The engineering company OES implemented an AI-based system to optimize the operation of over 240 mining vehicles, including BelAZ 75131 and 75306 dump trucks and more than 25 excavators. The solution operates as an overlay on existing monitoring infrastructure, utilizing data from Galileosky 7x and 10 terminals installed on the fleet.
The AI platform features over 40 specialized models, including a neural network for estimated time of arrival (ETA) and a balancing algorithm that recalculates vehicle distribution every 3 minutes. A swarm of 30+ AI agents monitors parameters such as fuel levels, route speeds, and loading quality, issuing directives directly to driver tablets.
The system addresses operational inefficiencies by managing refueling schedules and shift changes, which previously suffered from up to 25% early departures or late arrivals. It also predicts queue lengths 30 minutes in advance and automates the escalation of deviations across three levels: driver, dispatcher, and shift supervisor.
Performance results include a 10% increase in total rock mass extraction and a 20% reduction in excavator idle time. Dispatcher workload decreased by 60–70%, while 95% of AI-generated directives are executed without human intervention. The project generated over 20 million rubles in economic impact during its first month.
