ORIENTATION ESTIMATION USING KALMAN FILTERS WITH MULTIPLE SENSORS

File Type:
PDFItem Type:
Conference PaperDate:
2015Access:
openAccessCitation:
M.F. Cullinan, C. McGinn and K. Kelly, ORIENTATION ESTIMATION USING KALMAN FILTERS WITH MULTIPLE SENSORS, 32nd International Manufacturing Conference, Queens University Belfast, 2nd-4th September, Joe Butterfield & Paul Hermon, 2015, 29 - 38Download Item:
Abstract:
The ability to repeatedly and accurately position robotic end-effectors is a key requirement for modern industrial automation. Parallel developments in domestic and service robotics are opening up exciting possibilities for future manufacturing scenarios that would see robots and humans work far more closely together. These robots will typically be lighter, more mobile, less task specific and far more interactive than typical manufacturing robots of today. Reliable and precise detection of position and orientation in conjunction with the ability to plan and control complex kinematic motions will be vital. Currently, a large variety of sensors exist to measure orientation. The reliability and other attributes of these sensor technologies vary greatly. By using multiple sensors more accurate orientation estimates can be achieved. This paper presents the use of Kalman filters to realise sensor fusion of accelerometer, gyroscope and incremental encoder data. Four Kalman filters are implemented and tuned based on data from a system incorporating the relevant sensors. The filters use either gyroscope or encoder data for the predict step and accelerometer or encoder orientation estimates, or a combination of the two, for the measurement step. There are distinct differences between the responses of these implementations when the system is in operation.
Author's Homepage:
http://people.tcd.ie/kekellyhttp://people.tcd.ie/mcginnco
Description:
PUBLISHEDQueens University Belfast
Author: KELLY, KEVIN; MCGINN, CONOR
Other Titles:
32nd International Manufacturing ConferenceType of material:
Conference PaperAvailability:
Full text availableSubject (TCD):
Smart & Sustainable PlanetLicences: