closed-loop force-controlled grasping, slip detection, and tactile exploration — essential for research in physical human-robot interaction, delicate object handling, and sensorimotor learning.

Dex3-1 Tactile — Feel What You Grasp

The Dex3-1 Tactile builds on the three-finger dexterity of the standard Dex3-1 with an array of force-sensitive sensors integrated into the fingertips and palm surfaces. This enables real-time pressure mapping, contact detection, and slip prediction — critical capabilities for compliant grasping of fragile objects, tool use with variable force, and physical human-robot interaction. The tactile stream fuses with wrist camera RGB data on the Orin Nano 40 TOPS for multi-modal perception pipelines that close the loop between touch and vision.

Tactile-Visual Fusion — Multi-Modal Perception

Smart C is the first R1-A5 configuration to combine tactile sensing with dual wrist RGB cameras and binocular head vision in a single platform. The Orin Nano 40 TOPS processes synchronized streams from all sensor modalities: tactile pressure maps from each finger, close-up RGB from wrist cameras, and scene-level depth from the head. This multi-modal fusion enables research in sensorimotor learning, tactile exploration, affordance prediction, and force-aware manipulation — domains where vision alone is insufficient.

Orin Nano 40 TOPS — AI Compute for Tactile-Vision Fusion

The built-in NVIDIA Jetson Orin Nano delivers 40 TOPS of AI compute, handling simultaneous streams from tactile sensors, wrist RGB cameras, and head binocular vision. Run tactile-visual fusion networks, grasp stability predictors, and reinforcement learning policies locally with sub-millisecond latency. No external GPU server, no cloud dependency — just plug in power and start training.

Binocular Vision & Multi-Camera Perception

A binocular camera head delivers a 146° × 124° field of view with a 60 mm stereo baseline for scene-level understanding. RGB streams at 1280×720@30 Hz and depth at 544×448@10 Hz. Combined with the wrist RGB cameras and tactile arrays on each Dex3-1 Tactile hand, the R1-A5 Smart C offers a complete tactile-visual perception stack for research in multi-modal manipulation, active perception, and sensorimotor learning.

Voice Collaboration & Natural Interaction

Four-array beamforming microphones and dual 3 W speakers enable clear voice capture and synthesis in noisy labs or classrooms. The open audio framework supports custom wake words, multilingual TTS, and integration with LLMs for conversational command-and-control.

Compact. Lightweight. Desk-Ready.

At just 700 × 357 × 190 mm and approximately 11 kg, the R1-A5 Smart C fits on a standard lab bench or classroom desk. The fixed-base design eliminates footprint concerns while delivering full upper-body humanoid kinematics with tactile sensing. Power via the included external supply — an optional upper-body Li-ion battery is available for ~1.5 hours of untethered operation.

Full-Stack Open Development

Unitree provides a mature, Linux-based robot development framework with open APIs for the underlying system, robotic arms, audio, lighting, visual control, and tactile sensing. ROS 2 native support, dual encoders on every arm joint, and hollow internal wiring make the R1-A5 Smart C a clean slate for reinforcement learning, teleoperation research, embodied AI, and tactile sensorimotor learning. The built-in Orin Nano 40 TOPS handles multi-modal model training and inference without external hardware.

• Ambidextrous humanoid platform with human-like interaction and gesturing.
• Features Dex3-1 Tactile three-finger hands with force-sensitive sensors integrated into the fingertips and palm surfaces.
• Enables closed-loop force-controlled grasping, slip detection, and tactile exploration through integrated tactile sensing.
• Includes tactile sensing, dual wrist RGB cameras, and binocular head vision within a single platform.
• Features an NVIDIA Jetson Orin Nano delivering 40 TOPS of AI compute.
• Handles simultaneous processing of tactile sensors, wrist RGB cameras, and binocular head vision locally.
• Runs tactile-visual fusion networks, grasp stability prediction, and reinforcement learning workloads with sub-millisecond latency.
• Features a binocular camera head with a 146° x 124° field of view and a 60 mm stereo baseline.
• Captures RGB video at 1280 x 720 resolution at 30 Hz.
• Combines binocular head vision, wrist RGB cameras, and tactile sensing for multi-modal perception.
• Includes a four-array beamforming microphone system for voice capture.
• Features dual 3 W speakers for voice synthesis.
• Supports custom wake words, multilingual text-to-speech, and LLM integration through an open audio framework.