FusionPlanner: A Multi-task Motion Planner for Mining Trucks via Multi-sensor Fusion

TL;DR

This work targets autonomous vehicle motion planning in open-pit mining, a highly unstructured environment. It introduces FusionPlanner, a multi-sensor, end-to-end planner for mining trucks that fuses LiDAR and GNSS and jointly learns lateral and longitudinal control with task-aware uncertainty and evidential reasoning. It also presents MiningNav, the first mining-specific benchmark with three validation tasks, and the Parallel Mine Simulator (PMS), a high-fidelity platform for open-pit scenarios. Experiments in PMS show FusionPlanner reduces collisions and interventions, advancing trustworthiness and robustness for continuous unmanned mining operations.

Abstract

In recent years, significant achievements have been made in motion planning for intelligent vehicles. However, as a typical unstructured environment, open-pit mining attracts limited attention due to its complex operational conditions and adverse environmental factors. A comprehensive paradigm for unmanned transportation in open-pit mines is proposed in this research. Firstly, we propose a multi-task motion planning algorithm, called FusionPlanner, for autonomous mining trucks by the multi-sensor fusion method to adapt both lateral and longitudinal control tasks for unmanned transportation. Then, we develop a novel benchmark called MiningNav, which offers three validation approaches to evaluate the trustworthiness and robustness of well-trained algorithms in transportation roads of open-pit mines. Finally, we introduce the Parallel Mining Simulator (PMS), a new high-fidelity simulator specifically designed for open-pit mining scenarios. PMS enables the users to manage and control open-pit mine transportation from both the single-truck control and multi-truck scheduling perspectives. The performance of FusionPlanner is tested by MiningNav in PMS, and the empirical results demonstrate a significant reduction in the number of collisions and takeovers of our planner. We anticipate our unmanned transportation paradigm will bring mining trucks one step closer to trustworthiness and robustness in continuous round-the-clock unmanned transportation.

Figures (10)

• Figure 1: The difference between open-pit mining scenarios and urban traffic scenarios.
• Figure 2: Overview of FusionPlanner. The raw LiDAR point cloud (with visual colorization based on coordination and intensity) and GNSS data are propagated through a fully connected layer in the control module of FusionPlanner after feature extraction.
• Figure 3: The output control commands of FusionPlanner can be implemented in three modes. (A) Instantaneous model: executing all commands instantaneously; (B) Uniform fusion model: uniformly fuse the predictions from multiple past frames; (C) Evidential fusion model: incorporating uncertainty estimation in the output, intelligently weighting the predicted commands to enhance the trustworthiness and robustness, particularly in OOD events or on increased uncertainty of future time steps.
• Figure 4: Part of scenarios of the open-pit mine in PMS.
• Figure 5: The loading site in PMS, shown from the third-person view in three common weather conditions of open-pit mines.