UCSD researchers slash electric construction costs 7-96%
AI-powered scheduling cuts electric excavator charging costs by 7-96% with mobile stations
Researchers are tackling construction electrification with a data-driven framework. Using a real-world demonstration of a compact electric excavator at UC San Diego, they built a model that estimates per-subactivity power consumption from labeled video and battery state-of-charge telematics, achieving a 17% normalized mean absolute error on held-out test data. They then developed a mixed-integer optimization framework that jointly schedules construction electric vehicle work, charging, and mobile charging station logistics, while accounting for energy costs, demand charges, emissions, travel, and operational constraints. Across realistic scenarios, it delivered the lowest operating
- AI model predicts electric excavator power use within 17% error using video and battery telemetry
- Optimization system cuts operating costs 7-96% by coordinating work scheduling with mobile charging stations
- UCSD released open datasets and code to accelerate industry adoption of electric construction equipment
Why It Matters
Could slash $10B+ in annual construction fuel costs while cutting 30M tons of CO2 from diesel equipment