Operational planning of road traffic for autonomous heavy-duty dump trucks in open pit mines

摘要

This article discusses the issues of path planning for movement of an autonomous heavy-duty dump truck along haul roads in an open pit mine. Taking into account the heterogeneity and dynamic nature of the technological environment, the issues of the efficiency and quality of the path planning of are extremely relevant. As a basis for calculations, a digital open pit model is considered, with technological zones and haul roads described as their geometric surfaces fragmented into same-type regular polygons. In this case, the workspace represents a set of discrete points with specified coordinates — the centers of geometric primitives, and the segments connecting them represent the possible directions of travel of an autonomous dump truck. The combinations of these segments, which are assigned certain weights, form a sequence of reference points of a potential route for the robots to move along. Thus, the planning objective in this study is formulated as the search for the optimal value of a certain cost function that takes into account condition of a roadway. The procedure for constructing a graph which acts as a basic model for determining the best route for a mobile robot is briefly discussed. As an algorithm for finding the optimal sequence of points, it is proposed to use a modification of reliable Dijkstra’s algorithm, with the running speed being increased owing to the original implementation of parallel computing. In terms of the real-life fragments of haul roads, using the predefined terminal points and the constructed graphs of high dimensionality, the edge weights of which determine the values of the cost function, various routes for dump trucks are constructed and studied. The modeling results confirmed the efficiency of the proposed approach while limiting the use of standard computing resources.