Vibration Analysis for Speed Estimation of Wheeled Indoor Vehicles

Sammanfattning: In this thesis speed estimation of wheeled indoor vehicles based on vibrationsin the wheels is studied. Suitable applications include localization on wheeledindoor equipment, where wheel speed sensors are impractical or cannot be installed.Utilizing the inertial measurement unit installed on a tag, the accelerometerscan provide information about the fundamental frequency and harmonicsof the wheel. The information is provided through the vibration components,embedded in the accelerometer signal, and the speed of the vehicle is proportionalto the fundamental frequency. To enable inertial navigation of the vehiclea Kalman filter is used, allowing estimations to be performed in the time domain.In addition to tracking the frequency of the wheels, the accelerometers are usedto track fast speed variations and detect stand still. Compared to inertial navigationwhich has a cubic positioning drift over time, this setup limits the drift to belinear in time. The proposed inertial navigation system is evaluated on datasetsrecorded in a proof of concept environment, performing continuous frequencytracking with a low average error rate.