Identifying Operator Usage of Wheel Loaders Utilizing Pattern Recognition Techniques
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
The information about how wheel loaders are used by costumers is not well documented, but large quantities of sensor signal data are recorded. This data makes the starting point to develop a robust automatic pattern recognition algorithm to identify repetitive driving operations.
The algorithm is composed by models of events that appear in cycles. The cycles are pre-defined as short loading cycle and load and carry. They consist of the phases; load, drive loaded, unload and drive unloaded. In the load and carry cycle, the driving forward loaded phase can last up to 400 meters. Beyond these two is something called sub cycle defined. It contains cleaning cycle and reloading. A cycle is identified by transition automata of events. To make the identification more robust against disturbances, a probability function is connected to the cycle identification.
The developed algorithm uses signals from sensors available in series production wheel loaders. These signals are used to identify cyclic behaviour in the shape of short loading cycle, load and carry and sub cycle. After removal of completely standing still events, the algorithm calculates the characteristic parameters for the identified cycles and for their phases and present them in an excel file. Validation showed that the algorithm can find 75 % or more of cyclic behaviour for bucket handling.
Place, publisher, year, edition, pages
2012. , 54 p.
Other Electrical Engineering, Electronic Engineering, Information Engineering
IdentifiersURN: urn:nbn:se:liu:diva-78937ISRN: LiTH-ISY-EX--12/4591--SEOAI: oai:DiVA.org:liu-78937DiVA: diva2:537110
Subject / course
Systemet, Linköping (Swedish)
Frisk, Erik, lektor