Analysis of rocket propulsion test data using multi-sensor data fusion technique

Abstract

Multi-sensor data fusion is a technique that combines data from multiple sources (sensors) to extract unique features which cannot be achieved from a single sensor. The data fusion technique is widely used in areas like sensor networks, video and image processing, robotics, intelligent system design, and many such diverse applications, to name a few. In this paper, we present a method of analyzing test data of a liquid rocket engine using the data fusion technique. A deep neural network, machine learning-based algorithm LSTM- RNN, is used for the analysis. The available sample test data from ISRO is used for the analysis. One can also predict the engine performance under operating conditions for which the test data are not available. The dynamic behavior of the engine can also be predicted through the proposed data fusion technique. The proposed method is developed in a generalized form so that it can be used for any liquid rocket engine