Capstone Design 2025 Fall to 2025 Spring

Department of Embedded System Engineering

Incheon National University

This project presents the design and implementation of a semi-autonomous mobility aid device that recognizes the user's walking intention using a fixed-handle pressure sensor. It was developed as part of a capstone design project in the Department of Embedded Systems Engineering at Incheon National University.

The proposed system aims to support the elderly and people with walking difficulties by providing intuitive control through pressure applied to a fixed handle. Unlike traditional manual walkers or powered ride-on devices, this system seeks to balance cost-efficiency, safety, and autonomy.

• Manual walkers require physical force, posing safety risks especially on slopes.
• Powered devices reduce physical burden but are expensive, bulky, and reduce physical activity.
• This system offers an in-between solution: a cost-effective, semi-autonomous device that promotes active walking with intuitive control.

The device integrates the following components:

• FSR Array Sensors (Fixed Handle): Detects the user's grip pressure to infer walking intentions (forward, backward, turn).
• Gyroscope (MPU-9250): Measures slope angle to adjust motor output on inclined surfaces.
• Hall Effect Sensors: Measures wheel speed using custom-built encoders with neodymium magnets.
• MCU (Arduino Nano): Handles real-time sensor input and motor control via PWM.
• SBC (Raspberry Pi 5): Performs high-level processing and communicates with MCU over USB using JSON-formatted serial data.

• Two cylindrical FSR sensor arrays (16x16 each) are mounted on a PVC pipe used as the handle.
• Dual DC motors with gearboxes are mounted on the rear wheels of a modified walker.
• Hall sensors detect wheel rotation for speed calculation.
• 12V 18Ah lead-acid battery powers the motors and electronics.

• MCU reads sensor data and sends it to the SBC.
• SBC parses the data in Python and calculates the target wheel speeds based on pressure distribution and terrain slope.
• Final PWM commands are sent back to the MCU to control motor drivers.

A user test was conducted with 5 participants aged in their 50s and 60s. Test scenarios included:

• Forward and backward movement
• Uphill and downhill slopes
• Emergency stops

User satisfaction survey results (out of 5.0):

• Forward stability and speed control: 4.8
• Downhill safety and control: 5.0
• Braking reliability: 5.0
• Overall ease of use and safety: 4.6+

This project demonstrates a practical and scalable approach to mobility assistance for the elderly and physically challenged. The intuitive sensor-based control and simple architecture make it suitable for both personal and industrial use (e.g., logistics carts). Future work may include machine learning-based intent recognition and integration with vision systems.