Environmental Sensing Hub with Machine Learning Prediction of Ideal Drilling Site

Abstract

This project report presents the development of an environmental sensing hub for a rover, focusing on gathering comprehensive data related to ambient atmospheric pressure, temperature, humidity, altitude, soil moisture, inclination level, stability, and object temperature. The collected data is utilized to train a machine learning program that predicts the viability of drilling sites. A graphical user interface (GUI) is developed to showcase real-time sensed data in a user-friendly manner. The project aims to create a reliable and efficient system for assessing drilling site suitability based on real-time environmental data. By employing machine learning techniques, accurate predictions are made to assist in decision-making during drilling operations. The interdisciplinary approach incorporates robotics, environmental sensing, machine learning, and user interface design to achieve project objectives. The environmental sensing hub consists of various sensors strategically placed on the rover for comprehensive data collection. Sensors measure ambient atmospheric pressure, temperature, humidity, altitude, soil moisture, inclination level, stability, and object temperature. The collected data is transmitted for further analysis. A machine learning program processes the collected data, utilizing supervised training methods such as decision trees, support vector machines, and neural networks. By training the model with labelled data from known sites, correlations between environmental parameters and drilling site outcomes are learned. This enables accurate predictions of drilling site suitability based on environmental conditions. A user-friendly GUI is developed to visually represent real-time sensed data. The GUI presents data through graphs, charts, and maps, allowing users to monitor environmental conditions and make informed decisions regarding drilling site selection. Real-time updates ensure access to the most recent data, and features for data logging and exporting facilitate further analysis and documentation. The developed system undergoes rigorous testing and validation. The sensing module is calibrated against known reference values to ensure accurate measurements. The machine learning program is evaluated using historical and real-time data to assess prediction performance. The GUI is tested for usability and functionality, incorporating user feedback for iterative improvements. Results demonstrate successful integration of the environmental sensing hub into a rover system. The sensing module accurately collects data, processed by the machine learning program for predicting drilling site viability. The GUI provides real-time visualizations, enhancing decision-making during drilling operations.