MAROKO133 Update ai: China turns buried $50m tunnel machine failure into landmark feat of

📌 MAROKO133 Eksklusif ai: China turns buried $50m tunnel machine failure into land

Chinese engineers have managed to turn disaster into a masterful engineering feat by rescuing a stranded tunnel boring machine (TBM) using its twin. Conducted under the Yangtze River, this impressive engineering success is something the team, understandably, is very proud of.

The problem initially arose during the construction of the Jiangyin–Jingjiang Yangtze River Tunnel, a 4-mile (6.4 km) road tunnel under the Yangtze River. This was being completed using a 52.5-foot (16-meter) wide, multi-million-dollar TBM.

These machines can excavate the ground while supporting the ground above as the tunnel progresses. Such machines are also able to line and support the tunnel in their wake.

Work was progressing as expected until the TBM catastrophically failed. At this point, the TBM was around 177 feet (54 meters) underground and subject to enormous water pressure from above.

Write-off or recover?

Under these conditions, the TBM was unable to reverse and couldn’t be safely dismantled for recovery. It also couldn’t be repaired in situ. Things looked very bleak indeed for the stranded machine and the project as a whole.

The engineering team faced one of several equally disastrous choices. The first was to completely abandon the machine and absorb the loss. The second choice was to completely redesign or cancel the entire tunnel project.

In all cases, this would lead to years of delay and much embarrassment for all involved. However, the team decided to try a third option: rescue the TBM using its twin.

The idea was to launch the second TBM from the opposite bank of the river and drive it straight at the stranded one. While this sounds simple on the surface, the undertaking was not going to be a walk in the park.

To succeed, they had to predict ground movement under a massive river, while also controlling direction over kilometres with millimetre precision. The team also had to avoid even tiny vertical or horizontal errors that could cause collapse or flooding.

Above all, the target error margin was smaller than the thickness of a coin! Incredibly, the team managed to pull it off, and all with a vertical error of just 2mm.

Truly impressive engineering feat

The final horizontal misalignment was effectively zero and was completed under high pressure in soft sediments and water-saturated ground. The two machines met cleanly underground (a procedure called a mid-tunnel docking), which is one of the hardest operations in underground civil engineering.

This allowed engineers to not only access the failed TBM but also recover the tunnel project. It also enabled them to continue the project instead of scrapping it completely.

This nws in not just about saving the tunnel project, but is also an interesting showcase of what can be achieved with a little bit of lateral thinking and planning. Notably, it demonstrates that deep underground rescue is possible, even under rivers, and that large TBM failures don’t have to be terminal anymore.

It also highlighted how precision guidance systems for tunnelling have reached a new level. Looking ahead, the lessons learned from the near-disaster could be used for future projects like sub-sea tunnels, metro systems, and work in high-risk geological 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