📌 TOPINDIATOURS Update ai: China hosts world’s largest real-life ‘Quidditch’ drone

The Ablefly National Drone Soccer Championship Finals concluded on Sunday in Chengdu, marking a milestone for both competitive sports and the low-altitude economy.

The event brought together more than 10,000 participants, making it the world’s largest drone soccer tournament to date.

A total of 1,116 teams from China and overseas competed over three days at the Chengdu Airport International Convention Center.

The finals highlighted how emerging aviation technologies are moving beyond industry use cases and into organized mass-participation sports.

Drone soccer requires pilots to maneuver spherical drones through an opposing goal. The rapid aerial gameplay has drawn comparisons to a real-world version of “Quidditch” from the Harry Potter series.

The format blends engineering skill, hand-eye coordination, and tactical teamwork.

Held from February 7 to 9, the event was hosted by the People’s Government of Chengdu City and the Chinese Society of Aeronautics. Multiple municipal departments and industry partners supported organization and execution.

By the final day, all championship rankings had been decided.

Record-scale competition

Organizers positioned the finals as the first unmanned aerial vehicle football tournament to reach a “10000 people competing together” scale.

The structure emphasized both professionalism and accessibility, allowing large-scale participation without sacrificing competitive integrity.

The tournament followed a progressive elimination system. Matches ran daily from 8:30 a.m. to 8:00 p.m.

Group matches and early knockout rounds reduced the field from 1,116 teams to 558.

Cross-elimination rounds then narrowed the competition to 279 teams.

On the final day, teams advanced through successive elimination rounds to form a final 16. A round-robin and ranking format determined all placements, including first, second, and third positions.

The dense schedule tested pilot endurance and team coordination.

To ensure consistent officiating, organizers deployed 242 certified referees.

The referees oversaw match rules, scoring accuracy, and safety compliance. Their presence supported fair play across hundreds of matches.

Youth-driven tech sport

Drone soccer has moved quickly from a niche activity to a youth-driven competitive sport.

The Chengdu finals demonstrated strong interest from students, hobbyists, and technology professionals.

Many participants treated the tournament as both a sporting contest and a technical showcase.

The event also reflected Chengdu’s reputation as a city receptive to new consumer technologies.

Organizers viewed the city as a suitable testing ground for the integration of “technology+sports,” citing its innovation culture and youth engagement.

The competition served as a platform for hands-on exposure to low-altitude aviation systems.

Teams refined control algorithms, flight stability, and collision management under real match conditions.

More than 200 volunteers supported the finals through Chengdu’s “Little Green Pepper” youth program.

University students assisted with athlete registration, timing, scoring, technical operations, and venue guidance. Their work enabled smooth event flow across multiple venues and match cycles.

Beyond competition results, the finals offered a window into how cities may integrate low-altitude technologies into public life.

The event connected sports consumption, aviation innovation, and youth participation at an unprecedented scale.

As drone soccer concluded its largest championship yet, organizers framed the event as a preview of how emerging aviation sports could develop alongside the expanding low-altitude economy.

🔗 Sumber: interestingengineering.com

📌 TOPINDIATOURS Eksklusif ai: ByteDance Introduces Astra: A Dual-Model Architectur

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