Machine learning based prediction of Anamalous Radio propagation

Project Title

Machine learning based prediction of Anamalous Radio propagation

Project Area of Specialization Artificial IntelligenceProject Summary

An abnormal propagation echo is one of the echoes that are not precipitation-induced because of abnormal refraction of a radar beam. In change in radio propagation conditions due to unusual distribution of atmospheric variables with height in the atmosphere. The propagation of radio waves over distances greater than normal surface waves can reach, beyond those normally reached by surface waves. In anomalous propagation, the radar beam propagates from the upper atmosphere toward the ground when calm, stable conditions, commonly caused by super refraction occurring in a temperature inversion, are present. At normal atmospheric conditions, the warmest air is closest to the Earth's surface. The air gradually becomes cooler as it rises in altitude. Occasionally, however, an unusual situation develops in which warm layers accumulate over layers of cool air. The temperature inversion occurs when cool air ducts are sandwiched between two layers of warm air, or between the surface of the Earth and the warm air. VHF and UHF transmissions may be propagated far beyond normal line-of-sight distances if an antenna is extended into a duct of cool air, or if the radio wave enters the duct at a very low angle of incidence. It is not unusual for VHF and UHF television signals from distant stations to be widely received when ducts are present because of temperature inversions. It is possible to travel long distances because warm and cool air have different density and refractive properties. A radio wave is refracted back toward Earth when it suddenly changes density when it enters warm air above a duct. It is then reflected or refracted upward when it strikes the Earth or a warm layer below the duct, then proceeds on through the duct with several bounces. An example of radio waves propagating by ducting is shown in the figure 1.

Project Objectives

1. To conduct a survey of different machine learning algorithms to identify the relevant models

1. To utilize relevant machine learning algorithms for assisting in the propagation modeling of anomalous behavior of propagation
2. To recommend the most suitable ML based technique for prediction of propagation anomalous

Project Implementation Method

In our Approach we aim to utilize a relevant machine learning model which is expected to predict the anomalies in the radio signal propagation in a more accurate manner than current models. Machine learning analyzes the background data and studies the pattern which we call training that we will provide. Now based on this data the machine learning model will know that the anomaly is coming in the propagation. We will select the best solution for predicting radio propagation anomalies using different machine learning algorithms and deep learning techniques.

We will use data from radio propagation measurements and will preprocess it for training. After preprocessing the data, we will use different machine learning algorithms to achieve the highest accuracy for anomaly detection in radio propagation. We will then compare the predictions with those made by conventional/standard propagation models.

                                      Figure: 2 Flow chart of Methodology

Benefits of the Project

 It has major contribution in the field of radio propagation modelling. Radio propagation scenarios and challenges continue to grow, with the introduction of new services and frequencies. To date, as per information available in literature, ML based prediction models exist for normal propagation but not for anomalous propagation. So, this work is expected to focus on un unresearched area of radio propagation and be possibly the first of its kind for Anomalous propagation. It will help the Radio Network Design and Optimization Engineers to predict and model the scenarios of anomalous propagation during network design phase. It will help us to improve radio propagation by predicting the anomalies in transmission.

Technical Details of Final Deliverable

Activity	Oct-Nov            2021	Dec21-Jan-22	Feb-Mar.  2022	Mar.-Apr 2022	Apr-May 2022	May-Jun 2022	Jun-July 2022	July 2022
Literature Review, Machine Learning Study and Study of Relevant propagation Models
Comparison of Predictions of Relevant Propagation Models & Comparison of Relevant ML Algorithms
Training and Testing of ML Model
Comparative Analysis of ML Prediction with Standard Predictions & Model Refinement
Re-analysis after refinement
Thesis work & Dissemination of study
Correction in thesis

Activity

Literature Review, Machine Learning Study and Study of Relevant propagation Models

Comparison of Predictions of Relevant Propagation Models & Comparison of Relevant ML Algorithms

Training and Testing of ML Model

Comparative Analysis of ML Prediction with Standard Predictions & Model Refinement

Re-analysis after refinement

Thesis work & Dissemination of study

Correction in thesis

Final Deliverable of the Project Software SystemCore Industry TelecommunicationOther Industries IT Core Technology Artificial Intelligence(AI)Other Technologies Internet of Things (IoT)Sustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Elapsed time in (days or weeks or month or quarter) since start of the project	Milestone	Deliverable
Month 1	Collection of literature Study of Literature Analysis Preparation of Proposed Scheme Implementation of Scheme/Model Hardware Testing Analysis Hardware Simulation Data collecting Data Analysis	Collected Literature Review Completed Analyzed Completed Implemented Tested Simulated Collected Analyzed Completed