Weather Data Transmission Driven by Artificial Neural Network

Abstract

Nowadays, the trend of big data that can be describe as a massive volume and unstructured data have become more complicated because of its difficulty to be process using traditional database and software techniques. Due to increase in size of data, there also demand for big bandwidth for the transmission of big data. Data transmission is important in communication in providing information at different location. In this project we focus on big data transmission in the context of weather data. Weather data is important for meteorologist as it helps them to make weather prediction. Real time weather prediction is really important as it would help in making quick decision to react with the environment and planning for our daily activities. The purpose of this project is to develop a real time and low bandwidth usage for weather data transmission driven by an artificial neural network perform weather forecast using Adaptive Forecasting Model. This project seeks an application context offshore because the data transmission from offshore to onshore is very costly and requires high usage of network bandwidth. Other than that, offshore weather can change rapidly and cause offshore activity to be delayed.
This research is important as it would help in making faster decision for oil and gas daily activity and avoid loss of human life in natural disaster in the fastest way. In this research we propose a neural network based prediction to control weather data transmission. This system mainly has several parts which consist of data gathering module, neural network prediction module and transmission module. The data gathering module is developed using ODroid weatherboard sensor while the prediction module is developed using an artificial neural network technique and algorithm to make prediction. The transmission module is based on TCP/IP protocol. The prediction module works by using the data captured by sensor to be processed using neural network prediction. If the neural network predict major changes in weather condition or detect bad weather, it will send the weather data and result to the server so that the data can be published and alarm meteorologist and people on the oil platform to avoid loss of human life or another contingency plan on the current oil and gas activity. But if there are no major changes in weather prediction result, the module will only update the server on a daily basis. In this project, we consider to capture atmospheric pressure, temperature, humidity, visibility, Infrared index, altitude and UV index. Real time processing and transmission weather data will show improvement in predicting weather and saving the network bandwidth transmission.