Development and evaluation of a wearable biofeedback system for real-time gait retraining in stroke rehabilitation

Project Abstract

This study presents the development and comprehensive evaluation of a wearable biofeedback system designed to support real-time gait retraining in individuals recovering from stroke. The system integrates inertial measurement units (IMUs), electromyography (EMG) sensors, and pressure-sensitive insoles to capture spatiotemporal gait parameters, paretic limb asymmetry, and lower-limb muscle activation patterns during ambulation. A custom lightweight ankle-foot orthosis housing the sensors provides comfort and secure data acquisition in varied walking conditions. Real-time feedback modalities include auditory cues, haptic alerts, and visual displays aimed at promoting symmetrical step length, adjusted temporal spacing, and improved push-off phase to facilitate more normal gait patterns. The hardware is paired with a mobile processing unit implementing an adaptive algorithm that estimates gait deviations, classifies error types (e.g., step length discrepancy, cadence mismatch, stance symmetry), and generates personalized feedback to the user. The software environment supports clinician-controlled parameterization, enabling tailored target gait profiles based on individual impairment severity and rehabilitation stage. A two-phase study design was employed. Phase I involved system validation with healthy adults to establish measurement reliability, sensor fusion accuracy, and latency of feedback delivery under laboratory and simulated daily-activity scenarios. Phase II conducted a randomized controlled trial with chronic stroke survivors assigned to either the wearable biofeedback intervention or conventional gait training. Outcome measures included gait speed, stride length symmetry index, temporal-spatial parameters, maximum toe clearance, integrated EMG activity of key lower-limb muscles, and functional mobility scores such as the Timed Up and Go and 6-Minute Walk Test. Assessments were performed at baseline, mid-intervention, immediately post-intervention, and at a 4-week follow-up to evaluate retention. Additionally, user experience and adherence were assessed through validated questionnaires and qualitative interviews. Results from Phase I demonstrated high test-retest reliability for stride length (intraclass correlation coefficient >0.92) and acceptable real-time processing latency (<120 ms) ensuring seamless feedback without perceptible delay. In Phase II, the intervention group showed statistically significant improvements compared with controls in gait speed (mean difference 0.14 m/s, p<0.01), single-stance time symmetry (p<0.05), and stance-to-swing ratio balance (p<0.05). Improvements in toe clearance reduced the risk of tripping, while EMG analysis indicated more coordinated activation patterns in the paretic limb during push-off and pre-swing phases. Functional mobility tests revealed meaningful gains in the Timed Up and Go and 6-Minute Walk Test distances (p<0.05). User feedback indicated favorable acceptability, with participants reporting perceived autonomy, reduced cognitive load during walking due to intuitive feedback, and high adherence rates (>85%). The study demonstrates that a wearable, multimodal biofeedback system can effectively facilitate real-time gait retraining after stroke, yielding measurable improvements in gait quality, functional mobility, and user engagement. Limitations include a relatively small sample size and short-term follow-up, suggesting the need for longer trials and exploration of integration with broader rehabilitation programs. Future work will explore personalized adaptation algorithms, machine learning-based optimization of feedback strategies, and scalability for home-based long-term use.

Project Overview

What This Project Is About

---

A plain-language overview of the topic and what the project investigates.

---

---

The Problem It Addresses

---

What problem or gap this project tackles and why it matters to the field or society.

---

---

Objectives of the Project

---

1. Understand how real-time feedback can influence walking patterns after stroke.
2. Design a simple wearable device that provides gait-related cues to users.
3. Test whether real-time feedback improves walking speed, balance, and symmetry.
4. Evaluate user comfort, ease of use, and acceptance by stroke survivors and clinicians.

---

---

What You Will Do Step by Step

---

1. Review basic gait rehabilitation concepts and select suitable gait metrics (e.g., step length, symmetry).
2. Choose hardware components (sensors, processor, feedback method) and build a prototype.
3. Develop software to collect sensor data and deliver real-time feedback signals.
4. Recruit or simulate stroke patient data to test the system in a controlled setting.
5. Run experiments to compare gait with and without feedback.
6. Analyze data to see changes in walking performance and user experience.
7. Refine the device based on results and feedback from users.
8. Document methods, results, and potential for clinical use.

---

---

---

---

Expected Outcome

Expected outcomes include a validated wearable prototype, evidence of improved gait metrics, and insights into user satisfaction and practical deployment in rehabilitation settings.