R1-A5-D Ultimate G — mobile-base humanoid with LinkerBot O6 Tactile dexterous hands, wrist RGB cameras, MID360 LiDAR & stereo vision

What's in This Configuration

The R1-A5-D Ultimate G is the flagship high-DOF tactile manipulation platform in the Ultimate series. It pairs the same R1-A5-D mobile base with LinkerBot O6 Tactile six-DOF dexterous hands — adding force-sensitive tactile sensors to the already capable O6 architecture — plus wrist-mounted RGB cameras, all driven by the NVIDIA Jetson Orin NX 100 TOPS compute module. Unlike the Smart series, Ultimate G runs on an 8-core CPU with 16 GB LPDDR5 and dual DLA cores — delivering 2.5× the AI performance and enough memory headroom for large-model inference, multi-modal sensor fusion, and real-time control of complex hand kinematics with tactile feedback. The LinkerBot O6 Tactile hands provide five-finger dexterous grasping with 6 active degrees of freedom, 130 N grip force, and force-sensitive fingertips and palm arrays — enabling contact-aware grasping, slip detection, and force-controlled manipulation on the move.

LinkerBot O6 Tactile — Six-DOF Dexterity That Feels

The LinkerBot O6 Tactile takes the proven O6 six-DOF dexterous hand architecture and adds force-sensitive tactile sensing to fingertips and palm arrays. The underlying mechanics remain identical: 6 active degrees of freedom across 11 joints (6 active + 5 passive), worm-gear transmission for high torque density, 370 g per hand, and 130 N maximum five-finger grip force. What changes is the addition of tactile feedback — turning the hand from a position-controlled gripper into a contact-aware manipulation system. The tactile sensors detect normal and shear forces at the fingertips, enabling the robot to sense slip, adjust grip force in real time, and distinguish between object textures and compliance. Combined with the wrist RGB cameras, the O6 Tactile creates a closed loop: the wrist camera identifies the grasp target, the hand approaches and makes contact, the tactile sensors verify successful engagement, and the grip force modulates to prevent crushing fragile items or dropping heavy ones. This is the difference between grasping and feeling what you grasp — critical for physical human-robot interaction, delicate assembly, and robust mobile manipulation in unstructured environments.

LinkerBot O6 Tactile six-DOF dexterous hands on the R1-A5-D mobile base for contact-aware mobile manipulation

Tactile High-DOF Grasping — Force-Aware Dexterity on the Move

Ultimate G combines six-DOF dexterous hands with tactile sensing to create a platform optimized for force-aware mobile manipulation. The LinkerBot O6's 6 active DOF enable complex grasp formations — precision grips, power grasps, tripod pinches, and lateral holds — while the tactile sensors provide real-time feedback on contact quality. On a mobile base, this matters: the robot drives over uneven floors, accelerates and decelerates, and manipulates objects while in motion. Without tactile feedback, the hand must maintain excessive grip force to compensate for unknown disturbances. With tactile sensing, the hand dynamically modulates force based on actual contact conditions — gripping a wine glass with feather pressure and a toolbox with full strength, automatically. The Orin NX 100 TOPS runs tactile inference networks (slip detection, texture classification, force estimation) on the DLA cores while the GPU handles grasp planning and hand kinematics — all in real time, all on board.

Tactile-Visual-LiDAR Fusion — Four-Modal Perception on the Move

Ultimate G processes four synchronized sensor modalities on the Orin NX 100 TOPS: scene-level depth from the head binocular camera, 360° point clouds from the MID360 LiDAR, close-up RGB from the wrist-mounted cameras, and force/tactile data from the LinkerBot O6 Tactile fingertips and palm arrays. This is the richest sensor fusion stack in the R1-A5-D lineup. The dual DLA cores run vision and tactile inference networks in parallel — one DLA for head stereo and wrist grasp analysis, one for tactile feature extraction — while the GPU handles LiDAR SLAM, path planning, and dexterous hand control. With 16 GB RAM, the system holds multiple perception and control models concurrently: a scene understanding network for navigation, an object detector for grasp targets, a wrist-based grasp quality estimator, a tactile slip detector, and a force controller — all running simultaneously. This four-modal fusion enables research in multi-modal sensorimotor learning (training policies that fuse vision, touch, and proprioception), physical human-robot interaction (safe contact with humans), and robust mobile manipulation (grasping under uncertainty with tactile verification).

Autonomous Mobile Base — Navigate, Then Manipulate

The R1-A5-D chassis adds 3 degrees of freedom to the upper body: a telescopic pillar for height adjustment and a 2-DOF wheeled base for omnidirectional navigation. Combined with the MID360 LiDAR and stereo camera head, the robot builds real-time maps, plans paths, and avoids obstacles autonomously — then stops precisely to manipulate. The removable lithium-ion battery in the chassis delivers approximately 1.5 hours of untethered operation.

Orin NX 100 TOPS — Flagship Compute for Tactile Dexterous Manipulation

The built-in NVIDIA Jetson Orin NX delivers 100 TOPS of AI compute across an 8-core CPU, 16 GB LPDDR5, and dual NVDLA V2.0 accelerators. Controlling two LinkerBot O6 Tactile hands with 6 DOF each plus tactile inference, processing three visual streams (head stereo + two wrist cameras), running LiDAR SLAM, and fusing four sensor modalities simultaneously demands serious compute — and the Orin NX delivers. The dedicated DLA cores offload neural network inference from the GPU: one DLA for vision perception, one for tactile feature extraction, leaving GPU resources free for sensor fusion, grasp planning, and real-time hand kinematics. The 16 GB RAM provides space for large dexterous manipulation models, VLA policies, tactile-vision fusion networks, and multi-robot coordination algorithms that would not fit on an 8 GB module. No external GPU server, no cloud dependency — just power on and deploy.

Binocular Vision & Scene 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 MID360 LiDAR, wrist RGB cameras, and tactile sensors, the R1-A5-D Ultimate G offers the richest multi-modal perception stack in the lineup — with the Orin NX 100 TOPS processing all streams in real time for SLAM, object detection, grasp planning, tactile inference, and hand control.

Voice Collaboration & Natural Interaction

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

Collapsible. Mobile. Ready to Deploy.

At 683 × 520 × 440 mm collapsed (1323 mm elevated), the R1-A5-D Ultimate G navigates standard doorways, elevators, and lab aisles. The telescopic pillar lowers the upper body for transport and raises it for manipulation. At approximately 30 kg with battery, it is a self-contained mobile research platform — no external compute rack, no tether cables during operation.

Collapsed (683 mm) and elevated (1323 mm) configurations for navigation and manipulation

Full-Stack Open Development

Unitree provides a mature, Linux-based robot development framework with open APIs for the underlying system, robotic arms, mobile base, audio, lighting, and visual control. ROS 2 native support, dual encoders on every arm joint, and hollow internal wiring make the R1-A5-D Ultimate G a clean slate for reinforcement learning, tactile dexterous manipulation research, embodied AI, and autonomous navigation studies. The built-in Orin NX 100 TOPS with dual DLA cores and 16 GB RAM handles perception, tactile inference, grasp planning, and hand control locally — with compute headroom to grow into large VLA models, multi-robot coordination, and on-device reinforcement learning without hardware limits.

Upgrade Path — Grow With Your Research

Ultimate G is the top high-DOF tactile manipulation platform with flagship compute already installed. The Orin NX 100 TOPS, 16 GB RAM, and dual DLA cores provide headroom to grow:

• End Effectors — Upgrade to BrainCo Revo 2 for anthropomorphic manipulation or tendon-driven hands for biomimetic research. The compute is already ready.
• Large Vision-Language Models — Run VLA policies and multi-modal foundation models locally without quantization.
• Multi-Robot Coordination — Deploy fleets of Ultimate G units with distributed perception, coordinated manipulation, and shared tactile learning.
• On-Device Reinforcement Learning — Train dexterous manipulation policies directly on the robot using tactile reward signals.

Designed For

Technical Specifications

What's in the Box

• Unitree R1-A5-D Ultimate G Humanoid Robot (mobile base)
• LinkerBot O6 Tactile six-DOF dexterous hands (pre-installed, both arms)
• Wrist-mounted RGB cameras (pre-installed, both wrists)
• NVIDIA Jetson Orin NX 100 TOPS compute module with 16 GB RAM & dual DLA cores (pre-installed)
• Removable chassis lithium-ion battery pack
• External power supply & AC cable
• MID360 LiDAR module (pre-installed in chassis)
• Binocular vision head module (pre-installed)
• 4-array microphone + dual speaker module (pre-installed)
• Standard dual-hand remote controllers
• Quick-start guide & safety manual
• Secondary development documentation access