The field of human mobility and autonomous systems is rapidly evolving, with a focus on developing more accurate and equitable models for predicting human behavior and improving autonomous vehicle performance. Recent research has highlighted the importance of considering demographic disparities in mobility prediction models, with some studies proposing novel sampling strategies to reduce performance gaps between different user groups. Additionally, there is a growing interest in developing more realistic models of pedestrian-driver interactions, incorporating human perceptual-motor constraints and multi-agent reinforcement learning. Other notable trends include the use of multimodal feature integration and stacked ensemble-based architectures for predicting post-wildfire vegetation loss, and the development of unified models for human mobility generation in natural disasters. Noteworthy papers in this area include 'Mind the Gaps: Auditing and Reducing Group Inequity in Large-Scale Mobility Prediction', which proposes a fairness-aware sampling strategy to reduce demographic performance gaps, and 'Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving', which introduces a vision-language-action model that integrates Chain of Causation reasoning with trajectory planning.
Advances in Human Mobility and Autonomous Systems
Sources
Mind the Gaps: Auditing and Reducing Group Inequity in Large-Scale Mobility Prediction
Realistic pedestrian-driver interaction modelling using multi-agent RL with human perceptual-motor constraints
Beyond Demographics: Behavioural Segmentation and Spatial Analytics to Enhance Visitor Experience at The British Museum
MVeLMA: Multimodal Vegetation Loss Modeling Architecture for Predicting Post-fire Vegetation Loss
GEPOC Parameters - Open Source Parametrisation and Validation for Austria, Version 2.0
Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail
Forecasting Occupational Survivability of Rickshaw Pullers in a Changing Climate with Wearable Data
Physiologically Active Vegetation Reverses Its Cooling Effect in Humid Urban Climates
Dynamic Model Selection for Trajectory Prediction via Pairwise Ranking and Meta-Features
X-TRACK: Physics-Aware xLSTM for Realistic Vehicle Trajectory Prediction
Unveiling Uniform Shifted Power Law in Stochastic Human and Autonomous Driving Behavior
Embodied Cognition Augmented End2End Autonomous Driving
A Unified Model for Human Mobility Generation in Natural Disasters
Dynamic Population Distribution Aware Human Trajectory Generation with Diffusion Model
On Predicting Sociodemographics from Mobility Signals
Learning to Land Anywhere: Transferable Generative Models for Aircraft Trajectories
Studying the Effect of Explicit Interaction Representations on Learning Scene-level Distributions of Human Trajectories
SAFe-Copilot: Unified Shared Autonomy Framework