Inteligent Transportation Systems
With advancements in technology, autonomous transportation solutions are being developed for the average consumer. In-‐vehicle features are growing in popularity as is the idea of fully autonomous vehicles. The Google self-‐driving car is one such project that has gained a lot of traction. It is indicative of the direction for transportation development and can be used as a foundation for future systems. Applying the autonomous systems related to individual vehicles, intelligent transportation systems can be developed to increase efficiency by providing innovative services for traffic management, enabling vehicle operators to make more informed, more coordinated and overall smarter use of transport networks. The goal is to implement a fully autonomous system, designed to be anticipatory or reactive to changing environmental parameters. Such a system would have the effect of decreasing traffic, improving safety and reducing the environmental impact of vehicle transportation. Development in this area is critical because there are close to a billion vehicles on the road today, and a doubling projected over the next 15-‐20 years (Balakrishnan & Madden, 2011). Challenges resulting to even more cars on the road need to be met with innovative solutions to increase efficiency.
Today, intelligent systems are more common than one might expect. The introduction of GPS route planning, adaptive cruise control and autonomous steering are all examples of autonomous system integration and are a part of many new vehicles offered by the automotive industry. Less advanced than a fully autonomous vehicle, these technologies are the basis for future development and are a necessary step in the overall adoption of more intelligent systems. As a gradual process of decreasing the importance of the vehicle operator, the adoption of more intelligent transportation systems will be a slow and seamless process as to minimize negative reactions from drivers and to ensure that future systems are operational and reliable. As more technology is introduced to the automotive market, there is an opportunity to leverage this shift by introducing autonomous interactions between vehicles and with greater systems.
The introduction of in-‐vehicle technology and the increase in communication capabilities present an opportunity to develop more frequent communication between vehicles and with a larger system to improve transportation efficiency. Such an increase in information distribution coupled with an intelligent system will help drivers make better decisions on the road. The result of this will be increased traffic management capabilities, safer driving environments and reduced emissions because of more efficient use of transportation networks.
Traffic congestion in the United States has steadily increased over the past thirty years and is a serious metropolitan concern. Since 1982, annual time spent in traffic by American drivers has increased from 16 to 46 hours. This is relevant because a result of all that time sitting in traffic is an increase in unrealized earnings.
Since 1982 unrealized earnings as a result of traffic has gone from $14 to $63 billion dollars (Morgan, 2013). To combat this, many metropolitan areas have increased the size of roadways or build additional routes; however as the demand on our roadways continues to increase, continuing to do this is not economically efficient. Rather, it is imperative that we make better use of our current system by reducing traffic congestion and help drivers anticipate road conditions and alter routes accordingly.
Another reason to develop more intelligent transportation systems is to increase safety on U.S. transportation networks. As a result of increased vehicle-‐to-‐ vehicle communication, collisions can be reduced. This is a technology that is already offered in a number of commercial vehicles however broader implementation is necessary in order to have more of an impact. Additionally, communication between vehicles should not be limited to just avoiding collisions but rather be increased to have large systems of vehicles all communicating with each other. This concept is known as swarming which brings together artificial intelligence, robotics, and even biology, applying an understanding of swarming behaviors in nature to a system of autonomous vehicles (Kumar, 2009). As an incremental step towards fully autonomous systems, the driver is being made more and more obsolete. Decreasing the uncertainty of vehicle movement on the road will inevitably decrease the number of accidents. When applied to millions of cars, the number of accidents involving autonomous vehicles will be very small (Roberts, 2013). As evidenced by the Google car, which has only been involved in accidents when being human operated shows that the technology is reliable enough to operate safely (Murray, 2012).
The effect of using U.S. roadways more efficiently is a decrease in automotive emissions. Currently Americans burn 5.6 billion gallons of gasoline while idling in heavy traffic every year (Smith, 2012). That is 56 million tons of carbon dioxide emissions every year. Anything to reduce this number is very relevant when concerned with the future of vehicle impact on the environment. Autonomous systems can help reduce emissions by increasing travel efficiency by giving the driver more information on how to optimize their route and plan for traffic congestion.
Current methods of gathering data on vehicle operation are for the most part passive processes. Using embedded sensors, intersection cameras and radar are the most traditional ways to monitor cars on the road. It is a the foundation for a larger system to monitor vehicles and better coordinate road conditions. More recently, with the increased frequency of cell phone users, data about vehicle operators can be more accurate and predictive than passive sensors. Active mobile phones present real time information on traffic conditions and when recorded can present models of historical traffic conditions. Google maps currently generate real-‐time traffic conditions by using android phone users to monitor their location. To work effectively only a few phones are needed to gather data on traffic congestion, which is another benefit of using floating cellular data to build out a network of current traffic conditions (Burgess, 2012). Utilizing cellular data is very accurate for monitoring individual cars however has some limitations in that service can be inconsistent in “urban canyons” and uses a lot of energy. These limitations are not however enough to hinder the effectiveness of such a system on a large scale. Improvements n the connectivity of cell phones will inevitably lessen the effect of “urban canyons” further providing support for using floating cellular data for traffic monitoring purposes.
Another project to monitor vehicle movement is presented by the MIT CarTel project, which utilizes existing Wi-‐Fi networks. As a cell phone moves from one Wi-‐ Fi network to the next, data can be gathered on that individual’s movement. When coupled with existing road maps, an understanding of vehicle movement can be modeled. This system is beneficial in that it uses less energy and is widely available in urban environments, areas where vehicle congestion is more prevalent. Unfortunately this system is less accurate, and locates vehicles in the tens to hundreds of meters (Balakrishnan & Madden, 2011) .
Based on these methods of vehicle monitoring, data can be synthesized to anticipate road conditions and to monitor real time traffic conditions. Current ways to then optimize current transportation systems are with route optimization, light sequencing, variable speed limits and intelligent parking systems. The eventual goal is to link all systems and create a larger autonomous system for more efficient vehicle transportation.
Route optimization is a very common tool for making efficient use of existing roadways. As a feature on many GPS systems, routes can be designed around current traffic conditions, finding the fastest route to any given destination. As a basic technology that has gained in popularity, building this type of system out to become a constant monitor of the operators’ vehicle and other vehicles, a greater understanding of roadway use can be developed. Taking this a step further, integrating this technology with autonomous vehicles, use of existing roadways would automatically be optimized.
Another simple yet very effective tool for traffic management is the use of traffic light sequencing. Lights are timed to essentially group cars and enable more efficient use of the road by controlling right of way, thereby reducing traffic in congested areas by slowing other cars. Unfortunately these systems are designed to assume continuous traffic flow, not taking into account changing traffic conditions or accidents. In order for this system to remain relevant with the introduction of more intelligent systems, light sequencing must be adapted to monitor real-‐time traffic conditions and apply measures to decrease traffic in congested areas and maintain traffic flow at all times. In a study conducted at the International Islamic University Malaysia researchers looked at the traditional method of using beam interruption in detecting if cars were present at intersections. To build on this existing technology, they proposed a system, which evaluates traffic conditions real time by monitoring columns of traffic at intersections to develop a model of current traffic conditions. In their trial runs, their system worked efficiently and responded well to extreme traffic conditions, effectively enough they said to replace the need of an intersection police officer (Khalid A. S. Al-‐Khateeb, Jaiz A.Y. Johari & Wajdi F. Al-‐ Khateeb, 2008).
In the same vein as controlling traffic flow, another existing application for intelligent traffic systems is the use of variable speed limits. This means adjusting speed limits as traffic conditions change. As a more adaptive system than current dynamic light sequencing, variable speed limits are adjusted based on historical traffic conditions. The goal is to increase safety by limiting vehicle speed and prepare drivers for upcoming traffic. The result is less stop and go and therefore more efficient use of the road. Though limits can be set, it is difficult to monitor and enforce (Hart, 2012). In the case of autonomous cars, there would not be any way to violate the posted speed limit, thereby increasing the efficiency of such a system. Moving forward, greater implementation of speed limit optimization is inevitable. Road conditions are constantly changing so a more accurate way to measure traffic congestion is necessary for such a system to work to its full potential. Utilizing an existing method such as floating cellular data would be an accurate way to monitor this and ensure that the system is operating on accurate traffic data.
Making the most out of existing roadways is particularly applicable to intelligent parking systems. Using embedded sensors, drivers are given a real-‐time representation of parking availability. Additionally, long term, parking trends can be identified based on parking demands and use. Intelligent parking systems make finding parking easier. It also reduces traffic congestion, seeing that a third of all traffic is caused by vehicles circling or idling when looking for parking. An indirect effect is also a reduction in emissions. Currently intelligent parking systems are used in parking garages in the U.S., Japan and Europe. It catalogues all available parking spaces and notifies the driver upon entering the garage where to find the
closest spot. On the road, parking spot monitoring is done with sensors in the road, connected to a central system to monitor the availability. This system can monitor and then even price spots based on demand (Berg, 2012).
The application and further development of existing systems to a larger autonomous system is a future direction to be explored. Building out what is currently available, an autonomous system based on centralized control is feasible. It is a matter of integrating all available systems and increasing the communication between vehicles and that system. Ultimately however, a decentralized system is more dynamic and able to accommodate more individual vehicles. As the technology continues to improve and adoption of transportation monitoring systems becomes more prevalent, the reality of a fully autonomous system is within reach. Looking at the trend of self-‐driving cars and individually autonomous vehicles, the same technology can be utilized to build out a larger system. The ultimate goal is to have vehicles constantly communicating with sensors and with each other. The result of a large system would be significantly increased road capacity due to more efficient use of existing roadways and more efficient parking capabilities. Additionally, roadways would be safer, removing the importance of human-‐originated decision making and instead having cars reason through obstacles at a fraction of the time. For this to be a major result of an autonomous system, there would need to be a high level of adoption. With safer roadways also comes a decreased accident rate, thereby having an effect on traffic congestion. Furthermore, an effect of more efficient and safer use of roadways is a decrease in energy use. As evidenced by the current state of traffic congestion and immediate need for energy efficiency, adopting and funding research in autonomous transportation systems is very relevant. Reaching the level of full autonomy will come at a huge cost. More work needs to be done to improve the technology and increase communication capabilities. There also needs to be a shift in the public perspective. It is imperative that consumers are open to the idea of autonomous vehicles and learn to trust the technology. This will only come with time and through proving reliability, something that the Google self-‐car is doing.
Autonomous transportation systems are not limited to the roadways. Applications to air and rail transportation can further increase efficiency and decrease the need for human operators. Though air traffic control systems are fairly advanced and reliable, automation of the system will reduce the need for human operators and can help ensure continued safety of air travel. Doing so would also enable more efficient use of airspace, allowing for heavier use. This would translate to having less wait times on the runway and decreased time when approaching airports. Both of these improvements result in decreased emissions through more efficient use of current aircraft. On railways, better coordination of routes based on user and freight demands will result in more efficient use and therefore less emissions as well.
The application of existing monitoring systems to the future of transportation systems will result in more efficient use of U.S. roadways. The goal is to eventually achieve a fully autonomous system in which traffic congestion is minimized and safety is increased. A benefit of more efficient system is decreased emissions and less environmental impact from vehicle transportation. Current systems are a foundation for this future development. Adapting and combining them will be the key to creating a comprehensive system that is safe and reliable. Once the public is on board, vast implementation would be possible. The vision for the future of transportation may not be the once imagined hovering cars, or personal helicopters but rather a hyper-‐efficient system of decentralized autonomous vehicles.
References
Adler, J., & Blue, V. (2002). A cooperative multi-‐agent transportation management and route guidance system. Transportation Research Part C. 10. 433–454.
Balakrishnan, H., & Madden, S. (2011). CarTel [MIT CarTel]. Retrieved from
http://cartel.csail.mit.edu/doku.php
Berg, N. (2012). The Future of Intelligent Parking. The Atlantic Cities. Retrieved from http://www.theatlanticcities.com/commute/2012/03/future-‐intelligent-‐ parking/1573/
Burgess, R. (2012). Google Maps Get Real Time Traffic, Crowdsources Android GPS Data. TechSpot. Retrieved from http://www.techspot.com/news/48015-‐ google-‐maps-‐gets-‐real-‐time-‐traffic-‐crowdsources-‐android-‐gps-‐data.html
Khalid A. S. A, Jaiz A.Y. J. & Wajdi F. A. (2008). Dynamic Traffic Light Sequence Algorithm Using RFID. Journal of Computer Science 4 (7). 517-‐524.
Kumar, V. (2009). Scalable sWarms of Autonomous Robots and Mobile Sensors (SWARMS) project. SWARMS. Retrieved from http://www.swarms.org/
Morgan, D. (2013). The Cost of Highway Congestion. ABC News. Retrieved from http://abcnews.go.com/US/story?id=94064&page=1#.UYsdyjl1Pww
Murray, P. (2012). Google's Self-‐Driving Car Passes 300,000 Miles. Forbes. Retrieved from http://www.forbes.com/sites/singularity/2012/08/15/googles-‐self-‐ driving-‐car-‐passes-‐300000-‐miles/
Roberts, J. (2013). Autonomous Robotic Systems. CSRIO. Retrieved from http://www.csiro.au/en/Organisation-‐Structure/Divisions/ICT-‐ Centre/Autonomous-‐robotic-‐systems.aspx
Smith, B.W. (2012). Managing Autonomous Transportation Demand. Santa Clara Law Review. 52. (4).
Hart, A. (2012). Variable speed limit signs to go up on I-‐285. ajc.com. Retrieved from http://www.ajc.com/news/news/transportation/variable-‐speed-‐limit-‐signs-‐ to-‐go-‐up-‐on-‐i-‐285/nSGPs/