Excavator Robotization And Experimental Study

Excavators, called the king of construction machinery, is one of the mainstream products of construction machinery. It has a key role in the industrial or civil construction, local communication, water conservancy and hydropower, military construction and so on. However, the working environment of excavators is poor. For example, noisy and dusty. The operator has to concentrate on his/her operation to prevent accidents as a result of absence. Sometimes, the operator faces hazardous and unknown environments, such as earthquake and debris flow, to dredge river and clean road. These conditions pose a threat to the operator’s life. Therefore, to remove operators from hazardous environment is people’s wish and requirement, which is also the motivation of the dissertation.The excavator’s task planning is researched in this thesis. Behavior-based control which makes easy to decompose complex tasks into action sequences that can be composed of basic actions has been used in the excavator’s control architecture. The control architecture has two levels. The higher-lever is trajectory planning which is implemented with FSM and the lower-lever is using fuzzy logic control. Skilled operators’ actions have been observed, which are expressed with FSM. The transitions of statechart are triggered by sensors’ angle and pressure values. According to the cylinder’s pressure, the fuzzy logic controller gives the amplifier proper voltage to change the value opening area to adapt to soft or hard material.Digging sand model and rock encountered model are designed in this thesis. When rock encountered, the excavator tries to dig from the bottom of the rock by adjusting the motions of the bucket. The digging cycle repeats three times.The experimental platform is Hardware-In-the-Loop platform which based on MATLAB/xPC Target. Experimental prototype is the Komatsu PC02-1excavator after rebuild. The role of fuzzy logic controllers are highlighted by compared digging in the soil and gravel. Rock encountered experiment is also performed. Experiments show that the excavator could execute according to the planned action sequences and adjust the bucket three times when rock encountered trying to dig the rock out from the bottom of it.