Curated by UCL

BLOG: Autonomous cars – driving into the future, with no hands on the wheel!

Road accidents are currently the ninth leading cause of death globally, killing around 1.3 million people every year.[1] 90% of these accidents arise due to human error. As congestion on city streets continues to increase, these numbers are likely to go up. But what if there was no scope for any human error? What if a car could transport its passengers safely without the need for human input? This is the goal of car manufactures, through the invention of self-driving cars – an idea that is now becoming a reality. If everything goes as planned, we will soon be able to benefit from safer, more productive journeys and shorter commuting times. The realisation of autonomous cars would also mean that those unable to drive could commute independently.

The most crucial aspect of a self-driving car is ensuring that the system can see what a driver would. These cars therefore consist of various sensors and computers, which they can use to understand the landscape around them.

One of the most important components of these cars is a LiDAR[2] (light detecting and ranging) sensor. LiDAR sensors send out pulses of light, which bounce off of surrounding objects. The obtained reflection data then enables the creation of a high-resolution 3D map of the car’s surroundings,[3] making edges of roads and lane markings identifiable. Video cameras help the car to detect colour and hence interpret traffic lights.[4] The cameras also detect road signs and obstacles (such as pedestrians), as well as other cars. Being limited by visibility and other weather conditions, the video cameras are assisted by radar sensors[5] similar to those already in use in adaptive cruise-control systems (to measure vehicle speed).[3][4] Such cars are also equipped with ultrasonic sensors in the wheels that can detect the distance from curbs and neighbouring vehicles to a high degree of accuracy. This provides assistance during parking.[4] The data that’s fed to the car through all of these sensors must be interpreted to enable judgements to be made. For this purpose, a central computer—usually situated in the boot of the car—acts as its brain. It analyses data from the various ‘organs’ (e.g., cameras, RADAR and LiDAR) and uses the information to manipulate the vehicle’s acceleration, steering and braking .[6]

The above model for the working of a self-driving car is adopted by many major competitors, including Google’s self-driving division (now called Waymo/cite{bib3}), which is currently leading the race to driverless cars. Other key players include Uber and a Chinese web company, Baidu.[7][8][9]

Some car manufacturers that have opted for a slightly different solution. The most notable of these is Tesla Motors. Instead of LiDAR technology, it uses a “computer vision-based vehicle detection system”.[10] These cars are equipped with a forward-facing camera, forward radars, ultrasonic sensors and a GPS system which tracks the location of the car using high-precision mapping.[10] Unlike Waymo, these are normal cars (complete with steering wheel, pedals, etc.) with some self-driving capabilities. They can be switched to autopilot mode on highways, can change lanes, steer and park on their own, and even avoid collisions; but ultimately, the driver always has to be in charge of the vehicle. The features are undergoing continuous development but, despite Tesla’s promise of delivering fully self-driving cars by 2019, the company seems to be running behind on its deadlines.[11]

However, the idea of driverless cars is still an alien concept that must be accepted by the public before it can be successful. Even though most major competitors have begun extensive testing on city roads,[9][12] self-driving cars still face a number of challenges. According to the CEO of Volvo Group, Martin Lundstedt, “[it] will not be technology that will set the pace [for the adoption of driverless cars], it will be regulations and liability rules.”/cite{bib10} Additionally, autonomous cars will not be perfect. Even if their use reduces the number of road accidents, incidents can still occur. The question is, will people be as forgiving when the accidents are caused by machines as compared to humans? And in such an incident, who would be to blame? The car? The legal system would need to make changes to accommodate such instances. Finally, one popular belief among concerned citizens is that motor vehicle operators would lose their jobs. However, a contrasting view is that autonomous cars will not take away jobs, but rather make them better – and that although the popularity of certain jobs would decrease, new jobs would be created.

The above issues mean that the autonomous car is unlikely to be instantly adopted by everyone. The shift to driverless cars will be an extremely gradual one, but one with multiple long-term advantages. Road accidents, as well as noise pollution (due to car horns), would be drastically reduced. Furthermore, as most autonomous cars would be electric, atmospheric pollution would also decrease substantially. Some even believe that autonomous cars would imply shared ownership of cars, and hence the removal of large parking spaces, thereby enabling the expansion of parks and walking paths. It’s also thought that the cars’ ability to communicate with one another would reduce accidents and congestion, thus improving utilisation of road space. Indeed, according to Tesla CEO Elon Musk, the Autopilot system fitted in new Tesla cars “can help reduce accidents by as much as 50%.”[13] Self-driving car technology continues to develop apace, and may arrive sooner than we expect.

References

  1. Road Crash Statistics, Asirt.org. 2017. Accessed 3 December 2017.
  2. O. Cameron, An introduction to LiDAR: The Key Self-Driving Car Sensor, Voyage, 2017. Accessed 3 December 2017.
  3. T. B. Lee, Lidar tells distance, radar tells velocity, this new sensor aims to do both, Ars Technica, 2017. Accessed 3 December 2017.
  4. How does a self-driving car work?, The Economist, 2017. Accessed 3 December 2017.
  5. P. Ingrassia, Inside Google's push to shape the rules of the driverless road, Reuters, 2016. Accessed 3 December 2017.
  6. J. Armstrong, How do driverless cars work?, The Telegraph, 2016. Accessed 3 December 2017.
  7. A. France-Presse, Uber launches driverless car service, The Telegraph, 2016. Accessed 3 December 2017.
  8. M. Chandler, Google, Baidu, Tesla Gunning Self-Driving Car Development, Investor's Business Daily, 2016. Accessed 3 December 2017.
  9. C. Clover, E. Feng and S. F. Ju, Baidu launches public road tests of autonomous cars in China, Financial Times, 2017. Accessed 3 December 2017.
  10. G. Gates, K. Granville, J. Markoff, K. Russell and A. Singhvi, The Race for Self-Driving Cars, The New York Times, 2016. Accessed 3 December 2017.
  11. A. J. Hawkins, Tesla's Autopilot is supposed to deliver full self-driving, so why does it feel stuck in the past?, The Verge, 2017. Accessed 3 December 2017.
  12. Journey - Waymo, Waymo. 2017. Accessed 3 December 2017.
  13. C. Thompson, Here's how Tesla's Autopilot works, Business Insider, 2016. Accessed 3 December 2017.

About the Author

Student, University College London

Keen Undergraduate at University College London pursuing Mechanical Engineering owing to an interest in the automotive industry, manufacturing and product design.

Articles

Why Do Computers Use So Much Energy?
Geoengineering Is Inevitable
5 Tips for Building Fog Networks
A Short History of the Missing Universe