Self-driving cars are no longer a pipedream

It was in October 2005 when Sebastian Thrun and his team from Stanford University in the US created an autonomous robot that traversed 175 miles through the Mojave Desert. Little did they know then that their robot ‘Stanley’ (a modified VW Touareg R5) would shake up the auto industry. Google subsequently hired Thrun and founded X labs that spawned innovations such as Street View and Google Glass. The most interesting though was the self-driving car (SDC) and it was not too long before Google showcased a 2-seater (earlier this year). And when Ford’s ex-CEO Alan Mullaly joined the Google Board, it was crystal clear that the SDC is no pipedream.

Of course, autonomous driving technology is not unfamiliar to carmakers. Last year, Daimler gave a glimpse of the future with its 2014 S-class and followed it up this month with the Actros truck that can steer, accelerate and brake on its own. SDCs will alter mobility dynamics in a big way. Technologies that avoid crashes and take decisions based on traffic conditions will reduce fuel consumption and emissions. The entire vehicle design that revolves around crash avoidance will no longer need expensive and heavy steel structures or airbags. It will also offer more interior space for designers to get ultra creative.

On the flip side, though, SDCs might be used for longer commutes than current cycles which could defeat the purpose of easing traffic. Regulatory bodies are only too aware of this changing scenario. Last year, the National Highway Traffic Safety Administration (NHTSA) of the US released a preliminary policy on automated vehicles which listed three technologies: a) in-vehicle crash avoidance systems that warn or provide limited automated control of safety functions; b) vehicle-to-vehicle communications that support various crash avoidance applications; and c) self-driving vehicles.

Based on these technologies, NHTSA further segmented four levels of automation. The first is level 0 or no automation. Here, the driver is in sole control of the brake, steering and throttle. However, cars may have driver warning systems for lane departure or forward collision.

Level 1 involves function-specific automation. There is one autonomous control function which could either be electronic stability, adaptive cruise, self-parking, lane assistance or braking collision system.

Level 2 confers partial autonomy (combine function automation) like a traffic jam assist system where the car drives itself in stop-and-go traffic. Here two primary controls function in unison.

In Level 3, there is limited self-driving where the person behind the wheel is expected to be available for occasional control only.

Level 4 is complete self-driving.

While Google is expected to license the self-driving software rather than make cars, automakers have already geared up to launch systems up to Level 2 beginning next year. Automated driver assist systems are being offered as an added kit in premium models from Audi, BMW, Volvo, Tesla and Mercedes.

In the US and Europe, regulators have already made Level 1 systems like ESC mandatory and have included lane departure warning and forward collision warning in their NCAP (New Car Assessment Program).

Tier 1 suppliers are quite upbeat about SDCs and this is evident from the spurt in R&D investments as well as the spate of mergers and acquisitions. Recent reports of ZF planning to buy TRW clearly point to an objective of making products intelligent using TRW’s active and passive safety systems.

Or take Bosch whose profitable automotive technology division invests €400 million each year on development of electrical and electronics systems. Bosch expects annual sales of driver assistance systems to exceed €1billion from 2016.

However, companies keen on supplying systems to SDCs need to be careful in the backdrop of global vehicle recalls. Automotive electronics must perform in harsh environments which could involve longer high temperature cycles. This leads to fatigue, a common cause for electronics failure. Today suppliers have to comply with less than 10 part failures per million produced each year.

In the scenario of recall costs, Japanese supplier Takata comes to mind. Last week, it announced taking a $450 million hit due to costs arising from faulty airbags. Over the last five years, Takata had to cope with 12 million vehicle recalls with 2013 alone accounting for $300 million in costs.

For suppliers, the script will get even more complicated when SDCs enter the arena. Some challenges could include: a) sensors to recognise 360° environment; b) developing redundant system architecture; c) reliability in the event of malfunction and software hacking; d) high-precision map data; and e) legal regulations. With all cars tipped to become SDCs over the next four decades, it is a small price to pay now.

The writer is AVP, Symphony Teleca Corporation

COMMENT NOW