There is a basic scenario in which a bus travels from points A to B and in between some number of stops can be added. Bus has a mobile sensor and stops A and B has static sensors. In this project we have to create a set of docker containers which will play the role of agents. One container will act as stop A and a static sensor and B also acts as a stop and another static sensor. However,2 containers can also be created for one agent, one will give the general information of the stop including the parameters like geo coordinates of the stop, time stamp and the other container will generate the data including the parameters like location of the stop, time stamp and weather data. Another container will act as a mobile sensor and the Bus.
All the containers should be able to generate data according to the type of sensor they are acting as. Mobile sensor should create data for both the bus which includes variables like the speed of the bus, GPS location and time stamp and for the mobile sensor itself the continuous created data should include parameters like GPS value for the sensor, particulate matter values, temperature values and humidity values. The stationary sensors would always give the same GPS data, different time stamps and different weather report values.
The containers should be created in such a way that they can be used to generalise and to test for different use cases for example, if the route is changed or the numbers of buses or stops are increased (which means a greater number of both static and mobile sensors included). The data created has to be stored in text files locally so that it can be extracted from different containers for further processing.
The data should be created with a regular frequency along with keeping the accuracy of the data in mind which means that all the data is not supposed to be accurate, a percentage of incorrect data generated should also be considered so that simulation can describe the correctness percentage of data used. This has to be done in order to provide variability to the scenario plus the variability to the quality of the data created. The variability of data can be provided mostly to parameters like temperature, humidity, speed of the bus but not to the parameters like time stamps because that will create difficulty to compare and evaluate different data sets.
Another container has to be probably created which can process the data generated by the rest of the containers for data evaluation.
The ultimate goal is to make them all work in a simulation so that it can work as a testbed for Data Quality Management.