Reducing the effect of magnetic disturbance in inertial measurement based indoor localization

Using inertial measurement units (IMUs) has been one of the main techniques for indoor localization. In many cases, this type of methods uses accelerometer, gyroscope, and/or magnetometer that are available in commercial mobile devices, such as smart phone. One main advantage of using this type of techniques is that no separate infrastructure deployment is needed. However, its performance is affected by noise that exists in the measurement data. When tracking pedestrians, the moving speed and direction are usually estimated separately. Muldi Vision Ltd. (MDV), the industrial partner organization, has been providing localization services for workers in industrial environments. When using their current product in manufacturing plants, the accuracy of the IMU-based localization method is highly affected by magnetic disturbance. The objective of this work is to find methods to reduce the negative effect of magnetic disturbance on the accuracy of the sensor orientation estimate and therefore improve the localization accuracy.

Faculty Supervisor:

Dongmei Zhao

Student:

YiQiong Miao

Partner:

Muldi Vision Ltd.

Discipline:

Engineering - computer / electrical

Sector:

Professional, scientific and technical services

University:

McMaster University