CARS OF THE FUTURE

Google self driving car

When we hear the term an autonomous vehicle, a driver less vehicle or a self-driving car, the first thought that might jump into our mind would be the 80s television series ‘Knight Rider’.The self-driving feature and the artificial intelligence were quite thrilling to watch at that time. But how many of us would have thought that implementing this kind of technology to the common vehicles that we use daily will be possible in our lifetime? Today,with the technological advancements and thorough research & development processes in the field,technology of autonomous vehicles exists.

The concept of an autonomous car or a driverless car dates back to 1920s, where several experiments and trials were carried out. But promising prototypes were not developed until the 1980s. A group of individuals from the Navigation laboratory (NavLab) of the Carnegie Mellon University was able to develop a successful computer controlled vehicle for automated and assisted driving in 1984. Latera large project known as Eureka PROMETHEUS project was carried out successfully by the Bundeswehr University Munichwith the collaboration of Mercedes-Benz from 1987 to 1994. The DARPA (Defense Advanced Research Projects Agency-USA)Grand Challenge which was held from 2004 to 2013 focused on creating fully autonomous ground vehicles which are able to travel long distances. Theseprojects had opened many keyresearch areas towards the advancement of autonomous vehicles. Later, many companies and research organizations such as General Motors, Mercedes-Benz, Tesla Motors, University of Oxford, Toyota, Google, etc. continued researches and developments creating a fully autonomous street legal vehicle.

Tesla S

Through the advancement of different technologies such as radar, GPS and image processing, the fully human driven vehicles were made partially automatedallowing the vehicle to assist and make decisions for the driver in particular situations. Crash warning systems, Adaptive Cruise Control (ACC), lane keeping systems and self-parking systems are a few of such advanced technologies we use every day.

From the completely human driven vehicle to the completely autonomous vehicle, there is a hierarchy that was implemented to clarify the level of automation by National Highway Traffic Safety Administration (NHTSA-USA).

Level 0 (No Automation) – The driver is fully responsible for the complete control and following safety regulations of the vehicle at all times.

Level 1 (Function Specific Automation) – The driver has the overall control and has to operate the vehicle safely, but one or more specified automated functions such as ACC and dynamic break support can be used as an added control aid.

Level 2 (Combined Function Automation) – More than one functioncan be automated simultaneously. (e.g. steering and acceleration) But the driver must monitor the road and safety conditions alwaysand should be able to gain the full control of the vehicle incase of an emergency.

Level 3 (Limited Self-driving Automation) – The vehicle is responsible for the control of roadways and all safety-critical functions in specific environments whereas the driver can engage in different activities. At this level of automation, the driver is needed to gain the control of the vehicle occasionally.

Level 4 (Full Self-driving Automation) – The vehicle is responsible for the total control of the vehicle, inspecting the roadway and safety-critical conditions throughout the whole journey. At this level of automation, a driver may not be required at all.

Currently, a fairly advanced and a very sophisticated autonomous system isdeveloped in the Google self-driving car. It consists of a laser rangefinder, radars, a high-definition camera, a GPS, an inertial measurement unit and a wheel encoder. From them, the Velodyne 64-beam laser rangefinder is the most important sensor since it maps points in space using 64 rotating laser beams. The sensor takes more than a million measurements per second to form a three-dimensional model of the surrounding environment for a range of about 120 m andit is accurate to less than two centimeters. The details from the high-quality laser sensor are then combined with high-resolution maps of the surrounding to generate the accurate position of the vehicle and the positions of obstacles such as other vehicles, pedestrians, cyclists, etc.

The four radar sensors mounted on the front and the rear of the vehicle are used to calculate positions of the long range obstacles and fast moving vehicles. The high-definition camera mounted in front of the rearview mirror is used to track and detect the traffic lights. The GPS, inertial measurement unit, and the wheel encoder are used to keep track of the precise position of the vehicle. The approach does not use the GPS alone since the GPS data might not be precise always due to the environmental conditions.

After taking the required sensor data in, advanced control theories and probability theories are applied for positioning the car on the road, predicting the behaviour of the other vehicles, cyclists, and pedestrians. Safety of the passengers and the outside personnel is the most important factor when designing this autonomous system.

The autonomous vehicles will affect the safety, traffic congestion and energy usage enormously. Theymight be able toreduce the frequency of traffic collisions. Currently, the Automatic Braking System has reduced a considerable amount of rear-end collisions. Therefore, a fully autonomous vehicle will be able to function with almost no collision since the human error factor, which is the reason for about 90% of the collisions, does not exist. The autonomous vehicles of level 4 automation will enable the transportation of individuals which are unable to drive, such as children, blind, disabled or elderly. Vehicles above Level 3 automation will potentially reduce the traffic congestion and its costs drastically. Also,they will reduce the timespent drivingthe car, and the driver is able to engage in alternative activities.

There will be a considerablereduction of the energy usage and the amount of emission due to autonomous vehicles. The fuel economy can be improved by the smooth autonomous acceleration and deceleration done other than a human driver. Also, the roadway capacity can be increased by reducing the distance between vehicles. A group of closely positioned autonomous vehicles with less speed variation (especially on expressways) can improve both the fuel economy andthe travel timeand reduce the traffic congestion.

Along with the advantages, there are potential disadvantages of autonomous vehicles and their introduction to the general public. The customers are still doubt about the safety of fully autonomous cars. Also, establishing a legal framework and government regulations to them is difficult. Driverless cars might reduce the job opportunities in the driving sector and drivers may be inexperienced to handle a complex situation manually due to the lack of experience they might have. There may be disputes concerning the liability in case of a collision. The computer system and communication system can be compromised. Also, the sensors may not be functioning properly under different environmental conditions.

A Level 4, fully autonomous vehicle is not street legal up to the date. But some countries have legalized the level 2 vehicles with the driver assist features. Even though there are several technological and legal backgrounds that are required to be cleared, the research teams expect to launch fully automated, street legal vehicles by 2020. From the engineering point of view, the autonomous vehicles will have one of the most advanced electronic and mechanical real-time systems which will add many control, image processing, communication and robotics elements.

References

1.Wired (2016)A brief history of autonomous vehicle technology [Online] Available from :https://www.wired.com/brandlab/2016/03/a-brief-history-of-autonomous-vehicle-technology/
2.James Anderson (2016) Autonomus Vehicle Technology [Online] Available from: http://www.rand.org/content/dam/rand/pubs/research_reports/RR400/RR443-2/RAND_RR443-2.pdf
3.Google Self driving car(2016) [Online]  Available from:https://www.google.com/selfdrivingcar
4.Jamie Page Deaton(2016) How Driverless Cars Will Work [Online]  Available from: http://auto.howstuffworks.com/under-the-hood/trends-innovations/driverless-car.htm
5.Erico Guizzo(2016) How Google’s Self-Driving Car Works [Online]   Available from: http://spectrum.ieee.org/automaton/robotics/artificial-intelligence/how-google-self-driving-car-works

Thilanka Pathum Weerasekara
Third year,
Department of Electrical & Electronic Engineering, Faculty of Engineering, 
University of Peradeniya.