Improving GPS based distance measurement accuracy using Machine Learning: an empirical study

Shortage of workforce is one of the most critical challenges for agricultural domain. Automation of farm machinery can mitigate this to a greater extent. However, in a middle-income country like India, the cost of automation has to be minimized in order to encourage its wider adoption. An attempt has been made in this study to design a machine learning based system which improves the distance measurement accuracy of a low-cost GPS receiver that is deployed for vehicle localization and tracking in an autonomous harvester to particularly harvest the onion. Nine different machine learning based models are evaluated for their effectiveness in improving the distance measurement accuracies for step sizes ranging from 2-10 meters. Results indicate that the proposed machine learning based system is capable of improving the measurement accuracy of GPS system under test.