Autonomous Vehicles in Public Transportation

Abstract—The purpose of this research is investigating the usage of Autonomous Vehicles (AVs) in public transportation and proposing a new system using AVs and data sources to increase efficiency in public transportation provided by busses and taxis.
This paper states the parameters of AVs in public transportation and discusses how it will result in an efficient way of managing the public transportation according to collected data such as waiting passenger count, current traffic density, timetables etc. The words “autonomous” and “connected” can be used in this case, because the system that paper suggests, uses dynamic data also from current view (via multiple sensors) of the vehicle and other sources from the internet as it can come from other connected vehicles etc.

Index Terms—Autonomous vehicle (AV), traffic optimization, smart city, public transportation.


During the last years, the autonomous driving subject seems more popular than ever, as the technology improves and more companies announcing that the works have been started about this subject, it has become a hot topic. By 2020 companies like Audi, GM, Daimler, Mercedes-Benz, Nissan, BMW, and Renault expect to sell vehicles that are close to fully-autonomous driving. By 2035 it is predicted that most cars would be fully autonomous. [2] It has been a while for the semi-autonomous features that have been used in vehicles such as automatic braking, lane keeping etc. but the future of AVs seems more complicated. While the AVs world has some developments and works, there is no doubt that the autonomous vehicles issue would not affect public transportation. After widely usage of AVs in traffic, it will be impossible to consider using them in public transportation.

This research will be consisted of three main aspects of AVs in public transportation which are;
– Technology
– Optimization
– Safety

In Technology part, the features of AVs which provided by latest improvements in autonomous driving can be useful for public transportation and how they will be used in a matter will be discussed.

While debating the Optimization issues, beyond the general problems of AVs, it is especially will be investigated for public transportation and how it will affect the efficiency of managing it.

Lastly in Safety part, it will be discussed that if we can trust a vehicle that carries many passengers without a driver and what are the cases that should be considered.


Autonomous driving is one of the hottest topics in technology world in the last decade. Many companies work on new technologies and techniques to improve the efficiency and usage of the AVs.
The main concept is, an autonomous vehicle should operate without any human intervention. That appears new tasks to a vehicle, such as;

  • Vehicle should sense its environment with multiple sensors and identify appropriate behaviors according to them. From sensing cars, pedestrians, road objects to weather conditions, many input should be detected by the vehicle.
  • It should follow a navigation path. In the case of public transportation and proposed system, the navigation path can have multiple stops and instantaneous changes based on the information that is got from other data resources.
  • The vehicle should follow the traffic rules.
  • Unexpected situations should be handled. An odd behavior from a random pedestrian in the street or an accident should be recognized and the most suitable behavior of the vehicle should be applied.

This technology can be used in public transportation with a management system using different data sources and resulting an efficient system. These data can be used;

  1. Traffic Density. Bus routes are pre-determined and changes rarely. With the live traffic density data, bus route can be changed for maximizing the efficiency. In case of an accident which is on the bus route, or taxi’s route to destination, new calculated road can be a better solution to avoid the traffic jam and it definitely saves time.
  2. Passenger Density in Vehicles. This data can be gathered by using an observing technology in the vehicles. It will be used for distributing passenger numbers uniformly among the vehicles and lines.
  3. Bus Stop (Waiting) Passenger Density. This data can be gathered by using an observing technology at the bus/taxi stops. Measuring the crowd in the stops, will give data about demands for the numbers of vehicles that passes from the related stop.
  4. User Requests by a Mobile App. Waiting Passenger Destiny would not be useful when it is used solely, there can be many bus routes that passes from the same stop so destiny does not indicate how many passengers are waiting for a specific bus. Via a mobile application, bus line requests can be taken from the potential passengers.
  5. Vehicle Number and Timetable Data. For managing an efficient autonomous public transportation system, beyond the data that is collected from the sources must be coalesced with current numbers of vehicles and timetable data.

All these data that gathered and merged together, can be converted into a significant data which will be used in managing and optimization of an efficient public transportation system with autonomous vehicles.
In the system, buses and taxis are considered for the research and in the Figure 1, there are some differences mentioned in related vehicle types.


While comparing different type of public transportation AVs, there appears different properties. These properties are stated and compared as follows:

  1. Route. Bus routes are pre-determined and changes rarely. This provides an information about the behavior of AVs. An autonomous vehicle used for a taxi should be prepared for any destination that requested from the passenger and process it. In the case of buses, route is not a considerable problem as it appears in case of taxis. If we ignore the route changes in case of traffic jam or any related problems, changes in bus routes are not expected.
  2. Stops. Bus routes have many stops, although taxis have only two stop including departure and destination. In some occasions (for example in case of one of the passengers can get off at a location between departure and destination and taxi can continue its trip) there can me more than 2 stops but that will occur often. That results as different application styles of autonomous driving techniques in different type of vehicles when you consider about stops.
  3. Size. When we analyze in terms of maneuver capability and easiness of movement, taxis have much more advantages. As it is difficult for a human to drive a big vehicle like bus, same challenges (and more) appears when it becomes autonomous. For busses, it results as a disadvantage in the traffic on the contrary of capability of carrying many passengers. It will affect the behaviors of the vehicle, as it should maneuver differently in comparison of middle or small sized vehicles. For example; turning problems, changing lines, drawing close to bus stops.


Data from different sources must concentrate meaningfully, and can be used while targeting an efficient public transportation system with AVs usage. Without human drivers and with an internet connection to a management center, these AVs in a public transportation system can result in good scores.
We can catalogue optimized results within an efficient public transportation system with AVs;

  • Route Optimizing. This can be done also in human controlled vehicles, but costs will be remarkably less while managing AVs.
  • Human Factors. Both drivers and other staff, the cost of scheduling and asset utilization would be none.
  • Controlled Vehicle Behaviors. Better control of measuring speed, acceleration and braking. This will result as better passenger safety and comfort through improved driving style.


Beyond the issues of AVs about safety and reliability, public transportation vehicles have other cases that should be considered. There exists further points that can affect both passengers, vehicles and all traffic around the vehicles.
We can examine these issues under two separate headline, as safety and reliability;


In AVs, there are basic safety issues that are associated with driving assistance attributes together [8]:

  • Vision enhancement
  • Collision avoidance
  • Lateral and Longitudinal collision avoidance
  • Safety condition
  • Pre-crash restraint development
  • Automatic vehicle drive

These issues are also considered in public transportation vehicles, but there can vary other topics that should be studied.
Firstly, a public transportation vehicles (especially buses) have different maneuver styles and capabilities in comparison to personal AVs.
Secondly, Passenger get in/off the vehicle another issue. This includes, opening-closing doors and braking-speeding safely.
Lastly, another issue that related to ethical reasons, what should be the behavior of a 60 passenger carrying bus [10] in case of an unexpected situation? Discussions still continue what will be the behavior of an autonomous vehicle when to interfere with an accident. Essentially it is an important issue to consider, as it is explained in the paragraph below [9];

“Here is the nature of the dilemma. Imagine that in the not-too-distant future, you own a self-driving car. One day, while you are driving along, an unfortunate set of events causes the car to head toward a crowd of 10 people crossing the road. It cannot stop in time but it can avoid killing 10 people by steering into a wall. However, this collision would kill you, the owner and occupant. What should it do?”

And when it comes to discuss for public transportation it becomes more complex and hard to solve. What will happen when a 59 passenger carrying bus heading toward a crowd of 60 people crossing the road? Should we assume that as a correct decision when the bus steers into a wall to save 60 people, but killing 59 people? Unless final decisions have been made the AVs will be remained unsecure for the most of people with these vague ethical issues.


Reliability issues are mostly related to security issues, by the evaluation of the passengers. This results consequential effects on the future of AVs in public transportation also.
In a survey [11] of 2000 people, %80 of survey respondents said they would not buy an autonomous car if one were available today. This is a strong reluctance when you consider the people habits in last decades evolving with technology. People had been started to follow latest trends, buying latest technology stuff and if they withstand a technology, it can be better stop and consider there may be some things going not well. There can be many reasons to people think like this, but it shows us it is a susceptible topic and should be considered carefully.
Even people have some trust issues with personal AVs, there are more complicated and harder problems to solve when it comes to subject of public transportation with AVs. Even so most of the people keep at a distance while considering buying an autonomous vehicle, think how they reflect using public transportation with AVs.
Lastly, a fragile issue to contemplate is the test results in the road tests. When people start to hear bad accidents, uncontrolled situations, unexplainable behaviors of AVs that can result as a stillborn technology with disappearing promises and future plans. Namely, it is a delicate subject to many people with skeptical questions in their mind.


There would not seem a future without autonomous vehicles, as the improvements and numbers confirmed. [2] AVs in public transportation figures as an inevitable development after seeing many personal vehicles around over the next years.
All over the world, especially in subway systems [7], autonomous vehicle systems are being used and while some of them need a human operator as helper, some of them operates without any human assistant. With developing technology and growing numbers of companies that are working on AVs, it would not take so long time to be used in public transportation systems. After consideration of the problems of AVs in public transportation that are mentioned and applying the recommended system in this research may lead an efficient usage.




Robotics and Artificial Intelligence

Abstract—This research will present about the robots with artificial intelligence and in which areas that AI have been used in the history for robotics. It will be discussed if the robots can have the capability of dominating humans, as they get better intelligence. The paper also includes the history of robot science, robots anatomy and working principles of the robots, robots with AI, and the future with today’s robotics technology.


In the decades following the Industrial Revolution, machines has become an important part of human life. Hand production methods transacted to machines, manufacturing processes is handled by machines. That led to the development of machine tools which includes operating machines autonomously. Robots are not just in manufacturing anymore, they have become to take part in many areas of human life. The strength of the robots, working precisely, durability and many features has made them popular. Throughout history, functionalities and capabilities of the robots are not increased remarkably until 20th century. Now AI techniques are being tried to make robots more complex, to make them more intelligent machines which do not need human assistance to finish their tasks. Progress in computer science and mechanics, robotics science has a rapid growth and it seems that it will continue.


The term “Robot” was firstly used in a play by Karel Capek named “Rossum’s Universal Robots” [2] in 1921. The story of the play is about human making robots and these robots are destroying the people. As usual in the human history, the dream world of the man has led the way and then real world followed it. Especially after the Industrial Revolution in the start of 19th century, with complex mechanics and electrics the robotics science started to appear and develop. In the early stages of robots, they are built to complete tasks in producing without any human assistance. M. Turing asked “Can machines think?” [6] in his paper in 1950. It is a common question if the machines can have intelligence, even if they would become as complex as our body. Thus, AI community is in doubt between recreating the human intelligence on the machine, and creating intelligence that is unique to the machine. There are many events in robotics history, but a summarized list of some main events can be seen at the table on Fig 1.


Fig. 1. Timeline of Robotics History


Even though the variety of the robots’ types and features may be myriad, they have basic similarities between their constructions. These similarities can be grouped into three categories as; – Mechanical Construction – Electric Circuit – Computer Programming Basically a robot consists of these three parts. Complexity and numerous types of every component in these parts, gives us an infinite possibilities of combustion to build a robot.

A. Mechanical Construction

Mechanical construction provides physical movement to robots which includes a moveable body and a sensory system. According to the type of the robot which is building may need many different mechanical components. For example for building a space robot, that is needed a physical body that can resist high pressure and heat changings, or it is needed small robotic arms if you build a scientific robot to work on experiments with precisely.

B. Electric Circuit

Electric circuit is responsible connection between mechanical components that need electric. It consists of an actuator, a power source and a sensor system. These components may be electrical stuff which designed for a special tasks, microcontrollers, lcd displays, camera systems etc.

C. Computer Programming

A robot is just a motor and electric circuit, if there is nothing there to manage these physical components. This is handled by embedded software that installed into the robot. Robot software is used to perform tasks and automate tasks. There are many software systems and frameworks that can be used for programming robots. Artificial intelligence techniques are used in the software that will be installed to the robot. These include some basic methods like probabilistic inference, planning and search, localization, tracking and control etc. When it comes to robots, there may appear some other problems rather than programming a computer using AI. A computer compares info with the stored data decides about the situation, but a robot gathers facts about a situation through sensors and human point, which makes a robot to analyze the data and decide about it more complex. Some robots have the ability of learning and interacting socially. Not good as a human but robots can absorb the data and get information about the interaction. Techniques of AI are keep developing, also programming these techniques to the robots will come afterwards.


There are lots of different fields and tasks that robots are used in, so it will be better to categorize them by application. The robot types which will be discussed:

  • Medical Robots
  • Space Robots
  • Industrial Robots
  • Military Robots
  • Service Robots
  • Entertainment Robots

Each group works on different domains, although the theories of artificial intelligence may be common, domains may have different appliances.

A. Medical Robots

Usage of artificial intelligence in Medicine has increased in the most a few years. Primarily concerned subject is to construct robots which can perform diagnosis and make therapy recommendations. There are smart robots that can diagnose the illnesses, or talented robots can join the surgeries and they can even get injected into human body and collect data about the health and condition of the body. Technological advancements led the medical robotic systems to be used in many application areas in medicine.

There are some types of medical robots also, including Surgical Robot, Rehabilitation Robotics, Telemedicine, Assistive Robots, Orthotics, Prosthetics, Radio Surgery, Exoskeleton etc. Medical robots are generally used for helping doctors especially during the surgeries. Also robots are used in pharmacies to mix some drugs accurately. As the machine technology and computer science develop, these robots are getting smarter and have better capabilities, accuracies.

Works about AI in medical robotics systems are aimed to configure them as expert or knowledge-based systems or both. There are different types of clinical tasks:

  • Generating alerts and reminders. These tasks include an important aim as monitoring the patience situation. In the surgery or during the patience’s resting, it is vital to monitor and react as soon as possible. A robot that is equipped with necessary sensors, with knowledge and precise working as reduced to a minimum tolerance of error, it can monitor the situation and send reminders or alerts in case of emergency immediately.
  • Diagnostic assistance. When patience’s case is rare or complex, the robot with AI can help on diagnosis compiling the data of patience’s history and similar cases.
  • Therapy critiquing and planning. An intelligent robot can advise for revising a treatment plan based upon accepted treatment guidelines. Also it can formulate the treatment model according to the patience’s previous illnesses and history.
  • Agents for information retrieval. These robots are designed to get information through internet, especially from similar cases happened before relevant to the patience’s situation. It needs to medical knowledge to refine the information that matches the patience’s illnesses.
  • Image recognition and interpretation. Medical images can be analyzed by automated robots and in unusual situations, it can be asked for a doctor attention. Smart robots developed with AI can analyze X-rays, angiograms, CT and MRI scan and diagnose the problem.


Fig. 2. A robot developed for surgery operations.

B. Space Robots

Outside of the Earth is always a challenge for the human history, through the advances in robotics science it led new opportunities to discover extraterrestrial life. As well as the robotics has been developed, outside of the Earth has many though conditions in compare to our planet. First of all there is no gravity. Secondly the temperatures can change wildly in any moment and in any situation. Lastly, the biggest problem is the space is not a habitable place for humans, thus a robot is at its own when it is outside of the world. There comes the artificial intelligence and its applications that the robots must have while they are on their own in the space.

Expected attributes of a space robot are imagining, exploring, self-maintenance, data mining-reporting. Imagining is a must-have feature of the space robots, observing the space or the planet that they are visiting is one of the important events. Exploring attribute pairs with imagining, a space robot is expected to explore the place it visits, take pictures and videos. Data-mining and reporting includes the data which is got from exploring the area and it is a though aim for a robot also. Lastly feature of selfmaintenance is also a must-have for space robots. When a robot is sent to a few million kilometers far away from the Earth, it is not expected to get it back if there is a problem with the robot. Unless there is a big mechanical or electrical problem, it is aimed that the robots that are sent to the space have to fix or at least analyze what the problem is. All these tasks cannot be without artificial intelligence applicants. Computer science has an important part of space robots projects with electrics and mechanics.


Fig. 3. Robonaut 2 – developed by NASA [14]

C. Industrial Robots

Industrial robots have main attributes that differs from the other type of robots. These are, min failure rate, time consuming working and durability. In manufacturing, failure rate is a consuming problem and the systems are expected to work near “0” failure rate. The most considering problem of the industrial robots is decreasing the failure rate. With good programmed robot, this rate can be decreased to minimum even “0”. These robots are built to work fast. With a failure rate near zero, they are built to work in a fast way to make production faster. Lastly, after all of the costs that spent on an industrial robot on mechanics and programming, it expected to have a long durability time.

The applications of industrial robots are welding, painting, assembly, pick and place, product inspection and testing. [16]

In addition to the robots work precision and speed, they can work in dangerous task and use so much power than a human. With the capability of making decision via using knowledge, these robots can manage to do important tasks without any human assistance. Especially in manufacturing, cost of the workers are big expenses, instead of the workers many factory would prefer buying these intelligent machines.


Fig. 4. Industrial robots

D. Military Robots

One the usage area of the robots is military. Autonomous robots and remote-controlled devices are designed for military.


Fig. 5. A defensive military robot

In today’s world the majority of military robots are tele-operated and not equipped with weapons. Artificial intelligence is not related to these remote controlled robots, because they are just mechanics controlled by humans and it is not expected them to do tasks on their own. Especially the autonomous robots have very important advantages when it is compared to humans that is the reason why artificial intelligence is needed for military robots. In 1987 Major Kenneth Rose of the US Army’s Training and Doctrine Command said; “Machines don’t get tired. They don’t close their eyes. They don’t hide under trees when it rains and they don’t talk to their buddies … A human’s attention to detail on guard duty drops dramatically in the first 30 minutes … Machines know no fear.” [17] These words show us the importance of the robots for military. When it comes to reality, if you build 1.000.000 robots for the service and tried to control them by commands remotely, it will just lead to chaos.

Military robots are used in many areas, such as defense, bomb removal, ammo carriage, enemy detect etc. With artificial intelligence, smart robots can detect the movements around a certain area, remove the bombs landmines, or carry the ammo between places.

In another aspect, it makes sense to use smart robots in dangerous and suspicious situations which is too risky for humans to include.

E. Service Robots

Main goal of service robots is to perform useful services to humans especially on household operations. These robots can be studied in three main categories.

  • Domestic and household

Bill Gates claimed that robots will become more common in the future. [19] As well as the robots spread in many areas, domestic and household robots will reach big numbers eventually. These robots are built for helping people at their home. In recent years, to give an example robotic vacuum cleaners and cat litter robots started to find customers to buy. It is inevitable to see expert robots that can do household jobs in the near future with artificial intelligence these do not need any human assistance. These robots can learn the household’s habits and set time and other attributes of the services that it provides.


Fig. 6. Household robot is filling the dishwasher

  • Restaurant and Bar

Smart robots can be used for delivery and cleaning tasks in the restaurants and cafes. With AI, without human assistance these robots can complete daily tasks and also deliver foods to the customers’ tables.

  • Scientific

Robots are helpful in laboratory experiments such as in repetitive tasks. When they are programmed well, they have the capabilities of designing and running experiments, analyzing the data and even forming hypotheses. For example, a robot called “Adam”, became the first robot conducting experiments and make a scientific theory from it. [20] Adam made logical assumptions based on information that he was programmed into it about metabolism of yeast and the way proteins and genes work in other animals. After all that he proved its assumptions are correct.


From the science-fiction movies, books, people are used to hear about robots and have concern about if the robots get intelligence and dominate in the world. Near future does not seem like the robots conquer the world. The robots are already a parts of our lives. In manufacturing industrial robots are widely used, military robots are used for assist in dangerous situations, space robots are investigating the other planets and reporting, house robots vacuuming the floors at homes.

Joanne Pransky, who has been involved in robotics for twenty years and calls herself the World’s First Robotic Psychiatrist, “”Within the next 10 years, I hope to be able to afford to lease or purchase a domestic robot that not only does the household cleaning and prepare and serve my meals, but could carry me to the bathtub if I can’t walk, monitor my vital signs, and if I need a medical specialist from afar, could remotely become his or her eyes, hands, and ears.” Robotics science keeps developing beyond the mechanics and computer science. It will keep changing and the human will have the power to make a robot think some day.


Robots are already in many areas, and their numbers are increasing, their features are developing. With the improvements in mechanics, electronics, and computer science it seems like the robots will be the main parts of our daily life. Applying of the artificial intelligence techniques to these powerful machines will make them powerful and more useful, but robots dominating the world idea does not seem so realistic for the next 30 years.


[1] History of Robots,

[2] A short history of robots,

[3] Three Laws of Robotics, Robotics

[4] Renato M.E. Sabbatini, “Imitation of Life: A History of First Robots”, PhD,

[5] History of Robotics,

[6] A. M. Turing, “Computing Machinery and Intelligence”, 1950

[7] M. Du Sautoy, “AI Robot: How machine intelligence is evolving, The Observer” , 1 April 2012

[8] Types of Robots,

[9] “Global Medical Robots Market (SurgicalRobot, Rehabilition Robotics, Telemedicine, Assistive Robots, Orthorics, Prosthetics, Radio Surgery, Ex oskeleton) – Forecasts to 2018”, Normans Media Ltd., May 19, 2014

[10] Paolo Dario, “Smart Surgical Tools and Augmenting Devices”, IEEE, October 2003

[11] R. H. Taylor, “Robots as Surgical Assistants: Where We Are, Wither We Are Tending, and How to Get There” The John Hopkins University,

[12] E. Coiera, Guide to Health Informatics, Chapter 19, “Artificial Intelligence in Medicine: an Introduction”, CRC Press, October 2003

[13] E. Keravnou, C. Garbay, R. Baud, J. Wyatt, Artificial Intelligence in Medicine, March 199, pages 3-11.

[14] Robonaut 2, Robonaut: Home,

[15] Learn about robouts,

[16] Industrial robot,

[17] B. Pokorny, “Creating the ideal soldier; U.S. Sekkes a Pfc. Robot”, March 1987

[18] “Current use of Military Robots”, current-use-military.html

[19] W.B. Gates, “A Robot in Every Home”, Jabuary 2007

[20] “Industrial and Service Robots”, International Federation of Robotics,

[21] “Adam becomes first robot to make a scientific discovery after conducting its own experiments”, Daily Mail, April 2009

[22] M. Bay, M. Ford, “Robots: The future is now”, April 2006