Motion Design for Soil Excavation by Wheel Loaders Using Bayesian Optimization
Hiroaki Muranaka, Motoki Koyama, Masato Ishikawa
pp. 91-98
DOI:
10.5687/iscie.37.91Abstract
A wheel loader is a construction machine that consists of a four-wheel driving mechanism and a two-degree-of-freedom working mechanism of boom and bucket. This is mainly used to excavate sediment and crushed stones and load them into a dump truck. Currently, at civil engineering and quarry sites where construction machines are used, problems such as manpower shortages and worker hazards have made automation and remote operation of construction equipment desirable. Against this background, the objective of this study is to design an efficient excavation motion based on mathematical optimization, aiming at the automation of excavation by construction machinery. In particular, we used a Bayesian optimization approach that alternates modeling and optimization of experimental and input-output data. Based on the idea that it may be possible to tolerate deviation from the designed target trajectory and actively utilize “weak” tracking, we propose a flexible excavate operation strategy. Based on the above policy, the excavation motion design problem for a wheel loader is formulated as follows: the design variables are the parameters of the parametrized trajectory and the tracking accuracy of each drive element with respect to the reference trajectory, and the objective function is a linear combination of excavation volume and workload with a certain weight. Then, we investigated the effectiveness of this method by comparing it with track following control using an excavation simulation. As a result, we found that the proposed method can improve the excavation volume and workload depending on the conditions.