Using Step Size and Lower Limb Segment Orientation from Multiple Low-Cost Wearable Inertial/Magnetic Sensors for...

Area of Science:

• Robotics
• Sensor Technology
• Biomechanical Engineering

Background:

• Indoor positioning remains a challenge, particularly for pedestrian navigation.
• Existing systems often rely on expensive infrastructure or complex algorithms.
• Low-cost sensors offer a potential solution for accessible pedestrian navigation.

Purpose of the Study:

• To develop and evaluate a pedestrian navigation system using multiple low-cost inertial/magnetic sensors.
• To investigate dead-reckoning methods for indoor positioning.
• To compare different step size estimation techniques.

Main Methods:

• Utilized low-cost inertial/magnetic sensors for motion data acquisition.
• Employed a five-segment skeletal model for lower limb forward kinematics.
• Estimated limb orientation angles (pitch) from inertial measurements.
• Developed a sensor data logging system for comprehensive motion recording.
• Validated the system using treadmill walk experiments with optical motion capture.

Main Results:

• Mean error for estimated limb orientation angles was less than 6 degrees.
• Mean error for step length estimation was 3.2 cm.
• Mean error for left stride length was 12.5 cm; for right stride length was 9 cm.
• Achieved an expected positioning error of less than 5% of the total distance traveled.

Conclusions:

• Multiple low-cost sensors can effectively form a pedestrian navigation system for indoor positioning.
• The proposed method accurately estimates step and stride lengths using skeletal kinematics.
• The system demonstrates a low positioning error, making it practical for real-world applications.