Autonomous cars, though appearing to be a reality now, may still take much longer before these routinely drive themselves on the streets. It is a pleasure to consider a future filled with self-driving cars along with a world of safe commuting where robotic car navigators have made deadly crashes a thing of the past. An electronics researcher for Volkswagen at a worldwide conference on driverless cars held last year, pointed out (as an example) that even tumbleweed (it’s a plant of dry regions which falls apart near the ground during the late summer, forming a light globular mass which is tumbled about by the wind) can act a major obstacle.
Apart from the tumbleweed being a challenge, technology of the driverless cars can face more serious obstacles and I picked out some of those and listed them as follows:
Mapping for the self-driving cars
The creation and maintenance of maps for the self-driving cars are complicated to work upon. Two crucial systems that should work in a sequence are basically a combo of detailed & pre-made maps as well as sensors that observe obstacles in real time on the road.
But even this kind of a mapping system will put forward some major challenges. Like maps only for a limited number of areas whereas going mainstream requires mapping of millions of miles. Another major issue arises with mapping the constant changes of the roads.
To solve this issue, a company like Tesla is aspiring to build self-driving cars that prominently depend on imaging & sensor processing.
- You may also like : 6 reasons why driverless cars keep a transformative potential
Complexities of social interactions
The ability to drive requires several intricate social interactions which for now are definitely tough for robots. It’s far more intensely a social process involving complex interactions with others. Therefore, human drivers rely on the generalized intelligence & common sense that robots still lack.
But interestingly, autonomous cars like that by Google are trained on software to recognize some extent of hand gestures. But situations involving eye contact, subtle communication, and cop hand waving are all figured by the judgment and intelligence of a human but still a computer cannot make that call.
Fully autonomous cars and not partial ones will need to be capable of understanding the environment around. Also, understand why the people around are behaving in the way they are along with making the decision of how to respond.
Weather conditions make it trickier
Compounding to all other challenges, weather still poses as a major hurdle for the autonomous or self-driving cars. Alike our eyes, car sensors too don’t seem to work well in the complex weather conditions. On a sunny weather day, everything tends to work but do we have everything sunny every time?
It is a real hurdle but the difficulty isn’t that fundamental yet there is a doubt as to how long this challenge would take to overcome. Though weather conditions might not derail the technology and will come out first in those places where the weather is good and where the roads prove easy for driving.
Drivers would need to be disciplined & skilled
Drivers will need to be more disciplined as far as rules are concerned before the technology of self-driving cars become a routine sight. Also, if humans need to take over when the software isn’t coping, drivers would need to be suitably skilled.
This means that there are autonomous vehicles being built that have the freedom to switch between software and driver control.
Over-reaction towards obstacles and / or ignorance of real ones
Technology of the self-driving cars can easily spot obstacles ahead but may find it harder to tell apart the harmless and dangerous ones. The technology may also appear to overreact at times. Like some cars, while making the test drives made emergency stops even in front of the metallic chocolate bar wrappers.
The technology may ignore the socially acceptable incidents also like how a human driver wouldn’t stop to keep away from hitting a dog but might stop for a cow.
What about exceptions?
A valid point comes out in this discussion. At many times, traffic police officials wave the car drivers through red or green lights to stop them. I cannot understand how would a driverless car cope with that situation and what would be the following action?
- You may also like : No autonomous cars for India and Brazil
Technical & Cybersecurity issues
Providing a software sufficient information about the world around may become an issue for gathering data at a massive level. Apart from this, cyber security is also likely to be an issue. Though, it would be an attainable one.
How do we make sure that these cars aren’t hacked? There are infinite techniques to get into the software of these smarter & more connected robotic cars and disrupt what they are doing. Securing this system could be extremely difficult and is an important decision to consider.
Self-driving cars are unquestionably becoming the part of our future lives. Perhaps, not all at once, not too soon and not too conventionally.