Self-driving cars are here, and the some of the popular ones sports the Tesla badge. However, in the recent past, there have been multiple instances where self-driving Tesla cars have crashed while the drivers were reported to have been sleeping in the passenger seat or engaged in a conversation in the rear seat. While the victims might argue that Tesla advertises its self-driving feature as ‘autopilot’, however, it is the lack of awareness among some motorists which misleads them to belief that, similar to aircrafts, even the Teslas can drive themselves entirely without human intervention, and that is where things begin to go wrong. In this article, we hope to put those myths to rest.

Many of us often confuse the term 'autonomous' and 'self-driving' to be one and the same thing, but in reality, these are very different. Self-driving cars merely comprise those vehicles that can drive themselves but need commands from an external source, in this scenario, the source being humans. Whereas a fully autonomous car would be self-aware and capable of making its own decisions. So, a simple navigational command ‘to the workplace’ would be obeyed flawlessly by the self-driving car, but an autonomous vehicle might take you elsewhere.

The Tesla line-up comprise self-driving vehicles and not autonomous machines, regardless of the ‘autopilot’ term being used by the carmaker to describe its self-driving functionality. And not only just Tesla, but most carmakers who claim that their cars feature the autopilot mechanism, are mostly self-driving cars with Level 2 autonomy. Only those cars that possess self-awareness and can be operated without any human intervention can constitute autonomous cars with Level 5 autonomy. Although, before we jump into this classification, let us understand how a car manages to drive itself.

How do they work?
Unlike the conventional four-wheeler which is a mechanical maze, the self-driving car is brimmed to the gills with electronic devices and software. A host of cameras, sensors and actuators also get embedded to execute the software’s commands. If the software and powerful processors make up the self-driving car’s skeleton, the car’s exterior metal consists of cameras and sensors placed strategically all around it. 
While the cameras aid in detecting traffic signs, speed limits and pedestrians, the radar sensors help in monitoring the position of other moving vehicles in the vicinity. Lidar sensors can further measure distances and identify lane markings, whereas ultrasonic sensors come in handy while parking by detecting curbs and other vehicles. This array of cameras and sensors help create a virtual environment of the car’s surroundings. Using this data, the self-driving car’s software-driven brain plots a course to the desired destination while considering the predicted obstacles detected via sensors and further instructs the actuators to operate the vehicle’s acceleration, brakes and steering.

Classification of autonomous levels
The general confusion between self-driving and autonomous cars would tell you words don’t work best when segregating technology. Hence, the Society of Automotive Engineers (SAE) decided to set up some commonly agreed-upon standards or bifurcation levels to define and distinguish between the varying degrees of automation in cars.

 The bandwidth for this classification ranges from cars with no automation to cars with full automation. Ranging from Level 0 to 5, there are six levels in the automated driving spectrum. The base Level 0 comprises cars such as the Maruti Alto 800 or the Hyundai i10 Nios with no driver-assistance technology integration and requires full human intervention. Level 1 autonomy comprises cars where either the pedals or the steering wheel can be controlled by the car’s computer for a limited time and yet requires full intervention from the driver. Early-edition Volkswagen Polos with cruise control comprise this category. Cars equipped with Level 2 autonomy carry advanced driver assistance systems where the vehicle can take over the operation of both the pedals and the steering wheel for a limited time though human intervention is still required and the driver must be prepared to take over the control of the car when necessary. The recently launched MG Astor and the Mahindra XUV700 come equipped with ADAS, a suite of driver-assistance features consisting of Adaptive Cruise Control, Lane Keep Assist and Auto Emergency Braking, and fall under Level 2 autonomy.

 Human attentiveness was required at all times in cars with up to Level 2 autonomy. However, the requirement disproportionately decreases as we go onto Level 3 and beyond. With Level 3 autonomy, the vehicle’s software-driven brain achieves partial automation and allows the driver to cede all car controls. The driver can remain relatively more relaxed compared the previous levels though they will still be required to take control of the vehicle in situations where it gets overwhelming for the car onboard computers - an usual phenomenon on the streets of India. The major difference between Level 2 and Level 3 will be a larger time provided in Level 3 autonomy for the transition from automated operation to manual takeover, making it a tad bit more comfortable for the driver.
 Level 4 autonomy allows the driver to remain completely relaxed while the vehicle is in motion. Cars with level 4 autonomous features will be capable to tackle any road traffic situation by itself and without prompting the driver for aid at any point. So the part where you imagined taking a nap in the driver’s seat is possible with Level 4 autonomy cars and beyond. And finally, Level 5 autonomy comprises vehicles one usually sees in science-fiction movies with space for just passengers and no dedicated driver controls such as pedals or steering wheel present inside the car. But if one comes back to the worded differentiation as discussed above, cars with Level 4 autonomy or below would find themselves placed under the ‘self-driving car’ category, whereas the ones equipped with Level 5 autonomy can be described as fully ‘autonomous cars’.

Current situation of autonomous cars in India and globally
India’s mass-market cars have been a recipient of Level 1 autonomous features for over half a decade now. Cruise control, a level 1 feature that keeps the car’s acceleration steady at a set speed, was equipped in the majority of the premium segment offerings consists of the likes of Maruti Suzuki Swift, Hyundai Creta and Skoda Octavia just to name a few. In October 2021, MG Motor launched the Astor compact SUV, equipped with Level 2 autonomous features such as Forward Collision Warning, Adaptive Cruise Control, Lane Keep Assist, and many more. These driver assistance features form the ADAS or Advanced Driver Assistance Systems when clubbed together.

 Mahindra’s all-new XUV700 became the second car recently to feature ADAS. Following up on the advent of ADAS in Indian cars, Mahindra’s Chief of Global Product Development, R Velusamy, recently commented, “We anticipate ADAS features will be offered by all the major mass market players, at least in their flagship offering, in the near future. In India, Level 2 autonomy will be widespread in the next two to four years”. Substantiating on the same, reports of Hyundai introducing ADAS in its product line-up before the end of 2022 have surfaced online. The upcoming new Tucson is anticipated to be Hyundai’s first model to get the suite of driver assistance features.

 Among India’s massive pool of technological startups, a handful of companies have made significant headway in integrating autonomous technology in a different, more Indian-centric way. Minus Zero is a startup based in Punjab which claims to be the world’s first for building Level 5 autonomy for unstructured Indian roads. Its goal is to achieve Level 5 autonomy by 2023 and partner with an OEM to feature the tech in mass-market cars. Flux Auto is another Bangalore-based startup working on developing autonomous tech for Indian highways. Most importantly, they aim to provide such technology that can be retrofitted into existing trucks to make them driverless vehicles in the future.
 
In the global scheme of things, Mercedes-Benz is set to offer Level 3 autonomous capabilities in its flagship S-Class sedan by mid-2022 but operational within a certain geofence within Germany and under a certain speed limit. Polestar, an all-electric performance brand from Sweden, has claimed to equip the upcoming Polestar 3 electric SUV with Level 3 technology to be unveiled this year. Even BMW is prepared to join the Level 3 bandwagon with the 2022 MY 7-Series flagship sedan slated for a North America release by the second half of this year. American brands General Motors and Tesla are also looking to offer a similar suite of autonomous technology by early next year. Also, because self-driving cars are software-driven, technology tech giants like Google and Apple are trying to foray into the auto industry. The former is already conducting real-world tests with Waymo, Google’s self-driving project.

What does the future hold?
Developed countries such as the USA and the UK have taken massive strides in self-driving car technology. Testing of self-driving and autonomous cars is currently underway on streets of California. In contrast, India has just touched the tip of the autonomous ice-berg with the recent launches of Level 2 autonomous cars. The subcontinent is yet to witness autonomous technology in its purest form. Surprisingly, the Indian government has been disinterested in the past in allowing Tesla to set up shop on our shores. One of the major reasons for the same could be the loss of human resources, which is found in abundance in India. According to a recent statement by Nitin Gadkari, Union Minister for Road Transport and Highways of India, 22 lakh drivers are waiting to get themselves a job and cars with such human-eliminating technologies would further hamper the prospects of employment generation in India.
 Though the past is proof that technology waits for no one, and sooner or later, the masses will adopt the same. With India’s chaotic traffic conditions in the city, one should expect to see the initial application of autonomous technology in highway transport with autonomous trucks or the agriculture sector with autonomous tractors. Flawless execution of the same could lead the city traffic to be governed by autonomous friendly vehicles, but in all fairness that is still about a decade away in developed countries. In India, maybe few decades.

India has to have substantial preparations in place before the masses adopt autonomous technology. The most basic issue to resolve is a better condition of roads. Potholes and undulations on the road might influence the car’s computers to take sudden evasive actions, which might cause accidents. This also includes lane markings being made more clear and visible for the car’s onboard sensors. Moreover, the market feasibility for such technology is still on the lower side. Lidar sensors and cameras are expensive equipment, and India’s cost-sensitive crowd might not see value in them. 
But the most important aspect is accidents with self-driving cars involved. Whom or rather what is to blame in situations like these - the pedestrian, the driver or the manufacturer? Accident accountability raises a big question on autonomous cars, and every angle of the same has to be legislated before widespread adoption of the same. And the final cause of concern is us, or rather our willingness to trust the computers. Driving a car is an entirely different experience than any other activity replaced by a machine. Driving requires all limbs of the body and the constant attention of the mind, which would require a plethora of sensors for a machine to replicate. Moreover, there’s always the question of the computer’s ability to collate data from all the sensors and arrive at a decision, basically duplicating the human nervous system within a split second. The only workaround here is smart development of autonomous technology, repeated testing in all possible conditions and maintaining an environment suitable enough for these autonomous cars to work flawlessly. And that summarises the science-fiction dream we are working towards next.