A Low-Cost Adaptive Smart Sock for Assessing Walking and Running Level and Prescribing Progressive Cardiorespiratory Training

Technical Summary: A Low-Cost Adaptive Smart Sock for Assessing Walking and Running Level and Prescribing Progressive Cardiorespiratory Training

Problem Statement
The research addresses a critical gap in public health and exercise guidance: beginner users often lack an objective, low-cost, and simple mechanism to assess their current walking or running capacity. Without this assessment, individuals may progress too quickly, leading to fatigue, pain, and poor adherence, or remain at low activity levels with limited cardiorespiratory improvement. Existing mobile applications typically prescribe fixed plans based on time, GPS distance, or manual input, failing to capture real-time foot motion, rhythm stability, or smooth transitions between walking and running. Furthermore, traditional gait analysis relies on expensive laboratory equipment (motion capture, force plates) that is inaccessible for everyday use. The core research question is how a low-cost smart sock can objectively assess a user's locomotion level and generate an adaptive walk-to-run program to improve cardiorespiratory fitness without replacing clinical diagnostics.

Methodology and System Architecture
The proposed solution is a "minimum viable research prototype" consisting of a removable in-shoe foot module and a companion mobile application. The system prioritizes adaptive software logic over expensive sensing hardware.

• Hardware Components: The foot module integrates a single Inertial Measurement Unit (IMU) to measure acceleration and angular velocity, and a single forefoot Force-Sensitive Resistor (FSR) to detect foot-ground contact and push-off timing. These sensors connect to a Bluetooth Low Energy (BLE) microcontroller (e.g., ESP32-C3 or XIAO nRF52840) powered by a small rechargeable LiPo battery (100–150 mAh). The module is housed in a protective enclosure (TPU/silicone or textile pouch) to resist sweat and impact.
• Data Processing and Logic: The mobile application handles the complex processing, including baseline classification, training plan generation, and weekly adaptation. The foot module transmits raw or summarized metrics via BLE.
• Baseline Assessment: The system initiates with a 6–12 minute baseline test where the user performs comfortable walking, brisk walking, and potentially a brief light jog. The app records cadence, estimated speed (derived from cadence and calibrated step length), ground-contact timing, rhythm stability, and a Rating of Perceived Exertion (RPE) score (1–10).
• User Classification: Based on the baseline, users are assigned a conservative training level (Level 0 to Level 3). Level 0 users begin with comfortable walking only; Level 1 introduces short jogging intervals; Level 2 and 3 progress to longer running intervals.
• Adaptive Algorithm: The training logic follows the FITT-VP principle (Frequency, Intensity, Time, Type, Volume, Progression). The algorithm evaluates session completion (targeting >80%), cadence stability, contact timing, and RPE.
  • Progression: Occurs only if the user completes sessions, maintains stable rhythm, reports acceptable exertion, and has no warning symptoms (e.g., chest pain, dizziness).
  • Regression/Repetition: If completion is partial, pacing is unstable, or symptoms arise, the plan repeats the current week or reduces intensity, advising professional consultation if necessary.
• Interface: The application features four screens: Baseline Assessment, Today's Workout, Live Session Feedback (displaying gait state, cadence, and interval countdown), and Weekly Review.

Key Contributions
The paper's primary contribution is the integration of inexpensive foot sensing with adaptive exercise prescription rules to create a beginner-oriented coaching tool.

1. Sensor Minimalism: The study demonstrates that a minimal sensor set (one IMU, one FSR) is sufficient for classifying locomotion states and monitoring rhythm stability, avoiding the need for full plantar pressure maps or clinical-grade force sensors.
2. Adaptive Coaching vs. Data Display: Unlike standard step counters, the system functions as a "digital adaptive coach." It uses gait-derived indicators (cadence, contact timing, regularity) combined with subjective RPE to dynamically adjust training load.
3. Safety and Accessibility: The system is explicitly designed as a wellness tool, not a medical device. It incorporates safety protocols (symptom checks, conservative progression) to guide apparently healthy users while directing high-risk individuals to seek professional advice.
4. Cost-Effective Design: By shifting intelligence to the mobile app and using widely available development modules, the prototype aims to lower barriers to entry for cardiorespiratory training.

Expected Outcomes and Limitations
The expected outcome is a feasible prototype capable of detecting steps, estimating cadence, distinguishing between walking and light running, and generating a conservative adaptive training plan. The system is not intended to provide laboratory-level precision in speed or distance estimation, nor does it estimate VO2max or replace clinical cardiopulmonary testing.

The paper acknowledges several limitations:

• Speed Estimation: Derived from cadence and step length, it is approximate and affected by user differences, footwear, and terrain.
• Sensor Constraints: A single forefoot FSR detects contact events but cannot measure full plantar pressure or exact force.
• Physiological Data: The absence of heart-rate sensing limits physiological intensity estimation, relying instead on RPE and movement metrics.
• User Dependency: The system's efficacy depends on honest user reporting of exertion and symptoms.

Significance
The paper positions the smart sock as a public-health technology designed to lower barriers to physical activity. Its significance lies not in inventing new sensors, but in combining existing low-cost components with adaptive software to provide actionable, personalized guidance for transitioning from inactivity to structured aerobic training. By focusing on gradual, sustainable progression based on objective gait data and subjective feedback, the system aims to help beginners build regular aerobic habits consistent with global health guidelines (150 minutes of moderate activity per week). The work serves as a proof-of-concept for a graduation project, startup prototype, or applied wearable-health study, with future iterations potentially incorporating heart-rate integration, GPS fusion, and machine learning.