294 Automatic Operator Guidance System in Open Pit Coal Mining to Increase Truck Productivity using Adaptive Machine Learning Ikhsan Maulana1*, Julian Eryanto2, Sukrisno3 1,2,3 Management Science Department Engineering Division, PT. Pamapersada Nusantara, Jakarta,13220, Indonesia *) Corresponding author: ikhsan.maulana@pamapersada.com ABSTRACT. Waste removal production shares a foremost aspect in open pit coal mining system. Maintaining its fleet productivity, which consists balancing of loader and trucks performance, has always been a difficult task to do. Fleet Management System is implemented to help monitoring fleet condition to get real-time profile of loader and hauler performance. Integrated data from real time GPS, Telematics data from vehicles sensors and other production related data are acquired to support the Fleet Management System. Fleet Management System has a capability to interpret the real situation that happened on the truck cabin. Productivity and safety in mining operations are strongly influenced by operator driving behavior. A key challenge is that traditional monitoring systems depend on fixed thresholds, which overlook the complexity of human behavior and the variability of site conditions. This results in limited insight and generic feedback that fail to drive lasting performance improvements. To address this, we present an adaptive learning approach that continuously analyzes telematics and onboard sensor data covering acceleration, braking, idle time, and gear use through machine learning models capable of incremental adaptation. Unlike static methods, this system evolves with each operator and context, generating personalized recommendations such as reducing idle time, refining shift strategies, or optimizing speed ranges. Field validation demonstrates a clear impact, with measurable improvements in fuel efficiency, cycle times, and equipment utilization. By using integrated delivery system data, the authors conducted research to utilize the data to understand operator behavior that impacted to the productivity result. The data is modeled to obtain the most optimal truck speed and cycle time of dump trucks in every pathway using modified adaptive machine learning theory and adapted to actual conditions in the field. This approach offers a practical pathway to enhance operator training, optimize productivity, and strengthen safety in large-scale mining operations. Finally, through this research, the author is able to analyze and classify the pattern of determining the guidance recommendation for each dump truck operator personally in real-time. The result was a significant increase truck productivity up to 6%.
RkJQdWJsaXNoZXIy MTM0Mzk2