---
title: "Deep Learning-Based Estimation of Ground Reaction Forces in Parkinsonian Gait Using an Optimized Set of IMU Data"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.02408v1"
permalink: "https://feed.astralor.com/items/itm_fb6e41ecaa945d1be4de3e79"
published: "2026-08-03T15:49:50.000Z"
score: 78
channel: "ai"
tags: ["CNN-BiLSTM","Wearable Devices","Parkinson's Disease","IMU","Gait Analysis"]
---

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

## AI commentary

> This paper's value lies in applying deep learning to a real pain point in Parkinsonian gait analysis — obtaining reliable kinetic metrics with minimal sensors. For medical AI and wearable device developers, the sensor configuration optimization findings can be directly translated into product design guidance. Despite being a preprint, the data scale and consistency make the conclusions quite convincing.

## At a glance

This study pioneers deep learning for estimating bilateral vertical ground reaction forces (vGRFs) in Parkinson's disease (PD) gait, using a hybrid CNN-BiLSTM model trained on 13 IMU data from 61 PD patients and 65 healthy controls (HC). The model achieved high intra-subject accuracy (R²=0.98) and strong inter-subject generalization (R²=0.93 for HC, R²=0.91 for PD). Sensor configuration significantly affected accuracy; the optimal set for PD was four IMUs, while a minimal two-IMU setup still enabled robust estimation, offering a practical and scalable path for wearable gait analysis systems.

## Key points

- First application of deep learning to vGRF estimation in PD gait; hybrid CNN-BiLSTM performs excellently on both PD patients and healthy controls.
- Optimal IMU configurations identified as four for PD and two for HC; significant sensor reduction with maintained robustness.
- Enables remote monitoring and personalized rehabilitation, potentially replacing lab-based gait analysis for mobility-impaired PD patients.

## Full text

Accurate gait analysis in Parkinson's disease (PD) typically relies on laboratory-based systems to capture biomechanical data, such as ground reaction forces (GRFs). Estimating GRFs using inertial measurement units (IMUs) provides a feasible alternative. However, this approach remains challenging in pathological gait like PD due to its high variability and complexity. Moreover, existing monitoring approaches often require multiple body-mounted sensors, which limit practicality and reduce patient compliance. To date, no study has investigated the application of deep learning approaches to address this challenge. This study proposes, for the first time, a deep learning framework to estimate bilateral vertical GRFs (vGRFs) in PD using an optimized set of wearable IMUs. A hybrid CNN-BiLSTM model was trained separately on data from 61 PD patients and 65 healthy controls (HC) using 13 IMUs. The model achieved high intra-subject accuracy ($R^2$ = 0.98) and strong inter-subject generalization ($R^2$ = 0.93 for HC, $R^2$ = 0.91 for PD). Sensor configuration was found to significantly influence estimation accuracy, with optimal sensor placement varying between PD patients and HC. For PD patients, estimation accuracy dropped markedly when reducing to a single IMU. The optimal configuration for PD used four IMUs. We identified a minimal setup with only two IMUs still enabled robust estimation. This compact setup offers a practical and scalable solution. Overall, the proposed approach supports the development of wearable vGRF-based gait analysis systems for Parkinsonian gait and potentially other pathological conditions, enabling accessible clinical assessments, remote monitoring, and personalized rehabilitation.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_fb6e41ecaa945d1be4de3e79