An autonomous excavator for landscaping tasks

In the present day the use of big modern construction machines has considerably improved the building process and therefore increased the output. Autonomous machines could define the next milestone in this area, as such machines can prevent human beings from working in hazardous environments, improve many of the occurring processes and additionally lead to higher earnings for construction companies. This paper deals with problems in the field of developing an autonomous bucket excavator, which is equipped with control and perception hardware, and a complex software system. As the safety of the implemented algorithms has to be proved, a simulation environment is used for testing and adjusting the parameters. A realistic particle simulation of soil and a physics engine deliver this close approximation to reality and make soil excavation and environment manipulation possible. A novel behavior-based approach for motion control is presented which allows natural boom trajectories, achieves good environment disturbance compensation behavior, guarantees safety of the movements concerning human beings and structures, and keeps a high extensibility of the system for future improvements.