Levels of Autonomy of Self-Driving Cars4 min readReading Time: 3 minutes
As the concept of self-driving cars gains currency, auto manufacturers are wasting no time in jumping on the bandwagon. Both Tesla Motors and Google have tested a few prototypes, observing mixed results. Currently marketed as a revolution in personal transportation, self-driving cars pose many questions about urban transportation. Let’s go through the levels of Autonomy of Self-Driving Cars.
Levels of Autonomy of Self-Driving Cars
Human drivers rely entirely on their vision while driving, constantly interpreting their surroundings to make decisions. Self-driving cars instead rely on sensors (blind spot monitoring and collision warnings), connectivity (receiving latest data on other cars on the road, weather, traffic patterns and congestion, construction, maps and GPS) and software algorithms (decision-making for braking, speed, and routes).
To accurately define “autonomous driving”, SAE (Society of Automotive Engineers) International offers a helpful level-by-level guide. Since 2014, these guidelines have been adopted by several reputed bodies and have become the industry standard. According to SAE, there are six levels of autonomy. Each level has a unique set of requirements that a vehicle must meet. Today, the SAE’s approach remains the industry’s most widely accepted classification system.
Level 0 – No automation
Most of the vehicles we have today operate at this level. The driver dictates driving actions, manually steering the vehicle and applying brakes. The vehicle has no automated technology but has some basic ones such as blind spot and collision warnings.
Level 1 – Driver assistance
At this level, the vehicle has at least one driver-assistance feature. While the driver still controls mobility, the vehicle is able to maintain its speed. For instance, models with lane-keeping technology or an adaptive cruise control or lane-keep technology are a Level 1 vehicle.
Level 2 – Partial automation
At this level, the vehicle has more advanced assistance systems for steering, acceleration, and braking- adaptive cruise control, automatic emergency braking, lane-keep assist. It is the coordination between these functions that is important. Moreover, the driver must monitor the weather, traffic and road conditions and be ready to intervene even though they are not physically operating the vehicle.
Level 3 – Conditional automation
At level 3, the vehicle can take full control and operate by itself under certain conditions. It may, however, request the driver to take over if it senses something it can’t handle. For instance, the vehicle may be able to drive by itself on an expressway but not in the city traffic. When operating conditions are ideal, the vehicle can take full control. The driver is still responsible for overseeing the entire procedure. This level needs the vehicle to have advanced sensor packages and sophisticated software for safety. In was in 2012 that Google achieved Level 3 autonomy but found that drivers were slow to retake control in the event of trouble.
Level 4 – High automation
This is the level where full autonomy exists. At level 4, the vehicle can complete a journey and perform safety-critical functions without driver intervention. However, some constraints exist – geographical limitations, speed limits etc. For situations that need human intervention, the vehicle maintains driver controls such as a steering wheel and pedals. Currently, no level 4 vehicles are available to consumers.
Level 5 – Full automation
At this level, the vehicle is self-driving and makes its own decisions under all circumstances, including extreme ones. No provisions for human control exist which means passengers no longer need to concern themselves with driving. A level 5 vehicle can theoretically safely travel at all speeds with advanced vehicle-to-vehicle and vehicle-to-environment communication. At present, it may be hard to imagine a world such vehicles become prevalent. How would they change we travel, park, live? We wouldn’t need to worry about traffic and parking problems. We may not even need to own a vehicle, after all. However, it is a long time before such vehicles become a reality.
Why this matters
The SAE-defined levels serve as general guidelines for how technologically advanced a vehicle is. They make the distinction between ‘autonomous’ and ‘driverless’ clear – driverless is a more advanced stage of autonomous. The rise of self-driving cars is expected to transform business and productivity, replacing corporate fleets for deliveries and saving workers time spent driving during daily commutes. According to a KPMG study, accidents will drop by 80% by 2040, completely changing the car insurance industry. A view that enjoys general consensus is that with each level, the car will become safer. It remains to be seen how autonomy in vehicles will change urban transportation and mobility.