A ROBOTIC TESTBED FOR AUTONOMOUS DUMP POCKET CLEANING USING IMITATION LEARNING *Brik Henrry Meza Pinedo1,2, Brian Pajares Correa1 1Faculty of Science and Engineering, Pontifical Catholic University of Peru (PUCP), Lima, Peru 2NONHUMAN, Lima, Peru (*Presenting author: brik.meza@pucp.edu.pe) ABSTRACT Dump pocket blockages at primary crushers cause production downtime and require manual clearing in hazardous restricted areas, with documented engulfment fatalities in hoppers and crusher zones. Autonomous approaches for dump pocket cleaning and blocked- crusher clearing remain limited, and the feasibility of state-of-the-art imitation learning (IL) for this excavation task remains largely unexplored. This paper introduces an experimental testbed and benchmark for evaluating IL architectures on autonomous dump pocket cleaning under controlled laboratory conditions. We benchmark four IL architectures: Action Chunking with Transformers (ACT), Diffusion Policy, and Vision-Language-Action models (π0.5 and SmolVLA), using a lowcost SO-ARM100 platform ($250) with granular bentonite material. In a preliminary singlesession evaluation (10 minutes per model), π0.5 achieves the highest removal rate at 404 g, reaching 65% of the expert teleoperation estimate, followed by ACT (169 g), SmolVLA (113 g), and Diffusion Policy (57 g). π0.5’s lead is consistent with potential advantages from the larger Hugging Face checkpoint configuration evaluated in this study (3.7B vs. 450M parameters) and broader pretraining across diverse robot embodiments and tasks. ACT, trained from scratch on the same dataset, outperforms SmolVLA in this single-session benchmark, raising the hypothesis that for narrow single- task settings where the target domain differs from pretraining data, learning from scratch may remain competitive with fine-tuning a pretrained model. Diffusion Policy removes the least material, consistent with its visibly slower reactive behavior during rollout. This work establishes an open dataset (162 demonstrations) and a reproducible testbed to support future research on autonomous tasks for the mining industry. Project page: https://brikhmp18.github.io/dump-pocket-il KEYWORDS Autonomous Excavator, Dump Pocket Cleaning, Imitation Learning, Autonomous Mining, VLA Models, Operational Safety, Robot Learning. 1. CONTEXT AND PROBLEM STATEMENT Mining is a major economic driver in Peru, contributing around 10% of national output and approximately two-thirds of export value [1]. In large-scale operations, the
RkJQdWJsaXNoZXIy MTM0Mzk2