Can new technologies like bus automation support a shift to more sustainable mobility? How in practical terms could automation make buses more attractive? These were the key questions which motivated the Centre for Transport & Society’s (CTS) involvement in the five-year MultiCAV research and development project, co-funded by Innovate UK and the Government’s Centre for Connected and Autonomous Vehicles.
As well as operating autonomous electric bus services, MultiCAV, based at Didcot (Oxfordshire, UK) introduced electric bike sharing and an enhanced travel information system. The project was led by bus operator First Bus, supported by Nova Modus. Milton Park Technology and Science Park and Oxfordshire County Council hosted and facilitated the project’s services. Fusion Processing provided automation technology, and Zipabout, digital travel information integration. CTS, based at the University of the West of England in Bristol, was responsible for evaluative research to learn lessons from the application of the technologies and draw out messages for the future.
The research by CTS engaged users of the MultiCAV services, branded ‘Mi-Link’, as well as people working at Milton Park, who were seen as potential users. They took part in surveys and interviews. The ‘safety operators’, professional bus drivers who oversaw the automated buses and drove them when necessary, gave feedback on their experiences, and partners in the project also reflected on their ‘learning journeys’.
The two demonstration autonomous bus services ran in the Spring and Summer of 2023. As far as possible they were designed to feel like typical bus services, with route numbers, bus stops with printed timetables, and information also available online and in real-time.
A few things made the services a bit unusual: they were operated by electric minibuses, still quite rare in the UK at the time, and passengers needed to be seated, wearing seatbelts. Fares were not charged, but most people travelling on the chosen routes (first within Milton Park, and then later between Milton Park and Didcot railway station) already had prepaid bus passes anyway.
Interior of the Bus
View of bus interior from back seat looking towards front
Most Mi-Link passengers were very regular and frequent bus users. For the most part they found the services to be much like other bus services they travelled on. They found them to be quieter, probably because of the use of electric power. Our report on user experiences examines these and other views in more detail.
An indicator panel was put in the bus so that passengers could see when it was being driven by the autonomous system and when the human safety operator driving. For most typical manoeuvres during the journey users thought the autonomous system was driving as well or better than a human. But, even so, for now at least, passengers still wanted there to be a human presence on the bus, to cope with the unexpected. Our research also resulted in a report considering the views of people working at Milton Park, who also backed up this point: they were positive towards using automated buses provided they would be staffed.
Autonomous Mode Indicator
Sign illuminated with blue LEDs to read * Auto *
The experiences during the project of the safety operators were critical to record, and are covered in a third report. They had positive opinions about the training they received and their experiences ‘co-piloting’ the bus. Their confidence in the technology grew through the months of the demonstration services. One thing we had not expected was that the operators would work in partnership with the automated system, in order to give the best service to the passengers. They recognised when the bus was driving more effectively than a human driver would be able to, but also knew when it was the right moment to take control in order to avoid over-cautious braking, or to negotiate a junction where risk-taking by human drivers is the norm.
The last strand to our evaluation was to ask our MultiCAV project partners to share their experiences; the topic of a fourth report. On reflecting upon their experiences across the project, the partners described a technology which was maturing in capability. At the beginning of the project in 2018 the team had struggled to secure a sufficiently capable autonomous vehicle. Events including COVID-19 then slowed the rate of progress. There were ‘teething troubles’ to overcome such as cutting back roadside vegetation that could affect the autonomous driving system. But, in 2023, partners were nonetheless surprised and reassured to experience the bus running automated at 40 mph in normal traffic.
So where next for autonomous buses? In the near future, the project has identified passenger benefits from autonomous driving technology working in partnership with human drivers. The findings support the case for buses to be equipped with greater driver-assistance technologies, similar to the way that smaller vehicles like cars are evolving.
But a large benefit from lower operating costs is only possible if buses do not need to be staffed. The lower operating costs could mean additional bus services can be provided, or fares reduced. But this would only be possible if the need for intervention by safety operators is removed. The experiences in MultiCAV confirm the ecosystem around automated buses would need to evolve before unstaffed operation is possible. There are different views about how far roads will need to be standardised to allow automated vehicles to operate on them; some think at least some changes will be necessary, others argue that the evolution of autonomous driving systems will make changes unnecessary. Depending on how far the technology develops, roads may need additional maintenance and different ways of managing things like temporary road works. It is clear though that there will have to be a reliable solution for travellers with additional needs, such as wheelchair users, who are currently assisted by the bus driver. The confidence of the public in travelling on unstaffed buses would need to grow. They might receive remote support by telecommunications as they do on some trains.
In the meantime, fully autonomous buses may be able to operate first where it is possible to provide bus-only roads, perhaps in new residential or commercial developments where the land can be set aside. Special-needs passenger assistance would be bookable, as is the case on the railways. If it is not possible to provide a bus-only road for the whole route, then the service may have an automated section and a human-driven section, so that operating cost savings are made at least for part of the route.
This blog was written by Prof Graham Parkhurst, Professor of Sustainable Mobility and Director of CTS, University of the West of England, Bristol. Graham led the CTS research in the MultiCAV project between November 2018 and January 2024. For any questions or comments about this blog or the project please email graham.parkhurst@uwe.ac.uk.