The robot fleet
How much time do your employees waste on the road driving? For many businesses this time is significant, but vital: after all they have to get to their meetings and jobs. But they also could be doing something more productive in those lost hours on the motorways of Britain.
What if they could be sitting in their car or van, filling out job sheets, meeting reports or catching up on emails instead? There have been many studies on just how much time is lost by business drivers in traffic jams and congestion alone, with some quoting three to five working days per driver a year. But if much of this could be taken out of the working day, then time and motion become one and the same thing, as work and travel meld into one entity.
The advocates of driverless cars claim that this will be possible in the near future, perhaps in the next five years, when the car or van will take an instruction and set off all of its own accord. They also reckon this will cut the number of accidents – in approximately 90-95% of all road traffic accidents, human behaviour is partially or fully responsible – and that fuel economy and traffic will improve too.
However, there are many challenges around technology, cost, public acceptance, legality, liability and infrastructure.
Foremost is having the necessary technology available to do the job, and to define exactly what that job is. For example, there is full autonomy, where the driver could enter the destination into the sat nav, and take a nap until the car gets there, partial autonomy in a road train, or autonomous systems that can be switched on and off, or intervene automatically to avoid crashes or offer partial machine-only driving.
For the fully autonomous car, the future is further away, because the technology is more complex and more expensive. Google’s self-driving car, which has been developed intensively for the past three years, has now clocked up half a million miles of safe driving on US roads. But it requires a staggering amount of processing power, of nearly one gigabyte of data every second, in combination with GPS, laser rangefinding and mapping to determine its location and route and as a result, even Google admits that a projected cost of market-ready cars would be $150,000, of which nearly half is on the laser rangefinder. And a number of other technical obstacles remain. For driverless cars to work, every inch of road, every junction, road sign and signal everywhere will have to be mapped in perfect detail. This is being done anyway to support navigation systems for both cars and smartphones, but this takes time and a vast amount of money.
However, there are cheaper alternatives. Researchers at Oxford University claim to have built a system of self-driving-car cameras for £5,000, which could be eventually cost as little as £500.
Its system is less ambitious than Google’s more pervasive one, by using camera in concert with a database of stored 3D images. When the cameras pick up an environment and road system it has committed to memory, it tells the driver it can take over control. This would clearly be something business users and delivery firms who use specific routes or commutes would find of interest.
The price will be of interest to businesses too, because according to a JD Power & Associates survey published earlier this year, 39% of people would be interested in owning an autonomous car, but only 21% said they’d be interested if it cost them an extra £2,000. The Oxford University system, while some way off being market ready, could be retrofitted too, its makers claim.
The sensors are the key in these systems, because without good quality information, full autonomy can never be delivered. For example, Bosch’s high-performance long-range radar sensor can detect objects up to 250 metres away. In addition, a stereo video camera can detect objects in 3D, calculating how far objects are away from a vehicle, as well as the direction in which they are moving.
Fully autonomous driving will come about one step at a time. At first, driving on highways with an ever-greater degree of automation and at ever-higher speeds will be possible, until the highway pilot can take over the entire trip.
But two major challenges remain: first, inner-city driving, since automated vehicle functions have to deal with dense traffic involving a large number of road users traveling in every direction; and second, developing a concept to ensure that the system’s functions operate reliably in all types of driving situation.
It seems that before 2020, more simple – relatively speaking – automatic operating systems will become more prevalent, such as road trains.
Last year in Spain a road train comprising a Volvo XC60, V60 and S60, plus one truck automatically driving in convoy behind a lead vehicle operated on a public motorway among other road users. The SARTRE Project believes the road train offers the attractive possibility to do other things while driving, promotes safer transport, and reduces environmental impact.
Basically, a professional driver leads the vehicle platoon, for instance in a truck. All the following vehicles are driven autonomously at speeds of up to 60mph – in some cases with no more than a four-metre gap between them – thanks to a blend of present and new technology.
The cars drive close to each other and reap the benefit of lower air drag, and the reduced speed variations improve traffic flow, creating more efficiently utilised road capacity. Inter-vehicle reaction response times are very quick thanks to the co-ordinated technology.
Erik Coelingh, technical specialist at Volvo Car Corporation, who are partners on the project, said: ‘The energy-saving potential is 10-20%.
‘The road train is the best of two worlds. You can enjoy all the multi-tasking possibilities of public transportation, behind the wheel of your own car.’
A prototype Human-Machine Interface including a touch screen for displaying vital information and carrying out requests, such as joining and leaving the road train, while a prototype vehicle-to-vehicle communication unit that allows all vehicles within the platoon to communicate with each other and vary their speeds accordingly if needed.
An EU report in 2010 into autonomous driving saw the value in the development of these systems to the point of full machine control, but it recognised there were many hurdles before manufacturer, owner and operator all saw the benefits.
The report said: ‘The most challenging issue for the private sector is to find applications that bring value to all the players involved, from OEM and supplier, to fleet owner, insurance company and end user.
‘Then, a quick uptake will happen and significant penetration rates of automated driving vehicles will arise. Besides the individual benefit of increased safety and fuel efficient driving, a significant penetration rate of automated driving vehicles will increase the overall benefit of traffic congestion reduction, reduced emissions and traffic safety.’
Recently though critics have claimed that the US government is being more proactive in developing legislation to allow them to run on public roads than the European Union. There are significant legal hurdles around liability: just who is at fault should an accident occur? The OEM, “driver” or vehicle owner? It really is the philosophical stuff science fiction writers have mulled over for decades, when the relationship between robots and humans becomes entangled. Do drivers really want to be subordinate to machines while in the car? Certainly if you see the erratic way some rain sensing wipers or active cruise control systems work on cars at the moment, the future of fault-free robot cars is rather further off than the experts would claim.
But the National Highway Traffic Safety Administration is all but admitting that the day is coming when the robot is at the wheel.
‘America is at a historic turning point for automotive travel,’ the agency said in its preliminary statement on autonomous cars.
‘Motor vehicles and drivers’ relationships with them are likely to change significantly in the next ten to twenty years, perhaps more than they have changed in the last one hundred years.’
Driving autonomy in action
Audi – Piloted Parking
What is it?: Automatic location of parking spaces, and the ability to manoeuvre into them.
How it works: Using a wireless connection, the car is able to download a map of available parking spaces within a car park, and then use this to navigate to one of them. Once there, it uses ultrasound sensors to park automatically, then turns off the engine, locks the doors and contacts the driver. All the driver has to do is pull up outside and activate the system using a smartphone app.
When?: Audi demonstrated Piloted Parking at the Consumer Electronics Show in Las Vegas this year, but has yet to reveal when it will go on sale.
BMW – ConnectedDrive
What is it? Adaptive cruise control with the ability to change lanes.
How it works: Using data from camera, laser and radar systems, currently employed for accident avoidance and lane-keeping, BMW’s system is able to position the vehicle correctly in its lane and maintain a speed. Detailed GPS data forewarns the car about changing road layouts, while the windscreen-mounted camera means the car can overtake and change lanes to avoid joining traffic.
When?: Traffic Jam Assistant and automated parking will debut in the new X5 and i3 electric vehicle later this year.
Mercedes-Benz – Intelligent Drive
What is it?: Hands-free driving through slow traffic jams.
How it works: Stereo cameras at the top of the windscreen can recognise lane markings and groups of vehicles driving ahead together, intelligently deciding whether to follow the road or other road users in heavy traffic at up to 37mph. This can take control of steering, brakes, throttle and gear changes, allowing drivers to sit back and even let go of the steering wheel in traffic jams, reducing stress.
When?: On sale this summer in the new S-Class.
Google – Self-Driving Car
What is it?: Fully autonomous driving, using web-based maps.
How it works: Google is manually collecting detailed electronic maps of test routes , including junctions, lane markers and road signs, which are stored online. The car is able to combine this with GPS data, video and information from laser and radar detectors to avoid other traffic and pedestrians, navigating to addresses and points of interest stored on Google’s database.
When?: Google was the first to demonstrate an autonomous car, but as yet no manufacturers have adopted it.
VOLVO – Road Trains
What is it?: Groups of cars which autonomously follow a manually driven lead vehicle.
How it works: Volvo is the sole manufacturer partner in the SARTRE project, which allows a “platoon” of vehicles to follow each other automatically along stretches of road. The platoon is led by a vehicle with a professional driver behind the wheel, and following cars download and mimic the way this is driven while radar, camera and laser systems maintain a constant six-metre gap from each other.
When?: Volvo has recently unveiled a self-parking system similar to Audi’s, saying its next generation of modular cars starting with the XC90 next year will feature some autonomous driving features.
Volkswagen – Temporary Auto Pilot (TAP)
What is it?: Automatic cruise control which can recognise road signs.
How it works: TAP uses sensors from adaptive cruise control, lane-keeping and accident avoidance systems to monitor surrounding traffic and road signs at up to 80mph. Cameras at the top of the windscreen monitor road signs and lane markings, allowing it to obey overtaking rules and speed limits, while a radar allows it to maintain a safe distance from the car in front even in start-stop traffic.
When?: Volkswagen showed a concept vehicle two years ago, but has yet to announce when the system will be commercially available.