Automotive

Singapore hails driverless taxis

Singapore hails driverless taxis
nuTonomy is looking to test its fleet of autonomous cars
nuTonomy is looking to test its fleet of autonomous cars
View 4 Images
The team behind the driverless taxi
1/4
The team behind the driverless taxi
nuTonomy is looking to test its fleet of autonomous cars
2/4
nuTonomy is looking to test its fleet of autonomous cars
The system uses "formal logic" to make human-like judgments on the road
3/4
The system uses "formal logic" to make human-like judgments on the road
NuTonomy's cars use "formal logic" in order to make smart road decisions
4/4
NuTonomy's cars use "formal logic" in order to make smart road decisions
View gallery - 4 images

Autonomous car software designer nuTonomy is looking to bring driverless taxis to the streets of Singapore. Following successful obstacle course tests, the company is in the process of seeking government approval to test its driverless technology on roads in the OneNorth business district, with plans to eventually offer the service at a cost comparable with public transport.

nuTonomy originated as a spinout company from a group of researchers at MIT, where both the CEO (Karl Iagnemma) and co-founder (Emilio Frazzoli) taught as professors. The pair worked on driverless "golf cart" taxis in 2009 before NuTonomy became a standalone company in 2013 in order to further develop driverless car software. In January this year the company received $3.6 million in funding from investors including Samsung and Signal Ventures, and is currently developing an auto-park feature for Jaguar Land Rover.

Singapore has emerged as the key testing ground for the technology, however. It's here that vehicles equipped with nuTonomy's software recently passed their first driving test by successfully navigating a custom-made obstacle course without issue. The next big step is to get approval from the Singapore government, something that Frazzoli is hopeful of. "As soon as people start seeing this is real and this is working, and this is the benefit it can provide to the public, then people will start seeing that this can provide a solution to the mobility problem in big urban centers," he says.

The company says that its driverless taxis would reduce traffic congestion by choosing optimal pathways through the city. It believes that costs per trip could be similar to a bus fare, helped by the fact that there is no need to pay or insure drivers, and that waiting times could be kept to under 15 minutes, on par with public transport.

The team behind the driverless taxi
The team behind the driverless taxi

Part of the appeal of these cars is the reduction of pollution in heavily populated Singapore. The nuTonomy fleet will be electric, producing a smaller amount of local greenhouse gas emissions than a standard taxi. Singapore has long struggled with air pollution levels and today the city-state battles with a constant haze, regularly scoring a 60-70 (moderate to dangerous) on the Pollutant Standards Index.

In early 2015 the city's Land Transport Authority (LTA) gave the green light to driverless car testing in the One-North business district. "The 6-km long network of roads will give successful applicants the opportunity to test their vehicles' navigation controls in a real-world environment," stated the LTA.

NuTonomy is not the first company to seek approval for testing in One-North, with the National University of Singapore also applying to deploy driverless cars.

In 2014 Frazzoli and his colleagues published a paper in Road Vehicle Automation, outlining a coordinated fleet-based system for autonomous vehicles that borrowed algorithms from Frazzoli's previous work with US military drones. This research estimated that a fleet of driverless taxis "whose size is approximately 1/3 of the total number of passenger vehicles currently in operation" could meet the city's transport needs. "This was a big sign of impact for [the Singaporean] government," said Frazzoli. "At first we were asking them to let us test cars there – then they were asking us to come test."

Teaching a car road rules is is only part of the equation, as sometimes breaking minor road rules is the most efficient way to travel. nuTonomy's driverless taxis will use what it calls "formal logic" to navigate around Singapore's busy streets. Frazzoli uses the example of driving around a double-parked car, the taxi observes that there is no oncoming traffic and overtakes safely. "These are situations we encounter every day, and we use our judgment to understand the rules we can violate," said Frazzoli "We have these same judgments embedded in our algorithms."

NuTonomy's cars use "formal logic" in order to make smart road decisions
NuTonomy's cars use "formal logic" in order to make smart road decisions

Major companies such as Google, Volvo and Ford are experimenting with driverless technology, with Google launching a fleet recently in California. Google's cars achieve a level three classification on the National Highway Traffic Safety Administration's system, requiring that the driver can fully override the car's controls and operate the brakes and steering. nuTonomy's fleet will be the first to be classified as level four; a totally autonomous vehicle.

There are no firm details as to when the on road trial may start, but the company believes it could have thousands of driverless taxis in the streets of Singapore alongside public transport in just a few years.

Sources: MIT, nuTonomy

View gallery - 4 images
5 comments
5 comments
Mel Tisdale
The picture gallery shows that it retains a steering wheel. I wonder if it retains all the other controls. Imagine some drunken lout(s) gets in one of these in preference to walking (staggering?) their way home. They could wreak havoc in seconds. If the steering system were made so heavy it prevents such behaviour, it ceases to be practical from a road safety perspective. And what fun could be had by stamping on the brakes for no reason other than to scare those in following vehicles, especially if they are all part of the same group.
It is small wonder that Singapore is a chosen venue for testing. I wonder what this system is like in snow, both from a vehicle handling point of view and from a sensor performance one too. It won't drive round an obstacle that it cannot see because the sensors are blocked with snow.
If these vehicles are ever going to be more than just a gimmick, they need to all be operating to standard, yet one never sees mention of such in any of the articles about them. Mind you, they will need a large helping of common sense, too, which also appears to be lacking.
Martin Winlow
Mel,
Don't you imagine that someone intelligent enough to design an autonomous car is going to be able to think up your scenarios and imprelent a system to prevent them? What is there to stop any idiot in any ordinary minicab from yanking on the handbrake or opening a door in fornt of another car (or worse, a pedestrian)???
People seem to think autonomous cars are decades away and yet my car is at least 80% autonomous already and quite capable of safely driving hundreds of miles by motorway without any intervention at all (unless you want to change lanes where it requires the driver to decide when it is safe to do so and instruct the car accordingly by operating the indicator stalk in the relevant direction). Even Renault/Nissan CEO Carlos Ghosn stated only last week that his companies' cars will be capable of *full* autonomy within 5 years...
I'm just intrigued they chose to use an i-MiEV for their development.
splatman
@Mel Tisdale, don't worry, they will have a man walking in front with a red flag. There, feel better now?
Bob Flint
Again in the land of heat & sunshine, need real world testing Montreal in April 0 degrees C snow & blowing snow...
Stephen N Russell
Can we export, lisc this to the US alone, be super sweet Hire ex taxi drivers to service new taxis or input data for Maps alone Radical, Bring to the US & Canada.