Expert system for remote process automation in multiplatform servers, through human machine conversation

Abstract

Web portals and instant messaging applications have incorporated the use of bots and chatbots online in an accelerated way, computer programs that emulate the behavior of a human being to perform tasks, such as holding a conversation and answering the questions asked or executing specific orders made by users through a human-machine interaction; The present technical research consists of developing an Expert System for process automation in multiplatform servers, using a chatbot and bot at the same time, through artificial intelligence technology and virtual assistance (human-machine conversation), transmitting natural language to a multiplatform server, without restriction of its specific geolocation. The process began with the configuration of the multiplatform server to the telegram instant messaging server (human), then optimized the API predefined by this pattern (chatbot), this research is based specifically on the management of telegram messaging and their adaptation to the automatic management of the bot, facilitating the execution of corresponding processes in the remote automation of servers (multiplatform), a base of data in MYSQLite, allow the interaction between the created chatbot (machine) and the user telegram; after that the bot automatically manages the sending of statistics in real time to perform the requests made by the telegram user, using a matlab plug-in called Thinspeak that allows to obtain a graphical report of using the resources of Operating System. A study and validation tests were carried out on the use and performance of chatbot and bot implemented in GNU / Linux, Windows and MacOS servers from different platforms (Android, IPhone / IPad, Windows Phone, Pc's), Measuring the response time of RAM consumption, power-on time, hard disk consumption (HDD), CPU resource consumption, server-connected user monitoring, network connectivity status, remote server management and processor temperature, reflecting the results on the ThingSpeak open platform.