SkyMon PIA - SkyMon Predictive Interference Analysis

Status
Completed
Status date
2021-09-08
Activity Code
6B.075
Objectives

SkyMon Predictive Interference Analysis (PIA) investigates the use of AI/ML methods for the automatic identification of interference via spectrum analysis, prediction of satellite link quality degradation caused by bad weather and the automatic identification of interference caused by non-GSO satellites.
The objective of the project is to validate the benefits of using artificial intelligence (AI) and/or machine learning (ML) algorithms to improve existing satellite interference mitigation techniques and develop new applications to predict satellite interference before it even occurs.

Challenges

The search for interferences in a provided spectrum plot can be very challenging as every spectrum plot is different as well as the nature of the interference signal. For example, an interference signal can be hidden behind another signal, it can change in frequency, in bandwidth and/or in power. 
A huge number of high-quality sample spectrum plots need to be generated to train the AI model.

Benefits

The project investigates in the feasibility of the following new technologies:

  • Automatic identification of interference situations via spectrum analysis
  • Prediction of satellite link quality degradation caused by bad weather scenarios
  • Automatic identification of interference situations caused by non-GSO satellites
Features

Automatic identification of interferences is done in several steps:

  1. Blind scan identification by AI/ML. Within this step an AI model is trained such that carrier shapes can be identified within a spectrum.
  2. Anomaly detection, 1D model. This step identifies anomalies compared to a predefined mask from a single spectrum measurement
  3. Anomaly detection, 2D model. This step identifies anomalies compared to a predefined mask from a time series of spectrum measurements.
  4. Interference classification using 1D supervised model. This step classifies the interference within a single spectrum measurement
  5. Interference classification using 2D supervised model. This step classifies the interference within a time series of spectrum measurements
System Architecture

Within the project several models for AI/ML were evaluated. The architecture is split into two parts, part one generates a huge amount of sample data (spectra) using the SkyMon carrier monitoring system, whereas part two develops, trains and verifies different AI models. A separate analysis tool annotates the generated spectra for AI/ML into interference free and interfered recordings. Further the analysis tool allows adding interference for AI/ML development and training.

Plan

The present project represents the first phase covering feasibility analysis, evaluation of different AI models and prototype development. In a second phase different use cases for applying AI in SkyMon shall be identified based on the results of the present project. The third phase covers the development, integration and test of AI models as part of the SkyMon product realizing selected use cases (expected end of 2021).

Current status

Several AI/ML models were evaluated during the scope of this project with one selected showing the highest success rate in terms of carrier identification, interference detection and classifications of about 98%. Figure 1 shows an example of interference detection by AI/ML. The lighter the colour, the more confident the AI/ML model is that an interferer is present. As can be seen in Figure 1, interference was detected very well.
  
 


Figure 1: AI/ML detection result for CW interference

The model for detection of power degradation due to bad weather scenarios show similar prediction success rates.

Prime Contractor