Ben M. Chen

Large-scale ocean exploration with underwater robots is constrained by energy budgets, intermittent communications, and environmental dynamics. To address these challenges, this article presents a novel long-endurance underwater robot equipped with online spatiotemporal extended perception and adaptive dynamics modeling. The robot integrates efficient buoyancy propulsion and center-of-mass regulation, enabling effective underwater autonomous navigation. The online spatiotemporal extended perception strategy reconstructs large-scale environmental fields from information-optimal sparses observations. The environment-coupled adaptive dynamics incorporates explicit environmental couplings and performs online residual correction to account for implicit environmental disturbances. Field trials conducted in the South China Sea and the Pacific Ocean have demonstrated the robot's reliable long-endurance autonomy, its ability to achieve basin-scale perception with just 1% sampling density, and a reduction in motion error by over 71%. This research lays the foundation for low-cost, large-scale ocean exploration.

While coordination among multiple flying vehicles improves aerial logistics efficiency, safe and orderly operation requires robust detection and collision-avoidance capabilities. These requirements apply to passenger aircraft as well as to urban traffic and maritime environments, including emerging underwater flying vehicles. However, existing detection methods often fail under challenging conditions such as low illumination or cluttered backgrounds. Their progress is further constrained by the lack of large-scale benchmarks and the high computational and memory costs required to achieve high accuracy, which limits their deployment in resource-constrained scenarios, such as air-to-air collision avoidance in aerial vehicles. To address this gap, we introduce FT55k, an open-source benchmark comprising over 55,000 annotated RGB and infrared images across diverse environments. We further provide baseline approaches tailored for platforms with different computational demands. Extensive experiments on FT55k and three public datasets demonstrate the superior accuracy and efficiency of our methods compared with state-of-the-art approaches. Notably, our approach is the first flying vehicle detection method with a computational cost below 0.5 BFLOPs, achieving real-time performance at 62.3 FPS on an edge-computing device. This work presents the first comprehensive benchmark for flying vehicle detection in complex environments, establishing a practical and scalable foundation for future research and deployment in intelligent transportation safety. Our datasets is publicly accessible at https://github.com/chriszxk/Flying-Vehicle-Detection

Achieving carbon neutrality in the built environment requires advanced heating, ventilation, and air conditioning (HVAC) control strategies that optimize both energy efficiency and occupant thermal comfort. Co-simulation frameworks integrating building energy simulation (BES) with computational fluid dynamics (CFD) represent a promising approach for developing such control strategies, yet their implementation remains expertise-demanding, time-consuming, and error-prone. This study introduces AutoCoSim, a novel multi-agent framework that automates BES-CFD co-simulation by synergistically leveraging Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). Embedded within a digital twin platform, AutoCoSim translates natural language inputs into autonomous workflows encompassing model configuration, co-simulation execution, and performance analysis. The core innovation lies in a scalable architecture comprising a specialized agent-tool library, a hierarchical multi-level network linking distributed agents, and a unified execution protocol for orchestrating multiple simulation engines. Built upon the lightweight Model Context Protocol (MCP), the framework enables interoperable collaboration between users, agents, computational tools, and external knowledge bases. Validation across six categories of co-simulation tasks (60 experiments) demonstrates that AutoCoSim effectively decomposes user queries into executable subtasks, achieving 100% execution success rate with near-perfect syntax validity across all generated configuration files. The framework successfully formulates and evaluates energy- and comfort-aware building control strategies through parametric simulation. These results confirm AutoCoSim’s efficacy in orchestrating complex multi-physics co-simulations and its significant potential to streamline the optimal control strategy development in real-world, IoT-enabled buildings.

In this work, we present Tilt-Ropter, a novel hybrid aerial-terrestrial vehicle (HATV) that combines tilt rotors with passive wheels to achieve energy-efficient multi-mode locomotion. Unlike existing under-actuated HATVs, the fully actuated design of Tilt-Ropter enables decoupled force and torque control, greatly enhancing its mobility and environmental adaptability. A nonlinear model predictive controller (NMPC) is developed to track reference trajectories and handle contact constraints across locomotion modes, while a dedicated control allocation module exploits actuation redundancy to achieve energy-efficient control of actuators. Additionally, to enhance robustness during ground contact, we introduce an external wrench estimation algorithm that estimates environmental interaction forces and torques in real time. The system is validated through both simulation and real-world experiments, including seamless air-ground transitions and trajectory tracking. Results show low tracking errors in both modes and highlight a 92.8% reduction in power consumption during ground locomotion, demonstrating the system's potential for long-duration missions across large-scale and energy-constrained environments.

Fisheye cameras, with their wide field of view, are highly beneficial for applications such as unmanned systems. Solving the perspective- n -point (P n P) problem that estimates the pose of a camera is particularly important for fisheye cameras. However, most existing P n P methods are primarily developed for pinhole cameras, with limited research focusing on fisheye cameras. This study presents an efficient method for solving the P n P problem for fisheye cameras. The proposed method introduces a mapping from a fisheye camera image to a virtual image plane based on the geometric properties of the fisheye camera. Two-dimensional reference points on the virtual plane are then employed to solve the P n P problem. Error analysis validates the effectiveness of the mapping through numerical computation. The experimental results further demonstrate that the proposed method achieves better accuracy and efficiency than the existing methods.