We’re closer to achieving fully autonomous vehicles than ever before. However, are self-driving cars safer than humans and completely free from error? Here’s everything you need to know about driverless cars and how they can revolutionize the future of driving!
Twenty years ago, if you were asked to describe what vehicles would look like in the future, self-driving cars would be a staple! Pop culture and science fiction movies like Minority Report, I-Robot, and Blade Runner have also popularized driverless cars as iconic features of the future.
But is that enough for you to put your life in the hands of a machine? Let’s learn about how much self-driving cars have progressed, and whether they are safer than humans on the road!
What are ‘self-driving’ cars?
A self-driving car – also called an autonomous vehicle (AV) or a driverless car – is one that uses a sophisticated system of sensors and automation to move, preferably with very little human input. The car may use various sensors like radar, sonar, lidar, inertial sensors, thermo-cameras, and emergency braking. These features help it identify the best navigation paths and steer clear of obstacles.
Some features of self-driving cars include:
- Autopilot navigation
- Obstacle avoidance
- Automatic lane changing
- Smart Route Navigation
Currently, companies like Tesla, Waymo, and Uber are at the forefront of developing driverless cars. They are in the process of tweaking existing features of Advanced Driving Assistance Systems (ADAS) like adaptive cruise control, blindspot monitors, parking assistance, and emergency braking to develop smart cars that can navigate any complex situation on the road.
Is ‘self-driving’ the right term to use?
At the current stage of development, the term ‘self-driving car’ is actually a misnomer, because they are not fully autonomous and need the intervention of humans at some level or the other. Currently, no vehicle operating on roads anywhere is fully autonomous. However, they can be categorized based on the level of the Advanced Driving Assistance Systems (ADAS) built into them.
Levels of self-driving cars
Level 0 – No Automation: In such cars, the human driver is responsible for all actions. The automated features only assist the driver in cases that require warning or intervention – like lane changes and automatic braking.
Level 1 – Driver Assistance: Here too, the driver is responsible for the car’s actions. However, when the ADAS system is engaged, it can assist in accelerating, decelerating, and steering the vehicle. Examples of such automation include adaptive cruise control and lane-keeping assistance.
Level 2 – Partial Automation: In such cars, the driver remains attentive and engaged on the road. The system provides continuous monitoring and assistance for acceleration, braking, steering, and more.
Fun fact: Most self-driving cars of today only have up to Level 2 automation – which means even when the system is engaged, it continuously needs the attention of the driver to safely navigate roads. Cars with automation levels 3, 4, and 5 are still in development and are not available for commercial purchase.
Level 3 – Conditional Automation: Here, it is the system that is in charge of driving the vehicle. However, the human will have to respond immediately in the event of an anomaly.
Level 4 – High Automation: The car can navigate within a limited region and under specific conditions while the driver acts as a passenger. Even if the human responds late, the system can still pull over and stop safely by following a guiding system.
Level 5 – Full Automation: Level 5 involves a fully automated car that is capable of navigating any region, under any situation, and without the intervention of the human driver.
Benefits of self-driving cars
- Better road safety: Human error accounts for 95% of all road accidents. A proper self-driving system can help reduce driver errors behind most accidents. It could also benefit pedestrians and cyclists, who are frequently the victims of driving accidents.
- Mobility for disabled and senior citizens: Driving a car is often a hard task for persons with disabilities and senior citizens who have cognitive issues. With automated vehicles, they will have the freedom and mobility to go about their lives with ease.
- Reduced traffic: A well-connected driving system that also has excellent vehicle-to-vehicle communication can reduce frequent traffic bottlenecks and inner-city roadblocks.
- Improved productivity: Self-driving vehicles can also self-park, leaving their owners to get off at the doorstep of their destinations without needing to drive around for parking.
- Environmental benefits: Automated vehicles could also include smart pollution sensors and fuel efficiency monitors that can help reduce greenhouse gases and pollutants from vehicles.
Are self-driving cars safer than humans?
In certain aspects, self-driving cars are safer than humans. They are efficient at eliminating most accidents caused by human error. ADAS systems never get tired, do not text while driving, and can eliminate accidents under the influence of alcohol.
However, driverless cars still trip up when it comes to certain situations:
- Driverless cars perform worse in edge cases (extreme or highly improbable situations that are not expected in the real world).
- Bad weather can impair the car’s sensors and systems, making them more prone to errors.
- Driverless cars are still not able to navigate the chaos of inner-city traffic like in New York, Boston, and other megacities.
- Most safety regulations for automated vehicles are not mandatory, and regulators are also unsure of how to standardize laws across states.
- Computers also lack intuition and instinct which is common in human beings. This can affect their response in some situations. For example, a human driver may avoid an open manhole on the road, but a self-driving car may not see the need to avoid it and may hit a person coming out of it unexpectedly.
Thus, self-driving cars require further engineering to improve their response to complex situations – particularly edge cases. We may still need a few decades of improvements before they can be truly termed ‘safer’ than humans.
Can someone hack a self-driving car?
Remember the scene from Fate of the Furious (2017) where thousands of autonomous vehicles are hacked and hotwired to stop the protagonists? Well, turns out that’s a real problem!
Driverless vehicles are also vulnerable to cyber-security risks. The European Union Agency for Cybersecurity (ENISA) found that automated vehicles had a high risk of being hacked. The cyber-attacks could overload the sensors with light beams, use machine learning to find cracks in their systems, overwhelm the obstacle detection systems, and more.
However, most of these attacks were done in controlled environments. Until fully autonomous vehicles are made commercially available, we may not know the full extent to which they can be hacked.
How many deaths have self-driving cars caused?
Technical issues and oversight have caused some unfortunate deaths in recent years.
- Tesla’s self-driving cars have been involved in at least 10 accidents resulting in 12 deaths.
- Waymo’s autonomous cars have been involved in at least 18 accidents during the period 2019-2020.
- Uber’s self-driving cars were involved in 37 crashes. In 2018, a fatal accident resulted in 1 death.
It is still hotly debated whether the deaths were due to the autopilot system or due to oversight on the human driver’s part. Because of this, the exact numbers are hard to estimate.
Will self-driving cars cause more accidents?
It’s clear that driverless cars still have a long way to go before being completely accident-free. Self-driving cars are still prone to some common accidents like rear-ending and side-swipes.
Self-driving cars also have an accident rate of 9.1 crashes per million miles driven – which is more than double that of regular vehicles. The saving grace is that crashes involving self-driving cars have much less severe injuries than regular accidents.
Will self-driving cars replace drivers?
Self-driving cars will probably never completely replace human drivers. We still have a long way to go – at least 2040 – before we see some semblance of Level 5 automation in driverless cars. They could become popular modes of transport within specific regions that can be ‘geofenced’. Examples of such regions are theme parks, universities, and corporate campuses. However, driving on major roads and freeways will still need the supervision of a human driver.