Using NLP to analyze requirements for Agriculture 4.0 applications

Abstract

This contribution describes the use of Natural Language Processing (NLP) methods for the lexical analysis of requirements for control, sensors, and information systems in the Agriculture 4.0 domain. The analysis is presented on an orchard 4.0 concept.The proposed orchard includes a sensor network (containing mainly measurements of hydrometeorological and soil variables), camera monitoring of conditions, and yield, support for autonomous robotic care and harvesting based on machine vision, prediction of appropriate times for interventions, etc. Requirements specification for mentioned system is written in natural language.A sentence splitting, Tokenization, Lemmatization, and POS (Part-of-Speech) tagging methods are applied to the mentioned structured requirements of the system and Use Case description. From these and by means of NLP, the candidates of classes, attributes, operations, and associations of the UML (Unified Modeling Language) class diagram are filtered and the UML model is synthesized. This paper presents the application of software engineering methods to support the development of complex heterogeneous sensors, information, and control systems.