Currently 25% of all carbon emissions in the EU are generated by road traffic[1]. Given that 75% of all goods are delivered by road, it is unsurprising that commercial vehicles account for a significant proportion – one third – of those emissions.
Industry leaders and those in government have all committed to reduce emissions and to eventually decarbonise transport entirely, complete with ambitious deadlines. The UK government’s commitment to phase out all non-zero emission cars, vans and HGVs of 26 tonnes and under sold here by 2035, and all HGVs by 2040[2], is reinforced by the zero-emission vehicle (ZEV) mandate. For many in the transport sector however, timelines are even shorter, most notably members of the EV100 climate group[3], who are committed to converting their fleets to electric vehicles by 2030.
But how realistic are those deadlines? And what factors need to be considered if they are to be met? In many cases, significant action is required now to even stand a chance of success.
The UK Government is working closely with the tech sector to speed up innovation and, it hopes, to clean up the transport sector while making it more efficient. Recently AI has been a particular focus due to its unprecedented ability to drive operational efficiency using big data. Given the already data-rich nature of commercial fleet operations it is the obvious place to start.
AI in freight logistics
For many businesses and fleet operators, the transition to E-vehicle fleets is central to their net-zero goals. However, there are currently certain barriers to making this transition completely, such as lack of charging infrastructure to support longer-haul routes and the huge energy requirements of HGVs.
How and where AI can help bridge the gap between these issues and the pace of technology innovation, is an area that we are currently focussed on as part of the DESNZ (Department for Energy, Security and Net Zero) AI for Decarbonisation programme. FPS already uses big data and machine learning to deliver operational efficiency and resilience for fleets but is now developing a proof-of-concept on how AI can help accelerate the decarbonisation of the transport sector.
In the shorter-term, AI can bring meaningful benefits via deep learning, deep reinforcement learning, and large language models. Take the issue of e-vehicle charging. By using AI to enable ‘smart charging’ coordination of vehicle charging can be optimised to help negate challenges like grid constraints and varying electricity pricing in more complex or large-scale depots.
Similarly, AI can support Integrated energy management. As a natural extension of the smart charging use case, it can be used to optimally manage the energy flow between various producers and consumers of energy.
The technology can also be used to train systems in simulation to discover the most energy-efficient driving strategies, which can then be applied in real-world scenarios to reduce fuel consumption in conventional and hybrid vehicles.
In the longer-term AI also has the potential to relieve some of the uncertainty associated with shared charging infrastructure, specifically when vehicle downtime needs to be minimised. It could achieve this by enabling better demand forecasting for example, to ease congestion at shared hubs by using real time data to alert stakeholders. It could also minimise the prospect of breakdowns at crucial moments (which can wreak havoc), by improved predictive maintenance.
Predictive management and maintenance
Deep Learning AI models can also predict traffic conditions and optimise routes in real-time, reducing idle times and fuel consumption. By analysing patterns from vast amounts of data, and considering variables like weather, traffic, and vehicle performance to minimise fuel usage and emissions, these models help in creating more efficient logistic operations.
Just as traffic flows can be better predicted using AI, the same applies to ‘predictive maintenance’ – predicting when vehicles and machinery will need maintenance. This helps prevent inefficient operation that increases emissions, not only extending the lifespan of the equipment but also ensuring they operate at peak efficiency.
In the case of electric vehicles specifically, charger maintenance is central to this operational efficiency and improvements in, for example, remote charge point monitoring and predictive maintenance will be key.
Supply chain optimisation
Beyond fleet management and maintenance, Large Language and Deep Learning Models can also be used to analyse complex supply chain data, optimising for variables such as cost, time, and carbon footprint. AI can forecast demand more accurately, thus reducing overproduction and excessive transport and helping businesses meet Scope 3 requirements as well as their Scope 1 and 2 obligations.
These models also have a role to play in carbon emission tracking and reduction and can be used to process and analyse regulatory documents, scientific papers, and data from transport and logistics operations. This will help ensure compliance with environmental standards and to identify opportunities for carbon reduction.
Managing the transition
Taken as a whole, the immediate benefit offered by AI and related technologies is their ability to drive efficiencies and limit waste, while laying the foundations for a greener and more cost-effective transport sector in the future. Diesel and other fossil fuels cannot simply be replaced overnight and until the technology and infrastructure for complete e-vehicle fleets exists, innovative ways to minimise carbon emissions while using a mix of fossil fuels, biofuels, hydrogen, and electric power will be essential to the transition.
AI’s ability to optimise efficiency at almost every point in the supply chain will help businesses better predict and control their CAPEX while allowing them to plan ahead more effectively on exactly how they will decarbonise their fleets. This could include, for example, something as simple as knowing which routes can realistically be decarbonised now and which will need to be put on hold until the technology and infrastructure is more advanced.
Funding technology and innovation will only take us so far, however. AI has the potential to vastly improve efficiency but only with access to sufficient quality data. In order to make the necessary breakthroughs to decarbonise the sector in line with the proposed phase-out dates, policymakers will need to work closely with the industry to improve public data infrastructure and ease regulatory environments to free up Big Data sets, while obviously taking proper account of data security and privacy issues.
In practice, this will require governments, both in the UK and elsewhere, to take a more holistic view of transport regulation and infrastructure and be open to sharing cross-sector data and best practice to avoid industry working in cost-inefficient silos. In doing so they can help speed up the process of innovation and bring us one step closer to a net-zero future.
[1] https://www.eea.europa.eu/en/topics/in-depth/road-transport
[2] https://www.gov.uk/government/news/government-sets-out-path-to-zero-emission-vehicles-by-2035
