Self Driving Cars

Introduction

Self Driving Cars (SDC) are cars that can navigate without the intervention of the driver using various technologies like computer vision and GPS navigation. Navigation could be semi-automated where it assists the car or fully autonomous where it can drive on its own without input from the driver.

Keywords: Self Driving Car, Tesla, Navigation

Key Technologies Enablers

The possibility of self-driving cars is possible due to various forms of technologies coming together. Some of the key technologies are as follows (Sharda. 2020):

Faster wireless internet: Each generation of wireless internet is exponentially faster than the last giving the ability cars to communicate with servers and each other.
Increased computer power in modern cars: Cars now have a lot more computing power and sensors as compared to older cars. The technological improvements in the car for safety including devices like adaptive cruise control, automatic emergency braking, blind spot monitoring, lane keep assistance, active driver check, integrated sensors, and cameras are a few of the technological advancements that have helped the development of cars that can drive autonomously.
GPS Navigation and Maps: The easier availability of GPS maps and navigation is not only aiding drivers to navigate to any desired location but is also a great help in autonomous cars to navigate the real world.
Computer vision and Deep Learning: With advancements in deep learning, autonomous vehicles can now distinguish objects with greater ease. When acting in a moving vehicle, it's important to know how to tell a person apart from an item like a tree or if the object is moving or stationary. Computer vision in cars can now distinguish traffic signals and other obstacles while driving in real-time.

Regulatory Issues with SDC

There are some regulatory hurdles that car manufacturers need to overcome in order to be operational in the market. There are irregularities between different states in the US in terms of government regulations that manufacturers need to get compliance with eg: NY does not allow any form of hands-free driving but other states like California, Arizona, Nevada, Florida, and Michigan allow the same (Sharda. 2020). There are also some unanswered questions that the governments need to get clarification on as the market is fast evolving like:

Changes in driving licenses to accommodate autonomous cars.
Questions about the requirement of DL when cars become fully autonomous in the future.
Self Driving Cars of the future will be able to carry young, old and people with disabilities autonomously without requiring any input from the “driver”. This would need a rethink of the licensing terms again.
Insurance companies need to also figure out the cause of accidents and the liabilities in case of accidents.
Considering the real-world implication and the safety of public life at the stack, watertight regulations need to be implemented as there will be a lot of new cheaper “me-too” companies that are going to enter the arena of driverless cars.

Latest in Autonomous Cars

Tesla is leading the EV as well as the Self Driving Car segments currently. The self-driving car has been more difficult to achieve than companies thought initially because it needed to solve Real Worl AI, which none have been able to solve. Some companies Uber and Lyft have given up on the idea and sold their self-driving program because of the difficulties with the unpredictabilities of Real World AI. Tesla is going the supercomputer approach for its computer vision, Amazon is partnering with American start-up Plus for automated lorries, Ford is teaming up with UK delivery company Hermes and China’s biggest search company Baidu is investing heavily in self-driving technologies (Ajitha. 2021).

There are currently five levels of development in reaching full automation and the current positions of the companies are as follows:

Driver assistance: Single assistance like cruise control or steering assistance.
Partial automation: The vehicle can perform steering and acceleration. Tesla’s autopilot and GM’s super cruise are at level two.
Conditional automation: Cars are able to have Environmental detection. Honda and Audi are here after decades of work.
High automation: All the driving tasks that a driver can do can be achieved but in a specified condition. NYVYA and Alphabet’s Waymo are reaching here.
Full automation: The vehicle is able to perform all the tasks that a normal driver can do under all conditions. No one is here yet, officially at least.

Advantage:

Efficiency: Traffic jams will become a thing of the past. There might not be a need for traffic lights even. Transportation like trucks would be able to operate nonstop.
Reduction in Accidents: Majority of the accidents are caused by human error including distracted driving and DUI which can be reduced drastically with a robot behind the wheel.
Accessibility: Anybody would be able to “drive” a fully automatic car. They just might need to have the ability to specify the destination.
Reduced car ownership: Car ownership might reduce as people would be able to order a car as per needs right to their doorsteps.
Comfort: Long journeys would be much safer and more comfortable for the occupants.

Disadvantage:

A very small margin of error: SDCs are allowed a very small margin for error as the consequence of any error can be fatal and a question of public safety. It has to be 99.9999% safe to operate and get harder with each step of improvement.
Political pressure: Automation can have a very negative impact on the job market, especially in the Taxi and Trucking business, and hence might be faced with resistance.

Impact on the job market

The biggest impact of driverless vehicles is going to be felt in the Taxi and Trucking industries. Of this sadly the trucking industry is going to have the biggest job loss due to the replacement of truckers with computers due to the following reasons (Gupta. 2021):

24/7 operation: Automated trucks are going to be able to operate without halts and breaks, unlike the drivers who are required by law to have maximum driving hours monitors and frequent halts on long journeys.
Labor cost: Automated trucks are going to be cheaper to maintain compared to maintaining drivers and also companies are going to be eying on saving on worker benefits like medical, dental, and leaves.

Another impact might be the fact that owning a car might become obsolete as people might be able to order a car from a lot nearby as the need arises and the car can drive up to the person's doorstep.

Discussion

It is just a matter of time before the ability to drive will become an art or hobby just like calligraphy/writing being replaced by typing or horse-drawn cars with automobiles. The older classic car would be retrofitted with EV motors and autonomous driving modules. Car ownership will be optional and cars will become self-aware and will communicate with other cars on the road. Traffic lights will not be even required as all cars can coordinate in the intersections more efficiently. Although there are pros and cons to the new changes, the advantage outweighs the disadvantages and the term “auto-mobile” will come to live up to its namesake at last.

References

Ajitha, P. V., & Nagra, A. (2021). An Overview of Artificial Intelligence in Automobile Industry–A Case Study on Tesla Cars. Solid State Technology, 64(2), 503-512.

Sharda R. Delen D. & Turban E. (2020). Analytics data science & artificial intelligence: systems for decision support (11th ed.). Pearson Education.

Gupta, A., Anpalagan, A., Guan, L., & Khwaja, A. S. (2021). Deep learning for object detection and scene perception in self-driving cars: Survey, challenges, and open issues. Array, 10, 100057.

A General Discussion on Self Driving Cars (SDC)

Table of contents