Hardware and software system for analysis and monitoring of environmental climate and ecological changes
k.duisebekova@edu.iitu.kz
GOAL
Project Goal
To create a system for monitoring the climatic and ecological state of the environment using stationary and mobile sensors for collecting, transmitting and processing data (big data). Climate and environmental monitoring, as well as services collecting data related to climate change and variability, require the use and development of modern technical means of observation (remote and radio-electronic) and mobile and automated data transmission and processing systems.
ABOUT THE PROJECT
Abstract
The authors developed high-performance algorithms and tools for intelligent big data processing. The project created a toolkit for analyzing large volumes of data, which contributed to the further development of high-performance intelligent technologies in Kazakhstan. The successful achievement of the project results created the preconditions for increasing the competitiveness of the IT market of the Republic of Kazakhstan through the use of national scientific personnel and material and technical resources. The research results were used to meet the information needs of the scientific community and for other information and analytical purposes. The big data processing capabilities implemented in the project facilitate the work of company departments involved in customer service, marketing, competitive intelligence, external and internal corporate communications, processing of analytical financial information, etc., which confirms the practical significance of the project. The social impact of the project lies in training scientific personnel, increasing interest and broad involvement of young scientists in research, and using the scientific and applied results in the educational process.
RESULTS
Scientific Achievements
A LoRa data transmission network topology was developed. Measuring stations collect data from sensors and transmit it to the nearest gateway. Gateways act as a "transparent bridge" between the end devices of the LoRa network and the network server and do not process the transmitted data themselves. In addition to gateways, the LoRa network includes a management server that organizes data transfer from end devices to the consumer and back. The network server resolves collisions, adjusts transmitter power and data rates, monitors the battery charge of end devices and forwards the received data to the target subscribers if there are several in the network. The air quality monitoring network is planned to include 20 stationary observation posts in Almaty. Air quality observations can be carried out: under a partial program (3 times a day — 07:00, 13:00, 19:00 local time); under a full program (4 times a day — 01:00, 07:00, 13:00, 19:00 local time); or continuously. The capabilities of LPWAN networks for transmitting data from stationary and mobile posts collecting the main indicators were assessed. The developed circuit, built on a modern component base, will provide the statistical data needed to analyze an automatic network for collecting environmental and meteorological information at different times of day and in different weather conditions. The Semtech LoRa Calculator can be used to test the parameters of the data transmission module; it makes it possible to estimate bandwidth, transmission frequency, bit rate, current consumption and battery life. Google Colaboratory, a relatively new cloud service designed to simplify research in machine and deep learning, was used for visual statistics. Its advantages include free remote access to a machine with a GPU, which greatly simplifies training the deep neural networks used in the work. The air pollution data is provided as a CSV file containing 11,403 training instances. A real physical model of pollution dispersion and numerical modeling of pollution dispersion were considered. After the activation procedures are completed, the end node is ready to transmit sensor readings at a set interval. Once a connection is established, the data is converted into a payload format as hexadecimal (HEX) code.
The scientific and practical results of the project are reflected in 10 articles in journals recommended by the CQASHE of the MES RK and 11 articles in journals indexed in WoS and Scopus.
IMPLEMENTATION
Commercialization and Practical Application
Prompt decision-making based on big data processing using mathematical models of advance monitoring and forecasting leads to rational planning and use of natural resources and to cost reduction, thereby yielding a significant economic effect.
All research projects