Research Analyzer
← Back ICRA 2026

Error Fields: Personalized Robotic Training to Enhance Movement Accuracy across Speed

Arturo Ramirez, , Courtney Celian , James L. Patton

PDF

AI summary

Key figure (auto-extracted from paper)
Personalized Error Fields training significantly improves movement accuracy and induces motor adaptation after-effects more effectively than traditional error augmentation or sham conditions.
Robotic rehabilitation motor adaptation personalized therapy error augmentation neurorehabilitation 3D reaching

Problem

Robotic rehabilitation often lacks personalization, applying generic treatments that don't account for individual movement error patterns, limiting motor learning and recovery.

Approach

The researchers developed Error Fields, a robotic training paradigm that maps a user's specific trajectory errors during a 3D reaching task and applies personalized force perturbations to amplify those exact errors, compared against traditional error augmentation and a sham control.

Key results

  • Highest accuracy improvement across movement speeds
  • Significant motor adaptation after-effects upon force removal
  • Superior performance compared to traditional error augmentation and sham controls
  • Personalized error amplification drives stronger motor learning than constant amplification

Why it matters

This personalized robotic training paradigm offers a promising, targeted approach for neurorehabilitation, particularly for post-stroke patients needing speed-accuracy tradeoff improvements.

Abstract

We developed a personalized robotic training paradigm called Error Fields, which learns the user mistakes during an upper-limb reaching task and generates perturba- tions that drive the subject’s trajectories toward greater preci- sion. We hypothesized that force fields pushing the limb toward regions of largest error tendencies would intensify practice, accelerating motor adaptation and improving the speed and accuracy of the subject’s movements. To test this hypothesis, we analyzed error reduction across the experiment by comparing groups that received Error Fields with those subjected to tradi- tional Error Augmentation and a null-force control condition. In all conditions, participants were additionally exposed to a curl field to simulate motor impairment. Error Fields showed improved accuracy post-training and induced after-effects once the force was removed, demonstrating promising benefits for investigation in a post-stroke population.

Index terms

Rehabilitation Robotics Sensorimotor Learning Human Performance Augmentation

Related papers