Optimization of grading path planning for an autonomous construction machine

The construction industry, which faces an aging workforce and shortages of skilled labor, presents attractive opportunities for autonomous heavy construction machinery. To achieve this goal, many difficulties must be overcome. Most of the problems are nonlinear and cannot be solved through convex optimization. Bulldozer operation presents especially difficult challenges, as it is a skilled trade and obtaining an operational method for bulldozers analytically is difficult. In this study, we optimized the route planning of a simulated bulldozer by using a genetic algorithm. To properly evaluate the solution candidates, we developed a simulator that emulates the dynamics of a sand mound. This paper describes the developed simulator and discusses the optimization of bulldozer paths on construction sites.