Self Driving Car Project For Computer Science

Self Driving Cars

 

INTRODUCTION

The usage and production of these cars have become a leading industry in almost every area of the world. Over the years and centuries, this industry has gone through enormous development, as the first vehicles were only powered by the steam engine, then petrol and diesel came to the public mind, and currently, it seems that the electric propulsion will be the future. Of course, with this development, faster and more useful vehicles can be produced, but in our accelerated world with more and more cars, unfortunately, the number of accidents has increased.

In most cases, these accidents are the fault of the driver, therefore it could be theoretically replaceable with the help of self-propelled cars. The human presence is the most important part of transport at present, although there are many areas where you can use a tool or feature that helps people achieve greater efficiency. Some examples of these features are the autopilot on aircraft, cruise control in cars, and many other tools that help decision-making.

 

EVOLUTION OF SELF-DRIVING CARS

Autonomous cars are those vehicles that are driven by digital technologies without any human intervention. They are capable of driving and navigating themselves on the roads by sensing the environmental impacts. Their appearance is designed to occupy less space on the road to avoid traffic jams and reduce the likelihood of accidents.

The dream of self-propelled cars goes back to the Middle Ages, centuries before the invention of the car. A piece of evidence for this statement comes from sketches of Leonardo De Vinci, in which he made a rough plan of them. Later, in literature and several science fiction novels, the robots and the vehicles controlled by them appeared. The first driverless cars were prototyped in the 1920s, but they looked different than they are today. Although the "driver" was nominally lacking, these vehicles relied heavily on specific external inputs. One of these solutions is when the car is controlled by another car behind it. Its prototype was introduced in New York and Milwaukee is known as, the American Wonder" or "Phantom Auto".

Most of the big names – Mercedes Benz, Audi, BMW, Tesla, Hyundai, etc. – have begun developing or forming partnerships around autonomous technology. They invested sizable resources into this, and by making this step they wanted to be leaders at the market of self- driving cars.

Up to this point, numerous aids, software, and sensors have been put into these cars, but we are still far from full autonomy.

They use lasers that are testing the environment with the help of LIDAR (Light Detection and Ranging). This optical technology senses the shape and movement of objects around the car; combined with the digital GPS map of the area, they detect white and yellow lines on the road, as well as all standing and moving objects on their perimeter. Autonomous vehicles can only drive themselves if the human driver can take over control if needed.

 

These are those features that driverless cars already use:

•Collision avoidance

•Drifting warning

•Blind-spot detectors

•Enhanced cruise control

•Self-parking

 

Below we briefly present some companies that play the most important role in the innovation of this segment, to show how this industry has developed.

 

Tesla 

Elon Musk, the Chief Executive Officer of Tesla, claims that every Tesla car will be completely autonomous within two years. Tesla's "S" model is a semi-self-propelled car, where different cars can learn from each other while working together. The signals processed by the sensors are sent to other cars thus they can develop each other. This information teaches cars about changing lanes and detecting obstacles, and are continually improving from day today. From October 2016, all Tesla vehicles have been being built by Autopilot Hardware 2, with a sensor and computing package that the company claims to allow complete self-driving without human interference.

Google 

The Google team has been working on driverless cars for years, and last year a working prototype was presented (by them). Furthermore, Google also supports other car manufacturers with self-driving car technologies such as Toyota Prius, Audi TT, and Lexus RX450h. Their autonomous vehicle uses Bosch sensors and other equipment manufactured by LG and Continental companies. In 2014, Google planned a driverless car that would be available without pedals and wheels to make it available to the general public by 2020, but according to the current trends, its fulfillment is still unlikely.

 

nuTonomy 

A small group of graduates of the Massachusetts Institute of Technology (MIT) created the nuTonomy software and algorithm, especially to self-propelled cars. In Singapore, nuTonomy has already put sensors to the Mitsubishi i-MiEV electric car prototype, thus nuTonomy algorithms can control the car on these complex urban roads by using GPS and LiDAR sensors. Besides that, in November 2016, they announced that self-propelled cars will be tested in Boston as well.

The National Highway Traffic Safety Administration (NHTSA) adopted the levels of the Society of Automotive Engineers for automated driving systems, which provides a broad spectrum of total human participation to total autonomy. NHTSA expects automobile manufacturers to classify each vehicle in the coming years using SAE 0 to 5 levels.

These are the levels of SAE: 

Level 0: No Automation

In this case, there is 100% of human presence. Acceleration, braking, and steering are constantly controlled by a human driver, even if they support warning sounds or safety intervention systems. This level also includes automated emergency braking.

 

Level 1: Driver Assistance 

The computer never controls steering and accelerating or braking simultaneously. In certain driving modes, the car can take control of the steering wheel or pedals. The best examples for the first level are adaptive cruise control and parking assistance.

 

Level 2: Partial Automation 

The driver can take his hands off the steering wheel. At this level, there are set-up options in which the car can control both pedals and the steering wheel at the same time, but only under certain circumstances. During this time the driver has to pay attention and if it is necessary, intervene. This is what Tesla Autopilot has known since 2014.

 

Level 3: Conditional Automation 

It approaches full autonomy, but this is dangerous in terms of liability, so therefore, paying attention to them is a very important element. Here the car has a certain model that can take full responsibility for driving in certain circumstances, but the driver must take the control back when the system asks. At this level, the car can decide when to change lanes and how to respond to dynamic events on the road and it uses the human driver as a backup system.

 

Level 4: High Automation 

It is similar to the previous level, but it is much safer. The vehicle can drive itself under suitable circumstances, and it does not need human intervention. If the car meets something that it cannot handle, it will ask for human help, but it will not endanger passengers if there is no human response. These cars are close to the fully self-driving car.

 

Level 5: Full Automation 

At this level, as the car drives itself, human presence is not a necessity, only an opportunity. The front seats can turn backward so passengers can talk more easily with each other because the car does not need help in driving. All driving tasks are performed by the computer on any road under any circumstances, whether there's a human on board or not.

These levels are very useful as with these we can keep track of what happens when we move from human-driven cars to fully automated ones. This transition will have enormous consequences for our lives, our work, and our future travels. As autonomous driving options are widespread, the most advanced detection, vision, and control technologies allow cars to detect and monitor all objects around the car, relying on real-time object measurements.

Besides, the information technology built into the vehicle is fully capable of delivering both external (field) and internal (machine) information to the car

 

DECISIONS

Self-driving cars may be the future of transportation but we do not know whether it is safer than non- autonomous driving or not. There are unexpected events during driving that force us to decide, often these are only tiny things such as passing through the yellow light or not but sometimes situations arise where we have to decide on the lives of others or our own.

Trusting in new technologies is expected to be a significant challenge for the public. Few people feel comfortable about using a new and unproven transportation technology, which can be seen after studying aviation history.

 

These problems may arise:

• How should the car be programmed to act in the event of an unavoidable accident?
• Should it minimize the loss of life even if it means sacrificing the occupants, or should it protect the occupants at all costs?
• Should it choose between these extremes at random?

 

Answers to these ethical questions are important because they can have a great impact on the ability to accept autonomous cars in society. Who would buy a car that is programmed to sacrifice the owner?

 

Careful about:

1. An Unregulated Industry

Because information about the technology is limited, and although 200 car companies are jumping into the self-driving car space, there are not enough solid facts to create a baseline for safety standards. As yet, the industry is unregulated which is excellent for manufacturers but bad for consumers.

2. More Accidents Blending Self-Driving and Manual Cars

Sometimes self-driving cars give the passengers a sense of false security when really, they should be extra cautious and ready to take the wheel at any given moment should the need arise.

3. Vulnerability to Hacking & Remote Control

Any computer device connected to the internet is vulnerable to hacking. These cars also rely heavily on the software that runs their components, and if a hacker gets into the system, they can control every aspect of the car.

Other dangers to be aware of are the theft of private data and even gaining remote access to a cell phone connected to the car via Bluetooth. Self-driving vehicles may also be more susceptible to computer viruses.

4. Computer Malfunctions

Most self-driving cars are made up of not one but 30 to 100 computers. That is a lot of technology where things could go wrong. The software that runs self-driving cars is admittedly sophisticated. However, one of the more difficult challenges that engineers struggle to solve is how to operate smoothly in all weather conditions. Correctly controlling sensors on the rear camera is also an issue. A particularly dangerous glitch is how to know when to execute a quick stop when someone is in the crosswalk in front of the car. Other concerns that should be solved before these cars hit the road are freeze-ups during autopilot mode, and how to account for the unpredictable behavior of other motorists. 

5. Exposure to Radiation

With all the goodies on board like GPS, remote controls, power accessories, Bluetooth, Wi-Fi, music, and radio components drivers will be increasingly exposed to higher levels of electromagnetic field radiation. Exposure to electronic radiation can cause a myriad of serious health problems. Some of the more serious issues are high blood pressure, difficulty breathing, migraine headaches, eye issues, exhaustion, and sleeplessness.

 

 

CONCLUSIONS

This is a quite new topic, and this is still closer to a piece of science fiction literature, but several companies try to solve this task, even if there are many problems with it. We showed that the main problem is the fear of losing control. If a computer decides instead of us, we do not control the processes. Every computer and program may have a back door, and the question arising is what can be done if someone enters into the computer that can save our lives.