An Intelligent Autonomous Robotic System for Precision Farming

The decline in the agricultural sector is a massive problem for Pakistan, as it relies primarily on agriculture. This decline has come about due to the outdated and traditional methods of farming that are still being employed in the country. These methods are ineffective, lack precision, cause wastage of resources, and also lack productivity. In this paper we present an autonomous robotic system which aims to improve the situation of the agricultural sector by employing precision agriculture techniques reducing human involvement resulting in increased accuracy of the tasks and consequently an improvement in the agricultural yield. The system entails a mechanical movable structure of a robot that is capable of navigating a farm field; it creates a map of the farm field by obtaining pH values at various distances. A machine learning module is employed that uses convolutional neural network to detect the presence of weed plants in the field. A spraying mechanism is constructed that is used not only for drip irrigation, saving water, but also to eliminate the detected weeds by spraying herbicides precisely on them. The overall operational support for the system is provided through Robot Operating System.