Predicted safety & excavation progress algorithms for autonomous excavation

Abstract

Excavation is the process of moving earth and is considered to be one of the primary projects in the construction industry. The adoption of various cutting-edge technologies for full automation can be a solution to the various ongoing issues in construction equipment used for excavation such as safety, monitoring, and productivity. To address this, this project developed advanced safety algorithms and methodologies in ground mapping and estimation of excavation progress, which can accelerate autonomous excavation. For autonomous excavation, safety is a significant concern to reduce accidents and machinery damage. Considering this point, this thesis deals with tracking, motion prediction, and track management of the detected objects that can improve the safety function of autonomous excavators. The proposed safety algorithms can evaluate the degree of a potential collision risk by using the information of predicted motion of detected objects, working areas of the excavator, and safety indices calculation. The second component of this project covers the volume estimation for excavation progress estimation, occlusion problem for ground mapping, and 5D mapping. The volume estimation comprises of ground excavation volume and bucket volume estimation. To overcome the problem of an occlusion area that may result in incorrect mapping and estimation of excavation progress, sensing data of proprioceptive and exteroceptive sensors were integrated. Finally, we proposed the idea of 5D mapping to provide a broad spectrum of the excavated ground info that includes the coordinates and material type and properties using a 3D ground map, LiDAR’s beam reflectivity, and force index.