Two immediate pressures are reshaping public transport operations right now: electrification and the driver shortage. Both converge in the same place, the depot.
When I refer to BEBs, I mean battery-electric buses. Compared with diesel fleets, BEB operations introduce much stricter range constraints and far more variability in energy consumption. Ambient temperature is one of the biggest drivers of this variability, particularly through its impact on heating and cooling demand, together with factors such as passenger load, driver behaviour, and average speed.
Because of this uncertainty, many BEB systems end up over-dimensioned. In practice, operators often rely on conservative safety margins, which can increase infrastructure costs and lead to decisions that are robust but not necessarily efficient - such as selecting the largest available battery capacity or charging at full power whenever an opportunity arises.
Most BEB deployments already provide rich telematics data, including real-time battery state-of-charge information. The gap I see is that operational planning still rarely incorporates real-time energy prediction or predictive insights. As a result, decisions are often made with limited forecasting, and the system “protects itself” through additional buffers.
This is where digitalisation and data-driven tools start to matter. In our work, we explore AI-based energy prediction models trained on operational data. The goal is simple: vehicle- and route-specific energy forecasting that supports smarter charging strategies, balancing three things at once - operational reliability, battery lifetime, and energy costs.
Charging optimisation in real depots is a multi-objective problem. Operators need to avoid excessive full charge cycles and deep discharges, account for uncertain arrival times that shrink charging windows, and consider the effects of charging power on battery ageing. Another dimension is growing in importance: electricity pricing that varies over time. When tariffs become dynamic, the decision is no longer just “charge whenever you can” or even “charge during the cheapest hours”. Operators need strategies that jointly optimise energy cost, degradation-related cost, and operational robustness.
In that context, telematics needs to evolve from a monitoring tool into part of a real-time coordination layer: precise vehicle localisation, infrastructure status, predictive readiness, and decision logic that helps operators run the fleet with less oversizing.
In parallel, persistent driver shortages keep autonomy on operators’ agendas. Fully autonomous, high-capacity buses operating in mixed traffic remain a longer-term objective, often placed beyond 2032 in industry roadmaps. The near-term question is where autonomy becomes practical, safe, and scalable without compromising service reliability.
My view is that depot automation offers the most feasible early deployment domain. Depots are controlled environments: low speed, passenger-free, known layouts, and highly repetitive manoeuvres. These manoeuvres - parking, washing, charging moves, dispatch staging - consume significant driver time without generating revenue.
Automating them can increase depot throughput, reduce internal congestion, improve vehicle readiness for scheduled service, and optimise labour utilisation. Importantly, the benefits can be quantified with operator-specific data, which supports robust ROI assessment rather than relying on assumptions.
Depot automation only delivers at scale when it behaves like a system, not a standalone feature. That means integration across several layers:
This is also where I see telematics evolving again - from monitoring to operational coordination. When energy forecasting, charging optimisation, and depot movements share one operational picture, operators reduce manual workload and remove avoidable friction from daily depot processes.
A responsible rollout is staged. Operators typically start with a pilot vehicle, then a limited operational deployment, and only then move to larger-scale fleet integration once procedures and system performance stabilise. That sequence builds organisational readiness (training, supervision, override and recovery procedures) while generating operational evidence that operators and stakeholders can rely on.
Electrification demands better data and decision support. Driver shortage demands productivity and resilience. Depot automation connects both - delivering operational benefits now while building the maturity required for higher levels of autonomy later.