Multisensor approach to walking distance estimation with foot inertial sensing

Area of Science:

• Biomechanics and Human Movement Science
• Sensor Technology and Data Fusion
• Robotics and Navigation

Background:

• Accurate walking distance estimation is vital for applications like gait analysis, sports training, and pedestrian localization.
• Attaching portable inertial sensors to user shoes is a practical approach for gait monitoring.
• Estimating step length using biaxial accelerometers and gyroscopes on the sagittal plane is common, but direct signal integration leads to significant errors.

Purpose of the Study:

• To investigate a multisensor model approach for reducing uncertainty in walking distance estimation.
• To improve the accuracy of gait monitoring systems by minimizing error accumulation.
• To present experimental results validating the proposed multisensor fusion technique.

Main Methods:

• Utilizing a multisensor model that integrates data from multiple inertial measurement units (IMUs).
• Employing sensor fusion techniques to mitigate unbounded error growth inherent in direct signal integration.
• Attaching sensors to user shoes for natural and unobtrusive gait data acquisition.

Main Results:

• The multisensor model approach significantly reduces estimation uncertainty compared to single IMU systems.
• Experimental results demonstrate a lower standard deviation in walking distance estimation.
• The proposed method effectively addresses the error accumulation problem in step length computation.

Conclusions:

• Multisensor fusion is a viable and effective method for enhancing the accuracy of walking distance estimation.
• Attaching IMUs to shoes combined with advanced fusion algorithms offers a robust solution for gait analysis and localization.
• The developed approach provides a more reliable estimation of walking distance with reduced error.