Skip to main content

Hitachi

Social Innovation

Myth-busters: self-driving cars

With a driverless car future very much on the horizon, Hitachi myth-busters are here to help clear up three of the top myths about these vehicles

Self-driving cars are arguably one of most exciting technologies being developed today, with the potential to revolutionise personal transport and tackle environmental challenges.

Technology companies are climbing over each other to be the first to make a commercially viable vehicle, but why is it that so many consumers are reticent to relinquish control of the wheel?

The answer could be found in a number of myths about autonomous vehicles.

Here are the top three myths about self-driving vehicles:

1. Driverless cars are unsafe and will require the passenger to remain alert at all times to retake control

Whilst it is true that for some “semi-autonomous” cars the passenger will be required to stay aware, so that they can regain control of the vehicle during certain traffic situations, this will not be true of fully autonomous vehicles. These vehicles, referred to within the industry as level 5 autonomous vehicles, will remove all elements of human error and allow the passenger to sit back, relax and enjoy the ride. Experts expect this to become a reality within the next ten years.

This level of safety will be boosted by the development of ever more sophisticated vehicle-to-vehicle (V2V) communication. This technology will means cars will talk to other cars, constantly exchanging data that will prevent collisions. With the growth of the internet of things, vehicles will eventually communicate with sensors embedded into signs, traffic lights and the road itself (V2X communication).

2. Self-driving cars will not work in adverse weather conditions

Adverse weather presents a significant challenge to self-driving cars because they can lose their ability to sense the world around them, particularly other moving objects in heavy snow or rain. This has meant the majority of early testing has taken place in locations with favourable weather, such as California. The initial solution was to develop weather prediction capability in the cars, so that drivers could plan ahead and avoid using autonomous vehicles in low-visibility conditions. However, companies are already starting to develop technologies that will enable autonomous vehicles to be safe in all conditions.

Google has developed software that allows its cars to determine the severity of rain and drive more cautiously when visibility is poor. It is even building in the capability to stop on the side of the road if a storm becomes severe enough. In early 2016, Ford became the first company to test a self-driving car in snow. Their solution was for the vehicle to create high resolution 3D maps while driving in favourable weather. The maps pick up the road markings, the curvature of the roads and landmarks, which the vehicle can then use to pinpoint its location in bad weather. Once located, the car can follow the map to continue driving.

3. Self-driving cars are easily hacked

The potential for car hacking is not limited to autonomous vehicles. These days, most vehicles come with computer-controlled elements and many with in-built Wi-Fi. Cyber security is a growing challenge. As computer companies discovered long ago, the best defence is to allow researchers to attempt hacks, expose any potential dangers and then improve the software based on this. This happened last year when researchers exposed a weakness in the Jeep Cherokee system that enabled hackers to switch off the engine whilst the car was in motion. The flaw was quickly corrected.

In one sense, driverless cars are safer than traditional Wi-Fi enabled car systems. It is much harder for a hacker to take over the entire system given the complexities of these systems. The autonomous car network that allows a driverless car to make decisions is more resilient to attack, as the system doesn’t fully trust one sensor that’s telling it vastly different information from the others. A potential hacker must therefore either falsify every sensor simultaneously or take control of the central processing unit. Given the difficulty of the former, security engineers can focus attention on protecting the central processing unit, significantly improving their chances of effective protection.