Project_Objectives
- 1. 1 Objective The objective is to design and implement a Java application that can display traffic flow on a map according to existing sensor data, and display traffic events that will be calculated by external modules. There are several several sub-tasks to be fulfilled: 1. Visualize traffic information (can be toggled between traffic flow and average speed) on a road map. These traffic data are values over time and the application must offer a way to change the time. 2. Mark areas of the map specified by our algorithm, which will mark traffic anoma- lies. For visualize of traffic flow and average speed, the end result might look like the traffic information provided by Google Maps, as shown in Figure 1.1. Please be aware that the data changes over time and you need to implement a way to change the time, e.g. by a slider going from the first record to the last. Low traffic flow should be displayed in red, medium traffic in yellow and high traffic flow displayed in green color. The colors for low average speed are red, medium speed yellow and high speed green. The values can take different shades from red to yellow to green. There needs to be a way to toggle between these data. You are encouraged to design an application modular and make use of libraries you might see fit. 1
- 2. Figure 1.1: Example of traffic information visualization by Google Maps. See online: http://google.com/maps/place/Los+Angeles/@34. 014523,-118.2685669,11z/data=!4m2!3m1!1s0x80c2c75ddc27da13: 0xe22fdf6f254608f4!5m1!1e1 2
- 3. 2 Background As much effort is being done to avoid traffic events, such as accidents or obstructions on streets, it can not be prevented entirely. Because these kinds of events usually involve humans’ lifes, quick reactions to these events are necessary. Therefore, an Automatic Incident Detection (AID) system is crucial to identify and react quickly to any kind of traffic anomaly. A large amount of sensors were installed to monitor traffic in many regions. These sensors supply data, such as location of the sensor, average traffic speed or general traffic flow data. Analysing this large amount of data is a challenging task and even though there are existing approaches to identify traffic events, their ability to pinpoint the exact location of an anomaly is still improvable. AID systems objective is a precise traffic anomaly detection. When an estimated location and timestamp of the result is on or close to the ground truth, we regard it as better. To visualize the calculations and for better understanding of the mechanics behind a complex traffic system, we need a system that can display relevant data in a form that is easily be understood by humans. 3
- 4. 3 Detailed Description 3.1 Data Set We are using an existing dataset, Caltrans Performance Measurement System (PeMS)1, which contains archived and real-time traffic information data. The traffic data is col- lected from over 39,000 detectors. These sensors span the freeway system across all major metropolitan areas of the state of California, USA. We have downloaded some data to use locally, so we can access the data more quickly. Willy has a copy on his computer. 3.2 Relevant parts of existing program 3.2.1 Sensor.java The existing program is already able to read sensor data into Java objects. There are several classes relevant. The first one is the Sensor class, described in Listing 3.1. Listing 3.1: Attributes of the Sensor class (reduced to show only important parts) /∗ . . . ∗/ public c l a s s Sensor { /∗ . . . ∗/ private i n t id ; private i n t fwy ; // freeway id private double latitude ; // l a t i t u d e private double longitude ; // longitude /∗ . . . ∗/ private double absPM ; // absolute Postmile /∗ . . . ∗/ private String name ; // sensor name private i n t direction ; // the d i r e c t i o n of the freeway i t senses /∗ . . . ∗/ } 1 http://pems.dot.ca.gov 4
- 5. As you can see, a Sensor contains several relevant attributes for the visualization task. The attributes are on private visibility, but can be accessed using standard Getters and Setters. The important attributes for visualization are: id Unique identifier for this sensor. fwy Which freeway this sensor monitors. latitude and longitude Coordinate location. absPM The absolute postmile is the distance to the end of the monitored street. Can be used to identify the order of several sensors on the same street. name Same of the sensor. direction In which direction the sensor monitors (North, East, South, West are 0, 1, 2, 3 respectively) 3.2.2 Record.java Each Sensor will collect a series of Records. The important part of the Record class can be seen in Listing 3.2. The attributes can be accessed with standard Getters and Setters. Listing 3.2: Attributes of the Record class (reduced to show only important parts) /∗ . . . ∗/ public c l a s s Record { /∗ . . . ∗/ private TimeInterval timeInterval ; private Sensor sensor ; private i n t totalFlow ; private double avgSpeed ; /∗ . . . ∗/ } timeInterval The interval the Record is valid. sensor The Sensor that made this Record. totalFlow The traffic flow recorded. avgSpeed The average traffic speed recorded. Misc • For writing your code, we suggest to write it in a maintainable way. Maintainable means that it is a good idea to write it very clear, prefer a simple solution over the overly complex one and add comments where resonable. Comments should also include understandable JavaDoc. 5
- 6. • A consistent coding style is recommended. We suggest to follow the Java Code Conventions2. • We suggest to write brief summaries of whatever you do. In the end, you will need to hand in a report. This report will be easily done if you have written your daily archivements. • One paper that we are competing with is this paper3: “A New Framework for Traffic Anomaly Detection” 2 Most IDEs like Eclipse have an automatic code formatting function, having Java Code Conventions pre-installed 3 Available online: http://dx.doi.org/10.1137/1.9781611973440.100 6