Intelligent fleet: The long road to autonomous fleets

From ‘infancy’ through ‘adolescence’ and on to ‘adulthood’ the long road of autonomous vehicles can be likened to that of human development.

Autonomous cars will be driving between London and Oxford as early as next year, but it could be a decade before they arrive in company car parks, experts believe.

There is a global race to develop connected and autonomous vehicle (CAV) technology as governments believe it will play a key role in future industrial and business development and officials want the UK to lead the way.

Hundreds of millions of pounds is being poured into CAV projects that will bring autonomous vehicles onto the UK’s roads.

Already, fully autonomous cars are driving around the streets of Oxford under a Government-funded project run by the Driven consortium and by next year there will be trials of a fleet of driverless vehicles travelling between Oxford and the capital and back.

Vehicles will be operating at Level 4 autonomy – meaning they have the capability of performing all safety-critical driving functions and monitoring roadway conditions for an entire trip, with zero passenger occupancy. 

No CAV trial at this level of complexity and integration has been attempted anywhere in the world.

Graeme Smith, CEO of Oxbotica and a former director of telematics with Ford, is leading the project.

 

He says: “We are also starting to think about ecosystems of vehicles, how might fleets of vehicles work together in an environment, how might they interact, what might these vehicles chatter about, what information will they exchange.

“Our intention is to test these vehicles both in Oxford and in London and sometime next year we will start to run a fleet of six-10 vehicles backwards and forwards between Oxford and London.

“We have a very ambitious schedule, but we are already on the road testing and doing real-world validation.”

However, it is a big leap from closely monitored trials to fleets of vehicles being a common sight throughout the UK’s road network.

Industry analysts suggest autonomous vehicles are still decades away, despite several manufacturers including advanced autonomous elements in their current cars.

A pioneer is Tesla, which has led the global race to autonomy for years with its Autopilot system that allows autonomous driving while the driver is behind the wheel.

It includes a ‘summon’ function on the key fob so a driver can tell the car to drive a short distance to their location, for example, out of a tight parking space.

But Tesla’s experience also highlights the reasons why development will take so long, following a series of incidents while Autopilot was engaged, including several fatalities.

In the first reported autonomous vehicle fatality in Florida, US, a Model S in autopilot mode failed to distinguish a large, white 18-wheel truck and trailer crossing the highway in bright sunlight.

Tesla pointed out that the handful of cases represent the tiniest fraction of a percentage of miles travelled on Autopilot and that the system is much more likely to prevent incidents.

It also said the system is a driver aid and drivers must be ready to take control of the wheel at all times.

Therefore, the process of exhaustive testing required before a vehicle becomes safely truly autonomous will take years and it may be decades before self-driving vehicles can be used outside the tightly-controlled confines of pre-defined ride hailing routes and other public transport uses.

According to Mobileye chief communications officer and senior vice-president Dan Galves, autonomous cars are “precocious children” equivalent to the early stage equivalent of human development.

He told a recent Michelin Movin’On conference it would be 2020 before they became ‘adolescents’ that still require substantial training and 2025 before they are ‘fully-grown’.

When that happens, the benefits could be enormous. For example, studies suggest autonomous vehicles are anticipated to reduced fatalities by 99%, taking road deaths in the US from around 40,000 to below 350 fatalities per year.

This suggests that insurers, as well as legislators, could have a major role to play in encouraging acceptance and uptake by heavily reducing premiums for autonomous vehicles or even refusing to insure non-autonomous models.

Karl Iagnemma, (pictured below)  co-founder and CEO of driverless fleet software firm Nutonomy, says: “We tend to think autonomous vehicles will be on streets worldwide overnight, but it won’t happen that way.”

However, fleets still need to consider their strategic approach to the new technologies arriving on vehicles.

For example, there are training and disciplinary procedures to consider when drivers use the current generation of connected and semi-autonomous vehicles safely and correctly.

With modern smartphones able to connect directly to vehicle systems and display their screens on the dashboard, driver distraction can become a real risk as the connected world fights for their attention behind the wheel.

An additional concern is servicing. 

Modern cars now feature a range of complex software, coding and sensors, which requires careful calibration. 

If these sensors become misaligned during servicing, drivers could put their trust in an autonomous vehicle that is incapable of keeping them safe.

Windscreen replacement is a particular area of concern, as safety technologies, such as autonomous emergency braking (AEB) and lane-keeping assist, use radar, laser or camera sensors which are often located behind a car’s windscreen.

The same applies to more convenience-oriented features such as adaptive cruise control.

Vehicle safety consultancy Thatcham Research says windscreen-mounted ADAS technology is currently fitted to approximately 6% of vehicles on UK roads, with this proportion likely to rise to 40% by 2020.

 

The six levels of automated driving systems – from 0 to 5

In 2014, the Society of Automotive Engineers (SAE) outlined six levels of automated driving systems, ranging from complete human control to full autonomy.

This classification has been adopted globally as a benchmark for the development of self-driving cars.

 

Level 0 – full driver control (today)

An automated system issues warnings, but has no vehicle control.

 

Level 1 – hands on (today)

The driver and an automated system share control over the vehicles, but the driver must be ready to take full control at any time.

An example of this is cruise control, where the driver controls the steering and the vehicle maintains speed.

 

Level 2 – hands off

The automated system is in full control for accelerating, braking and steering.

The driver must monitor and be prepared to intervene at any time. Although it is called ‘hands off’, driver contact with the wheel is typically mandatory.

 

Level 3 – eyes off

The driver can safely turn their attention away from driving tasks.

The vehicle will drive itself and handle any situation that requires  an immediate response, even emergency braking.

The driver must still be prepared to intervene when called upon.

 

Level 4 – mind off

Like Level 3, but no driver attention is required.

The driver may safely sleep or leave the driver’s seat. Self-driving is supported in limited areas or under special circumstances.

Outside of these circumstances, the vehicle must be able to abort the trip and would park the car if the driver does not regain control.

 

Level 5 – wheel optional (2040)

No human intervention is ever required in any circumstances or environment, including isolated areas.

One example would be a driverless taxi or shuttle. 

 

Projected autonomous vehicle timeline

 

2017 Testing already taking place in major cities. Semi-autonomous systems available on growing range of cars.

 

2020s Testing in 100 cities worldwide, autonomous trucks on the roads. Driver-assistance features grows to a multi-million pound sector.

 

2030s Significant ramp-up of autonomous vehicle development. Up to 25% of miles driven in some countries could be autonomous. Millions      of autonomous vehicle sales globally.

 

2040s Combination of autonomous fleets and personally-owned cars – 75% of cars will be autonomous.

  

2040s The number of injuries and fatalities from road incidents is reduced by 90%.

 

2060s Cities restrict human driving

 

Source: Nutonomy/Michelin Movin’On