MAROKO133 Breaking ai: Watch: World-first convertible robot changes form to conquer all en

📌 MAROKO133 Breaking ai: Watch: World-first convertible robot changes form to conq

Hong Kong robotics firm Direct Drive Technology has unveiled the ‘world’s first’ fully modular embodied intelligence robot.

The 95-second video on YouTube shows the modular quadruped robot, named D1, traversing different types of terrains with human-like accuracy.

Its most interesting aspect is that the D1 robot can transform into a biped when needed, expanding its area of application across various terrains.

Quadrupeds are known for their stability and typically perform well on uneven surfaces or chaotic terrain. Bipedal robots are generally lighter in weight and smaller in size, known for their better performance on flat surfaces.

Because the real world contains both types of environments, Direct Drive Technology built the D1 to combine the advantages of each – using biped mode when efficiency and speed are possible, and quadruped mode when stability and sure-footedness are required.

The D1’s traversing capabilities

In the video, the D1 can be seen rolling as a scout vehicle over smooth terrains, showing its speed and ability to scout or patrol. This mode emphasizes the robot’s modular wheel capabilities.

The robot uses four legs when it has to traverse more complex terrain or carry supplies. It can also carry a full-grown human in such terrain.

The D1 robot shifts some modules to balance on two limbs, demonstrating its ability to adapt into a humanoid-like form or stand upright for interactions or observation.

In one instance, the robot took a heavy fall in rough terrain and managed to regain its balance with precision. It was also seen wheeling over a watery surface without losing its balance.

The robot configuration

Specs-wise, each biped bot section weighs 24.3 kg (53.6 lbs), has a maximum rolling speed of 11 km/h (7 mph), and can run for over five hours per two-hour charge of its 43.2V/9-Ah lithium battery.

Each section of the D1 robot operates with a Jetson Orin NX 8GB processor running the Ubuntu 22.04 operating system. This setup enables both remote and autonomous control.

When the modules are linked together, the two bipeds form a quadruped capable of handling a payload of up to 100 kg (220.5 lbs). Suggested applications for the robot include security patrolling, short-distance delivery, search and rescue, and mobile filming.

A single biped costs $7,499, and the double version is priced at $13,999, as shown on the company’s official website.

Locking horns with market rivals

Conceptually, the D1 robot is an upgrade over current market options, such as Boston Dynamics’ Spot robot dog. However, it’s still early to bank on the D1, given that the robot is still early in its lifecycle.

Spot remains a robot dog tested across various real-world environments, making it a much more reliable option for industrial and mission-critical tasks. With the D1 robot, however, limited field data and trials currently restrict it from overshadowing its quadruped counterpart.

That being said, the D1 does represent a significant leap forward with its modularity, multi-mode locomotion, and autonomous reconfiguration.

If Direct Drive Tech can validate the robot’s functional capabilities in various weather conditions, it could outperform current quadrupeds in terms of flexibility, efficiency, and mission adaptability.

🔗 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