An avatar behind the wheel
Making driverless buses a reality. It’s the logical next step, according to Dave Roat
Self-driving cars are the future of transportation. The UK has grand plans for autonomous vehicles and the government is striving to be at the forefront of the industry, estimating its worth at £28 billion by 2035. Most manufacturers are developing their own versions of autonomous cars, with both Honda and Toyota claiming to have self-driving vehicles on the road by 2020.
While conversations are focused on the driving aspect of these, such as road safety, the landscape is more complicated for buses. The industry must begin considering how to fast track these vehicles, so that people can take advantage of autonomous vehicles for public, as well as private transportation.
The need for a figure of authority
Bus drivers do a lot more than drive, they help with assisting vulnerable passengers, enforce how priority seating is regulated, recognise abnormal or sinister behaviour, and ensure passengers feel safe. They play two distinct roles, the driver and the conductor.
Many of these ‘conductor’ tasks cannot yet be automated fully. So, it is assumed that any early adoption of autonomous bus services by a transit operator will require staff to be on board the vehicles, even if they are not driving them. This seriously undermines the business case associated with the deployment of an autonomous vehicle. Some trials are getting around this by ensuring that staff are at the bus stops, or the trials themselves are so controlled that the risk of fraud or aggression are close to zero, which does not mirror a real-life situation.
Intelligent systems to police ticketing
Live Virtual Agent technology successfully tested in railway stations in the UK, Germany and Singapore use live video links and intelligent cameras to provide a human interface for rail passengers without needing to deploy staff on the stations.
The machine allows passengers who want help choosing which ticket to buy to talk to an expert member of staff, exactly as they would if they queued up at a traditional ticket window. The technology uses life size, eye-to-eye, full HD video and high-quality audio, providing a real time video link to the staff who would normally have sat behind the window in a station. These staff, however, are located off-site in a call centre and can handle incoming calls from multiple devices out in the field.
Although this is a slightly different environment, the technology can be used on buses in a similar way. With a live video to a remote ‘driver’ available, the autonomous bus could initiate a link whenever the vehicle arrives at a bus stop. The remote ‘driver’ would then take full control of opening and closing the door, based on what they can see from the integrated camera’s point of view. Once the bus is free to pull away the remote member of staff can then take control of a different bus.
This member of staff can also control the electronic deployment and recovery of a ramp for wheelchair users, while recognising vulnerable passengers and giving staff the ability to remotely override the ticket inspection kit, enabling vulnerable passengers to travel without a ticket.
The intelligent camera systems add further benefits, analysing and detecting human behaviour. Deployed on an autonomous bus, intelligent cameras can provide an added security component, ensuring passengers can enjoy a safe journey and staff get automatically alerted in case of misuse or other issues.
An autonomous future
In the future, artificial intelligence (AI) avatars could learn patterns of correspondence between a bus driver and passengers for the best possible user experience.
As machine learning (ML) takes over, AI avatars may become the solution that enables remote operators to only manage the rarest of situations. It is envisaged that ML could be used to learn the most common interactions between passengers and remote bus ‘drivers’ and one by one, replace the interactions with AI driven avatars. Only when rarer, more complex situations arise will humans in the call centre be required to take control.
The challenge here is to maintain a high-quality user experience while ensuring the link between the avatar and the real staff is seamless and responsive, not clumsy and disjointed.
By testing current, proven technology on autonomous buses and implementing AI avatars when they are ready, the UK will be in a position to implement driverless buses soon after autonomous cars hit our roads.
Dave Roat is Strategy Manager at Cubic