Stride Length Estimation Using ANN

Date
2018
Authors
Liu, Yu
Supervisor
Huang, Loulin
Item type
Thesis
Degree name
Master of Computer and Information Sciences
Journal Title
Journal ISSN
Volume Title
Publisher
Auckland University of Technology
Abstract

Measuring the stride length of a pedestrian is an essential task for many applications, such as augmented reality (AR) applications tracking devices and motion monitoring devices. In the traditional position estimation method such as the double integral of the acceleration signal, the signal noise of the acceleration from the Inertial measurement units (IMU) leads to a cumulative error. It is the bottleneck of the positioning perform- ance in most of Inertial Navigation System (INS). Moreover, in many applications, such as AR and tracking devices, sensors are attached to the human body and used in the indoor environment, which is hard to obtain the position of the device through GPS or traditional INS. In this paper, we proposed a novel method that transforms the position information to step length, and then we use Artificial Neural Network (ANN) to estimate the step length through the data from the IMU. The conducted experiments show that the proposed method achieved less than 2% error in a distance of 62.3m. Comparing to the traditional double integral method, it has superior performance and better ability to handle the signal noise from a low-cost IMU.

Description
Keywords
Neural network , IMU , Stride length estimation , Inertial navigation
Source
DOI
Publisher's version
Rights statement
Collections