Tesla aims for range-spanning self-driving capabilities

While there are still questions over whether or not electric-car-maker Tesla is right to label the pseudo-autonomous capabilities offered by its current crop of models as being equivalent to autopilot, the firm is pushing ahead with its plans to remove humans from the equation altogether.

Last week it announced that it will be setting out to integrate level five autonomy on board all of its new cars from 2017 onwards, effectively meaning that they will eventually be able to drive from one destination to another without requiring any input from the people on board.

Significantly safer

Chief executive Elon Musk explained that the autonomous technology it is looking to adopt will be able to ensure that its cars are significantly safer than any other vehicles on the road which are still driven by a human.

This feat will be achieved through a combination of cameras and sensors which monitor the entire area around the car, detecting objects, obstacles and other vehicles at distances of up to 250 metres and working out how to proceed based on the information that is received.

In a video published by the firm, some of the technology is demonstrated, with a Model S saloon taking to public roads in the US and not only navigating junctions but also making its way to a parking space and manoeuvring into it successfully under its own steam.

The current Model S has onboard sensors which use sound waves to detect a range of objects and identify them, with the ability to even work out how hard a particular obstacle might be. The newly upgraded models will double the distance over which this is achievable while promising to also work in adverse weather conditions which might normally have compromised purely camera-based systems.

All of this data being persistently captured by the car presents a challenge in its own right, since it is also essential to be able to process it effectively. To this end, Tesla promises that the computational power of its new vehicles will outdo the current range by a factor of 40 while also being better equipped to combine the analysis of multiple types of input and come to conclusions quicker as a result.

The first cars

The first cars to feature this enhanced autonomous technology are already rolling out of the firm’s factories at the moment and will be winging their way to customers in North America and Europe next year. Furthermore, the hotly anticipated Tesla Model 3, which is its smallest and most affordable saloon to date, will be getting this kit as standard when it comes to market.

This month it was publicly confirmed that in spite of the billions of pounds worth of pre-orders which were placed for the Model 3 after its announcement, it is now not set to arrive until 2018. This delay comes as Tesla seeks to crank up its manufacturing output and keep up with the stratospheric levels of demand that have been generated in the wake of its ascent into the mainstream car market.

The promise of full autonomous driving capabilities being available on every single Tesla model sold from 2017 onwards is an exciting prospect but one which is tempered by the fact that the actual level five features will not be available in the short term. This is because Tesla admits that it needs to carry out extensive real-world testing in order to bring it up to speed with expectations.

It will also have to navigate the legislative and regulatory minefield currently surrounding self-driving cars in any country where it wishes to sell its products. And overcoming public concerns about the safety of vehicles which are not under direct human control will be another challenge that needs to be tackled.

Autonomous cars need to learn a lot about the world in order to deal with it effectively, and Tesla will only be able to offer incremental access to these features by activating them with software updates.