MAROKO133 Hot ai: ByteDance Introduces Astra: A Dual-Model Architecture for Autonomous Rob

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

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

📌 MAROKO133 Eksklusif ai: US unlocks cheaper jet fuel with new catalyst that conve

Researchers in the US have designed a catalyst that could significantly lower the cost of producing sustainable aviation fuel (SAF), by converting ethanol into jet fuel precursors in a single step.

Colorado-based advanced biofuels company Gevo licensed two patented catalyst technologies from the US Department of Energy’s Oak Ridge National Laboratory (ORNL), to accelerate the commercial production of sustainable aviation fuel.

The innovation relies on a streamlined method that turns ethanol (also known as ethyl alcohol), commonly sourced from plant or waste feedstocks, into olefins (ETO). These are essential precursors used to produce jet fuel.

While this conversion typically involves multiple steps, the newly licensed catalyst supports a single-step ethanol-to-olefins pathway. This significantly improves the production process.

“This partnership will streamline the transition of ORNL’s catalyst technologies from lab scale to pilot-scale reactors,” Andrew Sutton, PhD, a senior scientist in the manufacturing science division at ORNL, explained.

Single-step fuel production

SAF is a cleaner, non-petroleum-based alternative to conventional jet fuel. It is produced from renewable waste, fats, oils and agricultural residues. It is widely regarded as a critical solution in the push to decarbonize air travel.

The International Air Transport Association, which represents over 80 percent of global air traffic, has signaled strong interest in SAF. Many airlines have already committed to large-scale purchases.

Nevertheless, production efficiencies still remain an issue, predominantly due to high production costs, limited feedstock availability, and complex infrastructure requirements.

To address this challenge, ORNL built a catalyst technology capable of improving carbon efficiency and cutting the cost of converting ethanol into fuel precursors.

At the same time, the olefins produced through this process can also be utilized in the production of plastics, solvents and surfactants. For perspective, the global plastics market is set to surpass USD 1.3 trillion by 2033.

“Gevo’s collaboration with Oak Ridge National Laboratory focuses on evaluating a novel catalytic process that converts ethanol into valuable fuel precursors and alternative chemicals like butadiene,” Andrew Ingram, PhD, director of process chemistry and catalysis at Gevo, said.

Unlocking cheaper SAF

The project is supported by a three-year cooperative research and development (R&D) agreement under the DOE’s Technology Commercialization Fund. Under the program, ORNL will develop and test catalyst pellets in advanced chemical reactors.

It will also build computational models to predict performance at industrial scale. “This work complements our broader ethanol conversion portfolio but is distinct from both our commercial deployment of Axens’ alcohol-to-jet process and our next-generation ETO platform,” Ingram continued.

In turn, Gevo will contribute process design and operational expertise, and guide how the technology is integrated in the pilot reactor.

“If the economics prove out, this pathway could provide a flexible, cost-effective option to scale US bio-based solutions, driven by American innovation that creates new markets and demand for farmers producing feedstocks for energy and materials,” Ingram concluded in a press release.

ORNL will also leverage advanced materials analysis capabilities at its Center for Nanophase Materials Sciences to better understand catalyst performance in large reactors.

Global demand for jet fuel is expected to surge to 230 billion gallons by 2050. Expanding SAF use could help the aviation industry meet this demand while advancing US energy independence and security, and reducing emissions.

🔗 Sumber: interestingengineering.com