This study presents KEPT, an AI system that helps self-driving cars predict their own short-term path more safely by combining video understanding with a memory of similar past scenes. Tested on the public nuScenes benchmark, KEPT cuts prediction errors and potential collisions compared with existing planning methods, while using a fast, lightweight retrieval module that is practical for real-time driving.

To address the growing conflict between personalized mobility analysis and data privacy, researchers have developed IPC-FM, a novel federated meta-learning framework. This approach enables accurate travel behavior prediction without centralizing sensitive user data. By integrating interpretable neural networks with rapid model adaptation, IPC-FM provides a customizable solution that significantly outperforms current state-of-the-art methods, ensuring individual mobility needs are met securely and transparently.

Researchers at Beihang University, China, introduce a new task setting: latency-aware trajectory prediction for autonomous driving, which explicitly accounts for the latency issue and transforms it from a hindrance into an opportunity for enhanced performance.

How can autonomous vehicles continuously learn new traffic scenarios without forgetting previously learned ones? Researchers from Tsinghua University have proposed a dynamically expandable learning framework for interactive trajectory prediction. The method enables models to adapt to evolving traffic environments while preserving performance on earlier scenarios. Experiments on real-world datasets show that the approach effectively mitigates catastrophic forgetting, especially for safety-critical driving cases.

As the demand for constructing lunar and Martian bases continues to rise, lava tubes—with their unique advantages such as natural shielding from cosmic radiation, thermally stable conditions, and ready-to-use subsurface living spaces—have become a core consideration for deep space exploration and the selection of long-term extraterrestrial base sites. Compared to traditional methods relying solely on surface rovers or single-sensor orbital identification, future scientific exploration of lunar and Martian lava tubes requires a systematic approach to address key questions: "Where are they?", "What do they look like?", "How do we explore them?", and "How do we use them?" This necessitates the establishment of a comprehensive, multi-dimensional detection system.

Recently, a study published in the journal Space: Science & Technology focused on the Jingpo Lake lava tube as a typical terrestrial analog site. Led by China University of Geosciences (Beijing) in collaboration with domestic and international research teams, including the Aerospace Information Research Institute, Chinese Academy of Sciences; Heilongjiang Second Surveying and Mapping Engineering Institute; Peking University; Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences; Chengdu University of Technology; and the University of Padova, Italy, a comprehensive five-year scientific investigation was conducted. Leveraging the Jingpo Lake lava tube network in Heilongjiang Province and taking advantage of the environmental conditions during winter when liquid water is absent—thereby simulating lunar lava tube exploration scenarios—this study carried out multi-sensor, integrated ground-air-space surveys. For the first time, an integrated ground-air-space exploration scheme for lava tubes was proposed. This scheme integrates multi-source detection technologies, including spaceborne synthetic aperture radar (SAR), UAV-based close-range photogrammetry, airborne LiDAR, in-tube GeoSLAM, hyperspectral LiDAR, and ground-penetrating radar (GPR). A multi-platform, multi-scale collaborative survey of the Jingpo Lake lava tube area was conducted, establishing a complete technical chain from surface skylight identification and subsurface void detection to the precise acquisition of in-tube geometric and spectral information. This work provides a robust terrestrial analog validation foundation and technical reference for future comprehensive lunar lava tube exploration.

Researchers have developed a rapid colour-changing test that can distinguish between different strains of golden staph, including those likely to be virulent and antibiotic resistant.

Golden staph is a major human pathogen and is a leading cause of infection-related deaths globally, with more than a million fatalities each year.

Mandatory lane changes at intersections often lead to intricate conflicts and traffic oscillations.  The advent of connected and autonomous vehicles (CAVs) is expected to mitigate these disruptions by coordinating acceleration and lane-change behaviors. Addressing this, researchers developed SS-MA-PPO, a novel Multi-Agent Reinforcement Learning (MARL) framework that assists CAVs in coordinating these critical decisions. Evaluated against a real-world dataset from Langfang, this method significantly improves traffic efficiency compared with traditional models and other Multi-Agent Reinforcement Learning baselines.