Sridhar, Nithya (TCS Research and Innovation), N S, Abhinay (Tata Consultancy Services), Bodduluri, Chaithanya Krishna (Tata Consultancy Services), Das, Kaushik (TATA Consultancy Service), Maity, Arnab (Indian Institute of Technology Bombay)

A Systematic Modelling Framework for Commercial Unmanned Hexacopter Considering Fractional Order System Theory

Scheduled for presentation during the Regular Session "Manned/Unmanned Aviation and Testbeds" (FrD3), Friday, September 4, 2020, 16:50−17:10, Edessa

2020 International Conference on Unmanned Aircraft Systems (ICUAS), September 1-4, 2020 (Postponed from June 9-12, 2020), Athens, Greece

This information is tentative and subject to change. Compiled on April 24, 2024

Abstract

This research work proposes a systematic methodology to model Commercial-Off-The-Shelf Unmanned Aerial Vehicle (UAVs) that belong to the class of multicopters. The modelling framework consists of meticulous analysis of frequency domain and time domain data acquired from the hexacopter. The aim of this work is to model a commercially available UAV to enable the development of an efficient outer loop controller, specific to the intended application. This framework makes use of sine and step response data of the hexacopter’s velocities in all the three planar axes corresponding to the global frame of reference. The framework considers the fractional-order model as a potential candidate for the UAV. Hence, it consists of a systematic method to determine the order of the model in integers or fractions, and its parameters. In the modelling process, control theory and fractional order theory have been sufficiently used to best correlate theoretical proposition with experimental outcomes. Also, real world factors and system constraints have been explicated and considered appropriately. The model obtained using the proposed framework has an average error of less than 5% when compared to the real system’s response. Additionally, the ill effect of considering a near integer order model for a fractional order system is demonstrated to show the importance of this modelling framework. Further, a linear controller is implemented for the system using the estimated model, and corroborated with real world results.