Three-dimensional quadcopter modeling and simulation using real-time brain machine interface

Abstract

The recent progresses in the area of neuro-robotics science and technology has provided the possibility to explore its practicability in aviation, especially in the area of flight controls through brain machine interfaces. This research explores the development and reliability of an outer loop electroencephalography (EEG) controller implemented on a quadcopter using steady P300 Speller technology. An inner-loop quadcopter state controller communicates through user datagram protocol (UDP) connections with the on-board computer that provides stability with an embedded controller. The creation and modeling of the inner-loop, outer-loop controllers and interface with BCI2000 occurred with the use of MATLAB/Simulink. In order to test the controllers, a user intended to complete a flight mission path by simulating the selection of a command letter with EEG signals by clicking a letter on the BCI2000 interface. Similarly, the interface and controllers were tested using EEG signals of a test subject. The test in these two cases implemented and transmitted commands to a quadcopter to perform them in real time. Prior to the flight mission, a user was briefed with the tasks to be performed and how a flashing letter would be associated with an action of the quadcopter. Observation of the flight and analysis of data acquired during the flight demonstrated a 100% accuracy of the quadcopter performing the actions coming from the brain-machine interface (BMI) software. However, this accuracy depends on the accuracy of EEG data acquisition from the brain and the selection of the letter targeted by the user. Further challenges, such as drifting and range of motion limits of the quadcopter need to be addressed to consider this research as reliable technology. Nevertheless, the modeling and simulation of a quadcopter BMI provides a promising future in the area of aviation and neuro-robotics.