You are on page 1of 7

WORKFLOW

 When human requests assistance, robot picks component and places it near human.
 When robot cannot locate object, human gives camera movement instructions to help the
robot.
 Parallel execution in normal mode.
THIS PAPER’S WORKFLOW

Platform that contains shared


Variables. OpenBCI and TMFlow
communicate by changing the
variables values.
OUR WORKFLOW Desktop Laptop

7
Object
EEG Cap
positioning
1
2 Acquisition Live video
Cykit
5

Processing 3
6

4 Prediction

Interfacing
COMPONENTS
THIS PAPER OUR WORK
 OpenBCI GUI: Acquired via LSL layer  CyKIT: Acquires data from EMOTIV headset via Bluetooth serial
interface
 Signal processing: Python, process for SSVEP.
 Signal acquisition, processing: Python stores signals in 2 s
 Sends values 0 – 4 windows.
 CCA for frequency peak analysis  Processes signal (filter, feature extraction)
 Butterworth filter – 4 to 30 Hz  Model prediction
 Notch filter: 50 Hz  Send 0/1 to robot
 GUI-ROS: TCP connection, sends parameter values  GUI: Displays red/green circle when corr. object in camera view
 ROS publishes to listening node  Acc. to input, variables set in OpCua. Pick and place task
initiated accordingly.
 TMFlow subscriber: tracks value updates in variables
 TCP communication between all nodes.
 After process, reset all flags and variables
SSVEP features in EEG signal
Worker assistance
Component A Component B

Component C Component D
Robot assistance
Move up Move down

Move left Move right

You might also like