MAROKO133 Update ai: South Korea commits $3.4 billion to 16,000lb-thrust indigenous KF-21

📌 MAROKO133 Hot ai: South Korea commits $3.4 billion to 16,000lb-thrust indigenous

South Korea has moved closer to powering its next-generation fighter jets with a homegrown engine, as the government finalised a long-term investment plan and development schedule for the program.

The Defense Acquisition Program Administration confirmed that the Advanced Aviation Engine Development Project will run from 2027 to 2040, backed by W3.4 trillion or roughly $3.4 billion in funding.

The effort aims to deliver a fighter-class engine producing 16,000 pounds of military thrust and 24,000 pounds with afterburners, exceeding earlier performance targets outlined by Seoul.

“This project will involve the development and production of a prototype of an advanced aircraft engine, conducting development and certification testing, and completing preparations for system-mounted flight tests,” says the DAPA.

“This will enable future flight tests on domestically produced fighter jets,” as reported by FlightGlobal.

The engine will serve as the powerplant for the KF-21 Block 3 fighter, which South Korea plans to field around 2040 as part of its long-term air combat roadmap.

Engine goals and scope

Engine development will begin in 2027 and progress through prototype manufacturing, testing, and certification before aircraft integration.

Officials describe the program as essential for improving performance and reducing dependence on imported propulsion systems.

The new engine targets a 15 percent reduction in specific fuel consumption compared with the General Electric F414 engine currently used on the KF-21.

Block 1 and Block 2 versions of the fighter will continue to rely on the F414, which Hanwha Aerospace produces locally under license.

Seoul has promoted indigenous engine development for years, citing both operational and economic concerns.

The Block 3 configuration marks the first platform intended to fly with a fully domestic fighter jet engine.

Government and industry alignment

To manage the program, South Korea recently set up an inter-ministerial coordination body bringing together defense, industry, and aviation authorities.

The council aims to prevent overlapping efforts and policy disputes that have delayed previous aerospace projects.

The Defense Acquisition Program Administration plans to convene the first meeting later this month.

Officials from the Ministry of National Defense, the Ministry of Trade, Industry and Energy, the Korea AeroSpace Administration, and the Ministry of Land, Infrastructure and Transport are expected to participate.

Hanwha Aerospace will lead industrial work on the engine, drawing on its experience producing foreign fighter engines.

Doosan Enerbility has also joined the program, signaling a broader push to strengthen domestic manufacturing capabilities.

South Korea’s combat aircraft fleet currently depends on foreign-made engines, mainly from the United States.

This reliance can limit export approvals and impose operational constraints during maintenance and overhaul cycles.

Officials believe a domestically developed engine will ease these restrictions, improve readiness, and support future exports of the KF-21.

The government also expects the program to generate spillover benefits for civilian aerospace technologies over the long term.

By setting clear funding and timelines, Seoul has turned its ambition for an indigenous fighter jet engine into a structured national program.

🔗 Sumber: interestingengineering.com

📌 MAROKO133 Eksklusif ai: ByteDance Introduces Astra: A Dual-Model Architecture fo

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