cs.LGNov 11, 2026

Subject-Held-Out Evaluation of Deep Ground Reaction Force Estimators Across Two Laboratories: Leakage and Cross-Laboratory Transfer

Authors: Hernan David Santos, Yinong Chen, Yu Ding, Xudong Zhang, Jian Tao

Organizations: Department of Electrical and Computer Engineering and the Texas A&M Institute of Data Science, Texas A&M University, College Station, TX 77843 USA · Department of Environmental and Occupational Health, Texas A&M University, College Station, TX 77843 USA · H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA 30332 USA · Wm Michael Barnes ’64 Department of Industrial and Systems Engineering, Texas A&M University, College Station, TX 77843 USA · 

Abstract

Deep networks that estimate ground reaction forces and centers of pressure from motion-capture markers are validated within one laboratory, in a seldom-stated coordinate frame, and sometimes with trial-level splits that put the same person in the training and the test set. We measured what each practice costs, and whether such models transfer between laboratories. Three sequence models (convolutional, convolutional-recurrent, and convolutional-transformer) were evaluated on two independent datasets, locomotion (6 subjects) and occupational patient handling (10 subjects), under one pipeline, one tuning procedure, three seeds, and four protocols: trial-level split, leave-one-subject-out, pooled two-laboratory training with matched held-out subjects, and cross-laboratory transfer. At a fixed configuration, trial-level splits overstated the subject-held-out coefficient of determination by +0.040 to +0.194; on occupational data the overstatement exceeded every difference between architectures. With subjects held out, the three models reached values between 0.583 and 0.804; the leader differed between datasets, and gaps lay within subject-to-subject variability. Pooling the laboratories changed neither dataset on matched subjects. Transfer between laboratories was below chance in both directions, including the vertical force, which no in-plane rotation can affect. Leaving force targets in the plate frame while inputs were heading-aligned cost 0.57 of locomotion anteroposterior-force explained variance, enough to make that force look hardest; a Newton's-law correlation detected the mismatch before training. Split protocol and target frame changed reported accuracy more than architecture did. Learned force estimators should be validated with subjects held out, checked for frame consistency before training, and tested across laboratories before deployment claims are made.

Figures & tables

Explore similar work

CardsList
  1. Gait2Hip-60: A Unified Deep Learning Benchmark for Predicting Hip Muscle Forces and Joint Moments from Multi-Cadence Gait Kinematics

    May 24, 2026Jiaqi Zhang, Ji Hou, Qing Sun +2Gait AnalysisSequence Modeling

  2. Deep Learning-Based Estimation of Ground Reaction Forces in Parkinsonian Gait Using an Optimized Set of IMU Data

    Aug 3, 2026Run Lin, Yingtian Tang, Jiawen Xu +6ground reaction force estimationGait Analysis

  3. Proprioceptive Force Estimation for Quadruped Locomotion and Human-Robot Interaction

    Sep 28, 2026Run Wang, Xu Yang, Alapati Tuerxun +1Robotic ControlHuman-Robot Interaction