MAROKO133 Eksklusif ai: ByteDance Introduces Astra: A Dual-Model Architecture for Autonomo

📌 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 Hot ai: Russia’s 1,250 MWe nuclear reactor unit with 100-year-lifespan

The first light-water reactor at the new Russian nuclear facility, Kursk II, has now reached full power as part of its commissioning. The 1,250 MWe unit was connected to the grid in December and was allowed to achieve full power after a series of checks and tests at each capacity level. 

Kursk II is a new nuclear power plant under construction in western Russia, a mere 37.5 miles (60 km) from the Ukrainian border. It derives its name from the previous nuclear facility at the site, which housed four RBMK-1000 reactors. 

The facility was designed for an initial lifespan of 30 years, which was subsequently extended by another 15 years through life-extension programs. The first unit of the Kursk nuclear plant was shut down in December 2021, and the second unit in January 2024. The power plant is being replaced by Kursk II, which consists of four VVER-TOI light-water reactors with a capacity of 1,250 MWe each. 

Russian light water reactors

Light water reactors (LWRs) are a type of thermal neutron reactor that uses ordinary water rather than heavy water as both coolant and neutron moderator. These are the most common types of nuclear reactors, usually deployed as pressurized water reactors (PWRs) and boiling water reactors (BWRs). 

The water-water energetic reactor (VVER) is a type of pressurized water reactor design originally developed in the Soviet Union and sold by Russia. VVER nuclear reactors have since been developed and installed in countries such as China, India, Finland, the Czech Republic, Jordan, Turkey, and many more. 

With water serving as both coolant and moderator, the VVER design is inherently safer. Power outputs for these reactors range from 70 to 1,300 MWe, with 1,700 MWe under development. 

First unit at Kursk II

The four units scheduled for Kursk II are VVER-TOI reactors, a third-generation design. Amongst its improvements are upgraded pressure vessels, low-speed turbines, up to 100 years of service life, when compared to its predecessor, the VVER-1200. 

Construction of the first unit began in 2018, and the reactor vessel was installed in 2022. The outer dome was completed in August 2023, while the unit was connected to the grid in December. 

“At each stage, our specialists conducted a series of tests to verify the parameters and performance characteristics of the systems and equipment met their design values ​​and ensure their reliable and safe operation,” said Alexander Uvakin, Kursk Nuclear Power Plant Director, in a press release. 

“At 100 percent power, we will once again examine the neutronic characteristics of the core, study the reactor’s performance under various scenarios, test the in-core monitoring systems, conduct a complete power-down of the unit, and check the effectiveness of the biological shielding and radiation situation at the NPP.” 

“We can now confidently say that the unit has confirmed the validity of the design and engineering solutions adopted in the VVER-TOI design, the high quality of the manufactured equipment, and the construction and installation work performed,” added Oleg Shperle, Vice President and Director of the Kursk NPP Construction Project at JSC ASE. 

The second unit of the light water reactor is also under construction at Kursk II, and all four units are expected to be completed by 2034. 

🔗 Sumber: interestingengineering.com