Operational efficiency #4: asset maintenance Eco-driving is one of the key advantages of autonomous trucking. Autonomous vehicles can be programmed to perform best practices in driving behaviour. Studies show eco-driving alone can achieve a 4-10% reduction in fuel consumption. Garrett Bray, former Product Director, Aurora Innovation; alumnus of Centre for Sustainable Road Freight, University of Cambridge 1.4 Regular and thorough asset maintenance plays a role in reducing emissions and prolonging asset lifespans. For example, in the road freight segment, a properly maintained engine ensures optimal combustion, while under-inflated tyres or poor alignment increase rolling resistance, requiring more energy to maintain speed and thus consuming more fuel. AI is enhancing asset maintenance through predictive maintenance solutions that monitor asset health, forecast potential failures, optimize maintenance schedules and monitor and maintain battery health. AI technologies can analyse vast amounts of historical and real-time data to identify patterns that humans could miss, such as engine wear, tyre degradation or brake performance, which could lead to costly repairs or inefficient power use if left unchecked. In EVs, AI-powered solutions can integrate many complex factors and predict battery lifespan with up to 95% accuracy.18 A battery management system (BMS) can collect data from temperature, voltage and charge/discharge cycles to predict battery degradation and optimize charging strategies. Several EV manufacturers are already applying this technology and recommending optimal charging and driving behaviour to prevent wear and tear. This is especially critical for electric trucking fleets, which demand high reliability to maximize uptime and enhance margins. For instance, in a tough commercial environment such as over the past two years, where average operating margins for US trucking (excluding less-than-truckload/ LTL) were below 6% in 2023, optimizing fleet performance becomes imperative.19 Predictive maintenance has also gained traction with the rise of AI. Such proactive maintenance in the rail sector costs around seven times less than emergency repairs done after infrastructure fails, so AI-driven optimization could help deliver emission reductions and further operational savings (see Box 2).20 For example, some rail operators use predictive maintenance platforms to prevent train delays by detecting early signs of wear and tear on switches and gears, which are common causes of rail disruptions. BOX 2: Hitachi Rail collaborates with NVIDIA to drive efficiencies BOX 2 Hitachi Rail is collaborating with NVIDIA to improve rail operations through AI solutions. This partnership aims to reduce maintenance costs, minimize idle times and enhance train scheduling and reliability. Building on Hitachi’s existing applications, which analyse data from 8,000 train cars across 2,000 trains, these tools provide computational ability to provide real-time insights into monitoring train fleets and infrastructure more effectively. Previously, such analysis took days to deliver results. While the emission-saving potential in the rail sector is relatively low given the already low emissions associated with rail transport, the increase in reliability is a crucial benefit. A more reliable rail network could encourage a switch from road to rail, a critical step in reducing overall emissions. This modal shift, while an indirect outcome of predictive maintenance, is a piece of the puzzle in lowering global freight logistics emissions. While emission- saving potential in the rail sector is relatively low, given rail transport’s already low emissions, a more reliable rail network could encourage a switch from road to rail, a critical step in reducing overall emissions. Intelligent Transport, Greener Future: AI as a Catalyst to Decarbonize Global Logistics 13

Intelligent Transport Greener Future 2025