Researchers at Tsinghua University developed PriorFusion, a unified framework that integrates semantic, geometric, and generative shape priors to significantly improve the accuracy and stability of road element perception in autonomous driving systems. The research addresses a long-standing challenge: existing end-to-end perception models often generate irregular shapes, fragmented boundaries, and incomplete road elements in complex urban scenarios.

Ride-pooling is widely recognized as a sustainable way to ease congestion, reduce costs and cut emissions, yet adoption remains limited.  When operators act independently, efficiency is low because requests cannot be matched across platforms.  Aggregation platforms seek to improve this by forcing all operators into a permanent coalition, but differences in size, cost and market position make such arrangements unstable.  To address this, researchers from Beihang University and Delft University of Technology developed a multi-level coalition formation game framework that enables coalitions to form dynamically in response to trip requests, allowing flexible cooperation without requiring all operators to remain in a single group at all times.

To answer this question: How to make AI truly scalable and reliable for real-time traffic assignment? A research team from KTH Royal Institute of Technology, Monash University, Technical University of Munich, Southeast University, and the University of Electro-Communications has developed a new framework—MARL-OD-DA—that offers a promising answer. The approach redesigns learning agents at the origin–destination (OD) level and utilizes Dirichlet-based continuous actions to achieve stable and high-quality solutions under dynamic travel demand.

To address the trade-off between accuracy and cross-city generalization in traffic flow estimation, a research team from The Hong Kong Polytechnic University and New York University proposes a novel framework based on global open multi-source (GOMS) data, including urban structures and population density. By developing an advanced graph neural network model that effectively fuses these static urban features with dynamic traffic data, the study achieves stable and accurate network-wide traffic estimation, as validated across 15 diverse cities in Europe and North America.

Researchers at National University of Singapore used multiple interpretable machine learning methods to predict traffic congestion in in Alameda County in the San Francisco Bay Area, USA, during the pre-lockdown, lockdown, and post-lockdown periods.

The development of bioinspired electronic skin plays a pivotal role in enhancing robotic environmental perception and interaction capabilities, with stretchable multimodal tactile sensors serving as the fundamental component. However, existing tactile sensors are often constrained by limited integration density and spatial resolution, hindering their applicability in complex scenarios. To address these challenges, this study proposes a multimodal tactile sensing strategy based on the synergistic integration of pressure and strain sensors. By innovatively embedding strain sensors into the gaps between pressure sensor units, both types of sensors are co-fabricated in a coplanar configuration, enabling simultaneous and high-precision detection of pressure and strain. Leveraging the dual-mode sensing data, the system further enables accurate evaluation of object hardness.

Lithium metal batteries hold great promise for high performance energy storage due to their high theoretical energy density. However, practical implementation is hindered by interfacial side reactions and dendrite growth at the Li metal anode, particularly in carbonate-based electrolytes. Hereby, the authors introduce a novel multifunctional group additive strategy using 2-fluorobenzenesulfonamide (2-FBSA) to address these challenges. The 2-FBSA additive plays a crucial role in modulating the solvation structure of the electrolyte, facilitating Li⁺ transport kinetics by lowering the desolvation energy barrier. Additionally, the preferential decomposition of 2-FBSA at the anode interface leads to the formation of a robust solid electrolyte interphase (SEI) enriched with inorganic Li salts, including LiF, Li₃N, and ROSO₂Li. This SEI layer effectively suppresses Li dendrite growth and mitigates parasitic side reactions, resulting in significantly improved cycling stability and rate performance of Li||Li symmetric cells and Li||LiFePO₄ full cells. The Li||Li symmetric cell achieves a remarkable lifespan exceeding 2400 h at 0.5 mA cm⁻²/1 mAh cm⁻² , while the Li||LiFePO₄ full cell demonstrates a capacity retention of 72% after 400 cycles at 1 C. This study highlights the potential of multifunctional group molecular additive 2-FBSA in interfacial optimization and provides new insights into additive design principles for high performance battery systems.

Developing multifunctional electromagnetic wave absorbing materials capable of operating in complex environments has become crucial for electromagnetic protection. Introducing dielectric ceramic materials with high thermal stability and oxidation resistance into carbon matrices has emerged as an effective strategy to address the high-temperature failure of carbon-based absorbers. In this study, a novel hollow SiC/C nanofiber aerogel was successfully constructed via a hydrothermal-carbothermal reduction method. The material exhibits an ultralight nature, good elasticity, fatigue resistance, high-temperature stability, and excellent electromagnetic wave absorption performance.

Heterojunction structures are favored for constructing photoelectrochemical ultraviolet photodetectors (PEC UV PDs), whereas lattice mismatches impede their optoelectronic performance. This work presents a novel homojunction consisting of two-dimensional (2D) In₂O₃ nanosheets (NS) and three-dimensional (3D) In₂O₃ microcubes (MC) with a suitable energy band alignment. 2D In₂O₃ NS not only shows an enlarged bandgap due to the quantum confinement effect but also effectively upshifts the conductive band and Fermi level stemming from the oxygen vacancy demonstrated by the theoretical simulation and experimental results. The photogenerated carrier dynamic of In₂O₃ photoanodes is boosted by the 2D-3D homojunction with a built-in electric field and more electrochemically active sites, leading to higher photogenerated carrier separation efficiency, faster interfacial charge transfer, and better self-powered capability. The In₂O₃ 2D-3D homojunction PEC UV PDs exhibit outstanding self-powered deep-UV photoresponse at 0 V, with an ultrahigh responsivity of 316.5 mA/W for 254 nm light, a fast response speed of 15/15 ms, high detectivity of 1.12 × 10¹² Jones, and an outstanding UV-vis rejection ratio of 1507, surpassing most recorded PEC UV PDs. This work demonstrates the pivotal role of morphology-controlled homojunction in modulating photogenerated carrier dynamics and offers a new strategy for designing high-performance PEC devices.