Advances in kinematic modeling and data-driven motion representation for biological and artificial systems

Kinematic modeling is central to understanding and interpreting motion across both biological and artificial systems. Traditionally underpinned by rigid-body mechanics, articulated chain models, and analytical formulations, the field has enabled major advances in how complex spatiotemporal behaviors are quantified in areas such as movement science, biomechanics, rehabilitation engineering, and robotics. Recent developments in high-resolution sensing, wearable technology, and computer vision have greatly enriched the volume and dimensionality of motion data available. However, these advances have also challenged existing models’ scalability, interpretability, and ability to capture individual and contextual variability. Existing approaches sometimes fall short in addressing uncertainty, identifiability, biomechanical realism, and generalization across domains. Cross-fertilization between fields—such as neuroscience and robotics—remains limited, signaling a need for methodologies and frameworks that unify traditional and modern perspectives.This Research Topic aims to foster the development of next-generation kinematic models that synthesize classical frameworks with data-driven and machine learning techniques. It seeks to encourage theoretical, computational, and applied research that bridges gap between analytical modeling and modern learning approaches, addressing core issues of accuracy, interpretability, adaptability, and cross-domain generalization. The goal is to stimulate interdisciplinary collaboration and accelerate the integration of novel sensor technologies, geometric learning principles, and hybrid physics-informed models—targeting diverse questions in biological motion, robotics, and human–machine interaction. Key questions include: How can we create unified representations that are robust across environments, morphologies, and individuals? Which approaches offer meaningful uncertainty quantification and biomechanical plausibility? How can emergent models inform diagnosis, control, clinical intervention, and intelligent systems?The scope of this Research Topic encompasses both foundational research and practical applications of kinematic modeling with a focus on unifying principles and transferable methodologies. Contributions may address models, algorithms, validation frameworks, or domain-specific applications, but should emphasize integration across disciplines and methodological innovation.To gather further insights in modeling, learning, and application boundaries, we welcome articles addressing, but not limited to, the following themes:o Analytical and numerical modeling of articulated and deformable systemso Data-driven or hybrid kinematic models incorporating physics-based and statistical constraintso Multi-sensor, multi-view, and 3D kinematic reconstructiono Kinematic modeling for human movement analysis, sports science, or rehabilitationo Motion representation, dimensionality reduction, and latent space approacheso Modeling uncertainty, robustness, and identifiability in kinematic estimateso Interpretable and kinematics-informed machine learning techniqueso Applications in robotics, human–robot collaboration, and assistive technologies Topic editors: Ziheng Wang, Guang Yang, Meiqi Wei Source: https://www.frontiersin.org/research-topics/77936/advances-in-kinematic-modeling-and-data-driven-motion-representation-for-biological-and-artificial-systems