HikingTTE: a deep learning approach for hiking travel time estimation based on personal walking ability
Reports and Proceedings
Updates every hour. Last Updated: 4-May-2025 20:09 ET (5-May-2025 00:09 GMT/UTC)
The research team at the University of Electro-Communications has introduced "HikingTTE," a deep learning model that accurately predicts hiking travel times. Traditional methods typically use slope-speed relationships but fail to account for individual ability and accumulated fatigue. HikingTTE integrates a modified Lorentz-based slope-speed function with LSTM and attention modules, adapting to each hiker’s walking data. Experiments showed a 12.95 percentage point reduction in Mean Absolute Percentage Error (MAPE) compared to established models and an additional 0.97 point improvement over other deep learning approaches. By enabling safer, more accurate planning, it has the potential to reduce mountain accidents and revolutionize hiking time estimation.
Primates scratch their skin or fur in a stressful situation (Figure 1). By using this self-scratching behaviour as a behavioural indicator of stress, our study on wild female Japanese macaques provided a novel and counter-intuitive finding: females exhibit higher stress levels when a closely related female is nearby. This pattern was found only during foraging but not resting, suggesting that food-related competition among relatives is a cause of their elevated stress. This unexpected finding provides new insight into the complex social dynamics of primates and their stress responses.
A Japanese version of the U.S. Esophageal Hypervigilance and Anxiety Scale questionnaire to investigate the psychological impact on esophageal symptoms has been developed and validated by Osaka Metropolitan University researchers.
Led by Assistant Prof. Kou Li, a research group in Chuo University, Japan, has developed an all-printable device fabrication strategy to resolve the existing technical limitations of multi-functional image sensor sheets for non-destructive inspections, with a recent paper publication in Small Science.
Researchers developed crossing matrices to classify doubly periodic weaves, addressing challenges in mathematical modeling of woven structures. This breakthrough enables systematic analysis of periodic entanglements, with applications in materials science, fabric design, and nanostructure development.
Mechanochemistry using a ball mill demonstrates versatility for generating academically and industrially significant organolithium compounds.