1543 This end-to-end cycle definition ensures that the performance evaluation reflects the actual operational efficiency, including travel time, dumping duration, return movement, and any delays occurring within the cycle. By standardizing the start and end points at the same front location, the system provides a consistent and comparable performance benchmark for operators and supervisors, supporting more accurate monitoring and continuous improvement of hauling productivity. 3. RESULTS AND DISCUSSION The results presented in Table 2 summarize the performance of the proposed adaptive machine learning system across three complementary dimensions: predictive accuracy, operational impact, and productivity outcomes. Evaluation was conducted using a combination of simulated scenarios and real-world telemetry data to ensure both methodological rigor and practical relevance. Predictive accuracy metrics quantify how closely the model’s outputs align with observed vehicle behavior under real operating conditions. Operational impact metrics capture changes in driving behavior, energy efficiency, and mechanical stress following the introduction of modeldriven guidance. Productivity outcomes reflect system-level performance improvements, measured in terms of effective throughput per shift rather than isolated vehicle-level metrics. Table 2. Model Performance, Operational Impact, and Productivity Outcomes Evaluation Aspect Metric Observed Performance Operational Significance Speed Recommendation Accuracy RMSE ±2.1 km/h Enables precise and safe real-time speed guidance Cycle-Time Prediction Accuracy MAE ±5.4 s per segment Supports reliable shortterm planning and forecasting Cycle-Time Consistency Variability Reduction 5–12% Improves predictability and operational stability Fuel Consumption Reduction 3–7% Lowers operating cost and environmental impact Braking Events Frequency Reduced Decreases component wear and safety risk Operator Behavior Variability More uniform across shifts Reduces dependency on individual driving skill Productivity Throughput per shift ≈ 6% increase Improves asset utilization and production output
RkJQdWJsaXNoZXIy MTM0Mzk2