Technical Concept Report
Technical Summary
AITOS uses numerous technical and scientific principles in its design and functionality. AITOS uses 4 main principles: Data centralization, Same size k means algorithm, Real-time vehicle detection algorithm, and coordinated traffic control.
Firstly, AITOS uses data centralization to accumulate data. Data centralization is a method of organizing data, where all data is collected and stored in one area but is accessible from many points. We will collect visual traffic data using TDMs(traffic data modules) and data packets containing the IP address, positional data, and velocity from each device with AINS. All of this data will be centralized to its respective location. Data centralization is necessary, as we will need to accumulate data in a single place to perform analysis with AI and Machine learning. This analysis will help us find the most efficient method of displaying traffic lights to achieve maximum traffic efficiency. Centralized data has many benefits. It provides data security, integrity, operational efficiency, and reduced costs.
Second, as our servers will be receiving an immense amount of data, we need to strategically partition our servers. Failing to do so may cause our servers to be overloaded, rendering AITOS impractical. We hope to partition our servers using the same size k means algorithm. This algorithm is used to cluster points into groups of similar sizes. For example, with this algorithm, it is possible to group 1000 points on a plain into groups of 20 points with similar areas. This algorithm will be simulated in real geographical regions to find the best location for our servers. We will consider all traffic lights and devices (mobile, AVs) that are able to install AINS as a point. Then, we will separate the points into a certain amount of clusters. The cluster will be formed such that the maximum amount of data produced by the devices in the group will not overload our servers. In each cluster will be a server, placed accordingly to address different variables and to ensure functionality.
Next, TDMs (Traffic data modules) play a vital role in providing AITOS with crucial real-time visual traffic data. TDMs are equipped with LiDAR cameras that allow them to view the state of a traffic lane. As they view their assigned traffic lane, they will continuously run the Real-Time Vehicle Detection Algorithm. The Real-Time Vehicle Detection Algorithm is an algorithm used with vision-based sensors, such as cameras, to allow them to detect vehicles in real time, as its name implies. This technical principle will be crucial, as it will allow TDMs to produce visual data that will be used for many different functions. For instance, our server will count the number of vehicles detected in a traffic lane to output the number of cars in that particular lane. It will also be able to track the average velocity of each detected vehicle. If there are multiple cars, each with a low average velocity, our server will recognize this lane as a congested region. Moreover, using the visual data produced by TDMs, our server will be able to create a replica of the traffic of its assigned regions, to the precision of each individual vehicle. This algorithm will also allow us to detect car crashes, one of the main causes of slow traffic flow.
Lastly, AITOS will use an improved version of coordinated traffic control. Coordinated traffic control is a method of organizing traffic by introducing drivers to a progression of green lights. Once vehicles encounter one green light, at subsequent intersections, they will be met with a continuous series of green lights. Coordinated traffic control permits drivers to travel long distances without coming to a stop. This produces increased traffic efficiency compared to standard traffic systems. However, this method of control has one fatal flaw: congestion. This method proves to be very inefficient during congestion.
We hope to base AITOS on this principle. We hope to strategically control traffic lights to ensure maximum traffic efficiency. However, unlike coordinated traffic control, we will be adapting to each situation using Real-Time traffic data. Coordinated traffic control guarantees efficiency (to a certain degree) for most, but not all situations. To ensure that AITOS is effective under all conditions, we will be using an AI that will determine the best method of controlling traffic using traffic lights. This will allow AITOS to change its method of control along with each variable it will face. This will certify maximum efficiency for all situations.
Need Statement
Problems: economic loss due to traffic, pollution, traffic jams in general, saving time for people going places, and pedestrians feeling safer, Lumino.net aspires to maximize traffic efficiency. Most traffic systems in the world are inefficient and pose several problems and challenges. They contribute to major time loss, as well as economic losses and pollution. Furthermore, traffic delays also pose safety issues. We hope to address these problems with our concept, AITOS (AI-based Intelligent Traffic Optimization System).
Firstly, traffic jams cause major personal inconveniences. They make people miss important meetings and consume an immense amount of time. According to recent studies, the average U.S. driver loses up to 97 hours in traffic per year. This time loss also leads to economic and environmental consequences. Statistics say that traffic congestion accounted for the loss of nearly $87 billion and is one of the main causes of economic stagnation. Moreover, this time loss also translates to more greenhouse gas emissions. Drivers typically do not turn off their cars when their vehicles are immobile, therefore, these gases will continuously be emitted into the atmosphere as the drivers are in traffic. Therefore, each driver will produce an extra 97 hours worth of GHG emissions. Studies have found that vehicles contribute to nearly 30% of all greenhouse gases emitted in the U.S. Additionally, upon further research, we have found that traffic delays render people more and more susceptible to accidents, as they induce anger and fatigue, one of the main causes of accidents. Statistics show that there are more than 1.35 million deaths due to traffic accidents per year.
By maximizing traffic efficiency with AITOS, we hope to address, to some degree, all of the following issues listed above. AITOS is an important innovation in the ITM (Intelligent Traffic Management) sector of the IoT (Internet of things) and the Transport Industry. Most businesses that hope to address traffic congestion propose temporary and costly solutions. Most of these solutions are temporary, as traffic jams are cumulative.
In a recent study, it was found that congestion accounted for the loss of over 100 billion dollars annually in America. This is due to the omnipresence of the largely inefficient traffic in America. Since congestion is cumulatif - it accumulates over time - the effects of an inefficient traffic system will be exponential by a factor of time. Simply put, traffic will become worse as more time passes. This will then lead to more time spent in traffic, which will lead to less productivity and thus will result in an economic drain. However, with the use of our concept, the economic losses can be minimized or even eliminated, for our concept maximizes the efficiency of traffic flow. Our concept will allow people to get to their destination faster than current systems. This will contribute to the economic growth of the country and will benefit both the companies and the workers. To companies, there will be more total productivity, for the loss of time will be minimized. For workers,
https://www.drivingtests.co.nz/resources/how-do-traffic-jams-form/
https://en.wikipedia.org/wiki/Gridlock
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6766193/
https://www.drive.com.au/news/road-rage-why-traffic-congestion-creates-angry-drivers/
https://www.who.int/news-room/fact-sheets/detail/road-traffic-injuries
https://www.smartertransport.uk/smarter-cambridge-transport-urban-congestion-enquiry/
https://www.sba.gov/business-guide/plan-your-business/write-your-business-plan
Background Technology
AITOS will be built upon several existing technologies such as AV autopilot, existing ITM concepts, navigation technologies, and coordinated traffic control.
Our competitors include ITM businesses such as Valerann and Miovison and navigation application businesses such as Google maps and Waze.
Companies such as Valerann and Miovison contribute to traffic management by offering insight on how to better manage roads. They strategically place cameras and sensors near traffic lanes to collect real-time traffic data. They upload this data to their cloud and accumulate detailed regional traffic data. Then, with the use of AI technology and machine learning, they predict future
Automated cars like the ones from Tesla usually use Lidar (Light Detection and Ranging) to detect objects in front of them and cameras to determine the state of the traffic signals. Lidar devices use light as a laser to measure distance. The nature of using light in both Lidar devices and cameras makes them prone to being inoperable in bad weather, like a foggy, rainy, or snowy day, where the traveling of light in the air is altered. Instead of Lidar, the product we offer uses radio waves to send two-way low radio frequency signals. Radio waves can still be impacted by bad weather like heavy rain too, where attenuation happens in the rain droplet-saturated atmosphere, but low-frequency radio waves below 5 Hz are impacted much less by attenuation compared to Lidar. Lidar scanners use electromagnetic waves that have a frequency ranging from 1 Hz to 100 Hz, which are heavily impacted by attenuation in bad weather, because of the high frequency. With radio signals, our device will allow automated cars to receive information from the traffic signals on the placement of the stop line, the traffic signal phase, and whether a pedestrian green light is active.
The traffic signals used now at intersections in urban areas coordinate interrupted traffic flows coming from different directions. A lot of accidents happen at these intersections, and they are not the most efficient. They are standardized and usually have a set cycle time, going from red light to green light with the same time duration forever. That means even if one side of the crossroad has significantly more cars waiting than another, they get the same clearance time. The result of this is traffic congestion. Some traffic lights nowadays have the ability to change their cycle duration. Most of them rely on embedded traffic detection devices like pressure plates and traffic cameras to survey the amount of traffic flow. Cameras and pressure plates can easily have failures like inefficient scanning programs and detection errors. They also don’t communicate with each other and don’t work together. Every camera of every different crosswalk has to detect the traffic flow again when a group of cars from another crossroad comes to a stop.
Our gadget allows traffic lights to receive signals from incoming cars and adjust their phase times according to the traffic flow. The signals come straight from the automated cars, so there is less possibility for errors since the traffic lights don’t have to analyze the number based on images like with a camera. They can also communicate with other traffic lights by uploading the data collected to our database. We will use deep learning artificial intelligence to come up with an adjustment of phase duration to apply to the traffic signals around the troubled crosswalk so that the next crosswalk the cars stop at will already have the information of the amount of traffic that needs to be cleared and how much clearance time is needed for every lane.
Some traffic detection devices like pressure plates and traffic cameras nowadays upload their data to Google Maps. Google Maps takes these numbers and displays the statistics as travel suggestions. Routes with heavier traffic will appear as a slower routes or have busy area warnings. This is very useful for people who want to get somewhere without being delayed by traffic jams, but there is still a big room for improvement. Google Maps doesn’t give information back to traffic lights. For example, a section of the road might have gridlock, and the traffic detection devices will report that to Google Maps. Google Maps will display that the route is having a gridlock to its application users, but the traffic signals won’t know what the data they gave means. They will still give the same time for each phase. The situation of gridlock will not improve. With our product, the traffic signals will be able to receive the analyzed data from Google Maps. With the programming in our software that’s incorporated into the gadget in the traffic signals, they will know how to solve the traffic jam by changing the duration of the phases of a traffic light or letting the line of cars that are blocking the intersection get cleared first.
https://www.youtube.com/watch?v=DP62ogEZgkI
https://oceanservice.noaa.gov/facts/lidar.html
https://www.everythingrf.com/community/what-is-the-impact-of-rain-on-rf-signal-propagation
https://www.merriam-webster.com/dictionary/attenuation
https://www.generationrobots.com/blog/en/what-is-lidar-technology/
Our concept aspires to achieve several goals
Our concept is intended to reduce significantly or even eliminate the dangers present in crossroads. With our concept, we aspire to assist in the transition to automated vehicles. Our concept has 2 key parts: the new traffic light and the software, which will be available on mobile devices. The new traffic light will work in conjunction with the software to serve its purpose. Our new traffic light will be equipped with lidar technology and sensors. The sensor will continuously send out signals to vehicles and devices that have the software installed, and the vehicles and devices will send a signal back to the traffic light upon receiving the signal, telling the traffic light that the device has successfully received the signal. Meanwhile, the lidar technology will create a visual representation of the lane.
More development about lidar
The lidar will then upload this data to our database. Our database will When the devices detect that there are too many cars in the lane, it will prioritize this specific lane to combat traffic jams. However, we recognized that lidar technology is highly unreliable in bad weather. Thus, to combat this, the traffic light will count the number of signals that were sent back to assist the lidar’s function.
https://miovision.com/blog/urban-congestion-impacts-and-improvements
https://www.hindawi.com/journals/js/2019/8473980/
https://www.techtarget.com/searchnetworking/definition/packet