MAROKO133 Eksklusif ai: China’s firm launches first full-size humanoid robot with touch mo

📌 MAROKO133 Eksklusif ai: China’s firm launches first full-size humanoid robot wit

Chinese robotics firm X-Humanoid has launched Embodied Tiangong 3.0, its next-gen general-purpose robot platform, on Tuesday. Designed to be “more open and easier to use”, the platform leverages the company’s proprietary Huisi Kaiwu embodied intelligence platform.

Tiangong 3.0 has achieved multiple technological breakthroughs, becoming the first full-size humanoid robot capable of touch-interactive, high-dynamic whole-body control. With an open ecosystem, Tiangong 3.0 aims to address the compatibility challenges humanoids face in industrial applications.

The hardware and the software

Tiangong 3.0 provides multiple expansion interfaces that allow peripheral tools to integrate flexibly for diverse scenarios. These scenarios include specialized operations and industrial manufacturing.

On the software side, Embodied Tiangong 3.0 supports ROS2, MQTT, TCP/IP, and other mainstream protocols. Users can work on secondary development without restructuring the underlying technology, as the system is paired with a comprehensive development toolchain and a low-code platform. This feature significantly reduces development costs and barriers.

X-Humanoid has also released several of its core technologies as open source, including the Embodied Tiangong robot hardware platform, the Embodied Tiangong vision-language model (Pelican-VL), and the RoboMIND dataset. This move aims to lower technical barriers and accelerate innovation across the robotics industry.

Analyzing the practical performance

Embodied Tiangong 3.0 features high-torque integrated joints that provide strong stability and enable complex actions, such as climbing over 1-meter obstacles. The robot’s degree of freedom enables it to achieve millimeter-level precision, which is especially useful for performing intricate tasks in small spaces.

The Huisi Kaiwu platform forms a closed-loop system linking perception, decision-making, and execution, enabling coordinated operation between a robot’s “small brain” for motion control and “large brain” for higher-level reasoning.

This architecture supports fully autonomous behavior while also allowing asynchronous coordination across multiple robots. In practice, a single central intelligence can manage several robots or multiple skill sets at once, helping move embodied AI from lab experiments into real-world industrial use.

A history of milestones

In April 2025, X-Humanoid’s Tiangong Ultra completed a 13-mile (21 km) half-marathon in 2 hours, 40 minutes. In doing so, it became the first humanoid win a half-marathon title.

Later in August last year, Tiangong Ultra and Tian Yi 2.0 won 2 gold, 6 silver, and 2 bronze medals at the World Humanoid Robot Games. Tiangong Ultra also emerged victorious in the fully autonomous 100-meter sprint final, clocking 21.50 seconds.

Tian Yi 2.0, powered by the Huisi Kaiwu platform, showcased strong general-purpose intelligence and autonomy, taking both first and second place in the material-handling competition. These results highlight the platform’s growing role in enabling high-performance, real-world robotic capabilities.

Delving into the future

The launch of Embodied Tiangong 3.0 represents a significant milestone in X-Humanoid’s construction of an embodied intelligence ecosystem.

As a national-level innovation platform, the company plans to leverage its two core technology platforms and work closely with industry partners to accelerate the translation of research into real-world applications, expand the use of humanoid robots, and advance human–robot collaboration in everyday environments.

🔗 Sumber: interestingengineering.com

📌 MAROKO133 Breaking ai: ByteDance Introduces Astra: A Dual-Model Architecture for

The increasing integration of robots across various sectors, from industrial manufacturing to daily life, highlights a growing need for advanced navigation systems. However, contemporary robot navigation systems face significant challenges in diverse and complex indoor environments, exposing the limitations of traditional approaches. Addressing the fundamental questions of “Where am I?”, “Where am I going?”, and “How do I get there?”, ByteDance has developed Astra, an innovative dual-model architecture designed to overcome these traditional navigation bottlenecks and enable general-purpose mobile robots.

Traditional navigation systems typically consist of multiple, smaller, and often rule-based modules to handle the core challenges of target localization, self-localization, and path planning. Target localization involves understanding natural language or image cues to pinpoint a destination on a map. Self-localization requires a robot to determine its precise position within a map, especially challenging in repetitive environments like warehouses where traditional methods often rely on artificial landmarks (e.g., QR codes). Path planning further divides into global planning for rough route generation and local planning for real-time obstacle avoidance and reaching intermediate waypoints.

While foundation models have shown promise in integrating smaller models to tackle broader tasks, the optimal number of models and their effective integration for comprehensive navigation remained an open question.

ByteDance’s Astra, detailed in their paper “Astra: Toward General-Purpose Mobile Robots via Hierarchical Multimodal Learning” (website: https://astra-mobility.github.io/), addresses these limitations. Following the System 1/System 2 paradigm, Astra features two primary sub-models: Astra-Global and Astra-Local. Astra-Global handles low-frequency tasks like target and self-localization, while Astra-Local manages high-frequency tasks such as local path planning and odometry estimation. This architecture promises to revolutionize how robots navigate complex indoor spaces.

Astra-Global: The Intelligent Brain for Global Localization

Astra-Global serves as the intelligent core of the Astra architecture, responsible for critical low-frequency tasks: self-localization and target localization. It functions as a Multimodal Large Language Model (MLLM), adept at processing both visual and linguistic inputs to achieve precise global positioning within a map. Its strength lies in utilizing a hybrid topological-semantic graph as contextual input, allowing the model to accurately locate positions based on query images or text prompts.

The construction of this robust localization system begins with offline mapping. The research team developed an offline method to build a hybrid topological-semantic graph G=(V,E,L):

• V (Nodes): Keyframes, obtained by temporal downsampling of input video and SfM-estimated 6-Degrees-of-Freedom (DoF) camera poses, act as nodes encoding camera poses and landmark references.
• E (Edges): Undirected edges establish connectivity based on relative node poses, crucial for global path planning.
• L (Landmarks): Semantic landmark information is extracted by Astra-Global from visual data at each node, enriching the map’s semantic understanding. These landmarks store semantic attributes and are connected to multiple nodes via co-visibility relationships.

In practical localization, Astra-Global’s self-localization and target localization capabilities leverage a coarse-to-fine two-stage process for visual-language localization. The coarse stage analyzes input images and localization prompts, detects landmarks, establishes correspondence with a pre-built landmark map, and filters candidates based on visual consistency. The fine stage then uses the query image and coarse output to sample reference map nodes from the offline map, comparing their visual and positional information to directly output the predicted pose.

For language-based target localization, the model interprets natural language instructions, identifies relevant landmarks using their functional descriptions within the map, and then leverages landmark-to-node association mechanisms to locate relevant nodes, retrieving target images and 6-DoF poses.

To empower Astra-Global with robust localization abilities, the team employed a meticulous training methodology. Using Qwen2.5-VL as the backbone, they combined Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO). SFT involved diverse datasets for various tasks, including coarse and fine localization, co-visibility detection, and motion trend estimation. In the GRPO phase, a rule-based reward function (including format, landmark extraction, map matching, and extra landmark rewards) was used to train for visual-language localization. Experiments showed GRPO significantly improved Astra-Global’s zero-shot generalization, achieving 99.9% localization accuracy in unseen home environments, surpassing SFT-only methods.

Astra-Local: The Intelligent Assistant for Local Planning

Astra-Local acts as the intelligent assistant for Astra’s high-frequency tasks, a multi-task network capable of efficiently generating local paths and accurately estimating odometry from sensor data. Its architecture comprises three core components: a 4D spatio-temporal encoder, a planning head, and an odometry head.

The 4D spatio-temporal encoder replaces traditional mobile stack perception and prediction modules. It begins with a 3D spatial encoder that processes N omnidirectional images through a Vision Transformer (ViT) and Lift-Splat-Shoot to convert 2D image features into 3D voxel features. This 3D encoder is trained using self-supervised learning via 3D volumetric differentiable neural rendering. The 4D spatio-temporal encoder then builds upon the 3D encoder, taking past voxel features and future timestamps as input to predict future voxel features through ResNet and DiT modules, providing current and future environmental representations for planning and odometry.

The planning head, based on pre-trained 4D features, robot speed, and task information, generates executable trajectories using Transformer-based flow matching. To prevent collisions, the planning head incorporates a masked ESDF loss (Euclidean Signed Distance Field). This loss calculates the ESDF of a 3D occupancy map and applies a 2D ground truth trajectory mask, significantly reducing collision rates. Experiments demonstrate its superior performance in collision rate and overall score on out-of-distribution (OOD) datasets compared to other methods.

The odometry head predicts the robot’s relative pose using current and past 4D features and additional sensor data (e.g., IMU, wheel data). It trains a Transformer model to fuse information from different sensors. Each sensor modality is processed by a specific tokenizer, combined with modality embeddings and temporal positional embeddi…

Konten dipersingkat otomatis.

🔗 Sumber: syncedreview.com