There were days when cars were mechanical wonders. Car driving was an experience of its own. The driver has to feel the car to drive it to its full potential. A connection has to be established where the car and the driver became a single body. This was the beauty of driving a few decades ago. Cars started their journey as a machine but as we step towards the future, computers have started taking over a big part of our tasks and this is true in terms of driving a car as well.
Now the cars come with multiple computers and controllers, sensors, codes and onboard diagnostics that monitor each and every aspect of the car. Cars have become more of a gadget now. With so many assisting and connectivity systems it has taken away the pleasure of sheer driving a machine.
Are Self Driving Cars Safer than Cars Driven by Humans?
As you are reading this, automakers and technology companies are working together to make self-driving cars for the future. The day is not far when we will see cars ferrying on roads without any one of the driver’s seat. This has already seen in aircraft when autopilot mode is engaged they can fly themselves on a predetermined path. Though the difference lies in the chances of facing an obstruction or traffic is much higher when traveling on land than being in the air. Autopilot software already has millions of lines of codes and in case of a car to drive by itself, it needs a much smarter system with even more complex codes.
Things that to be taken care of when the car will be self-driving on roads
- Understanding the surroundings and path.
- Monitoring speed of other vehicles, their shapes and sizes, their direction of movement.
- Identification of pedestrians and their activities whether they are stationary or going to move on the pedestrian lane. They might be crossing the road as well or maybe plainly standing in the middle.
- Understanding road signs, signals and understand what is a roadsign and what is not.
- Monitoring speed, where it is going, the internal systems of the car and the signals it may be receiving from other cars around it.
These are a few very basic things to be taken care of. There are many unknown variables that are needed to make self driven cars safe to be on the road. Complication arises when all these conditions are needed to be framed in codes because that is how the computers understand and react accordingly. These are further discussed later in the article.
One major point of debate is the ability to make decisions like a human brain in cases of unknown scenarios. Though this is true humans are not perfect but still they can take alternate decisions not like machines that are programmed to work on a set of predecided rules.
The biggest question rising is that can we trust the passenger’s lives in the virtual hands of the technology that will drive our cars in the near future. It is extremely important to learn about the safety and security concerns related to self-driving cars. First, let us take a look at road crash figures. Then we can understand how safe are these autonomous cars when compared to the ones that are human-driven. Let us find out whether self-driving cars safer than cars driven by humans or not.
How safe are human-driven cars?
Experts say that self-driven cars are going to safer than human-driven cars. Records indicate that each year 20-50 million human drivers get into accidents with around 1.35 million are fatal. This is about 3700 every day. These figures of people losing lives in road accidents, each year questions the driving abilities of humans. This number is a lot but statistics show that these are diminishing each year. For instance, in the U.S a small decline in deaths on roads has been observed. In 2019 it was estimated that 38800 people died in road accidents which are 2% lower than that of 2018 at 39404 deaths and a 4% decline than that of 2017 at 40231 deaths. A seriously high number of people at 4.4 million were injured who required serious medical attention in road crashes but still, this is a 2% decline over 2018 statistics.
Risk factors (Human-driven car accident statistics)
- On average every 1% increase in mean speed relates to a 4% increase in the fatal crash risk and a 3% increase in the serious crash risk.
- Pedestrians’ death risk to be hit by car fronts increases sharply by 4.5 times when the speed goes up from 50 km/h to 65 km/h.
- At a speed of just 65 km/h the car-to-car side impacts, the fatality risk for car occupants is 85%.
Driving under the influence of alcohol and other psychoactive substances
- Driving under influence of alcohol/psychoactive substances or drugs raises the hazard of a serious or fatal crash.
Nonuse of motorcycle helmets, seat-belts, and child restraints
- A 42% decrease in the risk of fatal injuries and a 69% decrease in the risk of head injuries is observed when using a standardized helmet in the right way.
- The use of seat-belt lessens the risk of death among front-seat occupants by 45 – 50%. The risk of death/serious injuries among rear seat occupants reduces by 25% when wearing seatbelts.
- Child restraints can prevent deaths by 60%.
- The use of mobile phones is 4 times more likely to be causing an accident. Using a phone when driving slows reaction times making it difficult to keep in the appropriate lane, and maintaining the correct following distances..
Unsafe road infrastructure
Ideally, roads should be designed with adequate facilities for pedestrians, cyclists, and motorcyclists. Traffic calming measures like footpaths, cycling lanes, safe crossing points, etc reduce the risk of injury among the road users.
Safe vehicles are crucial in averting crashes and reducing serious injuries. There are regulations and standards on vehicle safety that are to be met in order to manufacture and sell cars in various countries. These may differ from country to country but if there are many global standards that maintain a basic level of safety of vehicles around the world. Tests like front and side-impact, the inclusion of electronic stability control, anti braking system, and ensuring that airbags and seat-belts are fitted in all vehicles.
Inadequate post-crash care
Delay in detection, first aid and then required medical attention for those involved in a road crash may add to the severity of injuries. This is very time sensitive as delays of minutes can cost someone’s life. Better post-crash care ensures timely prehospital care and acceptable quality of both prehospital and hospital care.
Inadequate law enforcement of traffic laws
Traffic laws on drink-driving, seat-belt wearing, speed limits, helmets, and child restraints when strictly enforced, bring about the projected fall in road traffic fatalities and injuries related to specific behaviors.
How do self-driving cars prevent accidents?
The reasons listed affect the possibility of a road accident. Well as we know in the case of computer-aided driving most of these human errors can be eliminated making cars on road less prone to such accidents. All these parameters are fed to the car’s computer and control units in the form of codes so that they can prevent accidents in such conditions. Though driving on roads is much more varied than just these conditions. Self-driving cars are equipped with many sensors and controllers that are calibrated to observe and act on many variables that are invisible to a skilled human driver who relies on his own senses and assumptions combined.
These systems are continuously tested to make them better suitable for real-world driving. We are still unclear whether these autonomous vehicles are ready to be used on roads or they are still to be supervised actively by humans. The autonomously is in its nascent stage learning, growing and evolving every day as the tests collect results. They adapt accordingly to get better day by day. One day it will definitely be better and safer than humans driving their vehicles but till then there is a long way to go.
In March 2018, in Arizona, there was a fatal accident involving an Uber self-driving test with a biker in Arizona. The self-driving car botched to identify the biker as a human as well as the human safety driver did not notice the biker hence the crash could not be avoided.
But the good this is that vehicles driven by humans are using such technology in Advanced Driver Assistance Systems (ADAS). These aid in driving the car and using it at its full potential as well as making them less prone to human errors as these systems override any delusion by humans and correct them on their own unless and until it has already been fed in the codes.
Today’s Advanced Driver Assistance Systems (ADAS), How ADAS works?
Almost all new cars come equipped with ADAS now, which includes:
- Lane tracking
- Autonomous emergency braking
- Enhanced vision systems
These systems rely on the combined performance of sensors and actuators that communicate over local networks, which are in turn controlled by microcontrollers.
The cars are connected with each other sharing their learning and updates collectively improving all the vehicles in the network via a communication system known as Vehicle to Vehicle Communication (V2V). Another type of communication network is the Vehicle to Infrastructure (V2I) or V2X, where the vehicles also communicate to the infrastructure — such as traffic lights, road signs, or satellites.
Security Concerns with Self-Driving Cars
As earlier mentioned there are numerous security concerns related to these systems.
As V2X communication becomes more prevalent in cars, they are more susceptible to malicious attacks. There have been incidents where hackers take control of cars, overriding the driver.
Almost all automakers use On-Board Diagnostics (OBD), which gives control over various engine parameters for fault-finding and diagnostics at the time of service.
It is possible that OBD’s are publicly available which can easily enable anyone to access engine parameters using a smartphone with Bluetooth connectivity.
Automotive Regulations and Standards
With the time, we will see the ADAS grow and self-driving cars turn out to be the norm, then the automotive compliances are of utmost importance. At all stages of this development, automotive standards will necessary to follow.
ISO 26262 is one of the fundamental safety standards for vehicles focusing on the functional safety of electrical and electronic systems. It pertains to all actions within the lifecycle of safety-related systems including requirements that are relevant to the quality of software.
The measure of such risks is provided by the use of Automotive Safety Integrity Levels (ASILs) ranging from A to D where A is the lowest and D the highest safety integrity level with max requirements. These risk assessing parameters include the severity of the risk, probability of exposure, and controllability.
Controllability presupposes that the driver:
- Is in a fitting condition to drive.
- Has the apt driver training and/or driver’s license.
- Act in accordance with all related legal regulations.
Are Self-Driving Cars Legal?
Laws will call for adaptations to put up with ADAS and self-driving cars. It’s significant that legislation accounts for the post actions in the case of an ADAS failure. ADAS in made to notify drivers and control falls back to humans.
Software Design Standards
If the human driver may not be attentive and when ADAS fails then they will be not able to avoid harm.
The Society of Automotive Engineers (SAE) standard J3016 diversifies driving automation into six classes, from no automation to fully automatic.
ADAS at SAE level three or higher rely on software to:
- Collect data from sensors.
- Create a model of the environment.
- Settle on how to assist the driver or control the vehicle.
ADAS at these levels also determine whether sensors are functioning correctly, when to alert the driver, and when to trigger a fallback to human control.
Traffic laws will change to contain ADAS chiefly in the area of liability and privacy as every country has its own traffic laws.
In future cars are going to be much more reliant on technology. Constant learning and updation of these hardwares and softwares will at some point make them capable of being autonomous and finally being self-driven. Until then humans will be needed to monitor autonomous driving.