Singapore will introduce a second driverless taxi pilot as it bids to solve congestion problems.

The island city state has signed a deal with Delphi Automotive, a vehicle electronics supplier, to test a fleet of six Audis in a prominent business district.

The scheme will begin later this year and initially include drivers who could take the wheel if needed.

However they will be phased out by 2019.

If successful, when the pilot ends the following year the plan is to roll it out across the Southeast Asian state.

Customers would book the taxis through software similar to that of disruptors Uber and Lyft.

It could see Singapore, which has a population of 5.5 million, become the world's first ‘smart nation’.

"We are starting in a fairly small controlled manner but the expectation is we will continue to build… moving to an operational fleet of 30, 40 or 50 vehicles,” said Delphi's vice-president of services Glen De Vos.

"It will get a lot of attention and that is good because it helps socialise the technology."

Kevin Clark, president and CEO of Delphi, said: “We are honored to partner with the Singapore LTA on advancing innovative mobility systems, which will put Singapore at the forefront of autonomous vehicle adoption."

The cars will travel along a 3.5-mile route in the One North business area of Singapore.

Delphi, which tested self-driving cars in the US last year, plans to hold similar trials in Europe. Jaguar Land Rover will hold self-driving car trials in the UK this year.

Delphi said the technology could reduce the average price of $3 a mile for a Singapore taxi to less than a dollar.

City authorities had already signed a deal with MIT-based start-up nuTonomy to test autonomous vehicles. 

In 2014, Singapore Prime Minister Lee Hsien Loong said of its 'smart city' plans: "Our advantage is that we are compact, we have a single level of government, we can decide efficiently, we can scale up successful experiments and pilots without any delay."

Germany plans to include ‘black boxes’ in for autonomous cars to help determine responsibility in the event of an accident.