Using Physiological Measurements to Analyze the Tactical Decisions in Human Swarm Teams
In this work, we explore the physiological correlates with the user’s tactical decisions in a simulated search and rescue mission.
Transfer Learning of Motor Difficulty Classification in Physical Human-Robot Interaction Using Electromyography
In this study, we present a convolution neural network (CNN) model to predict motor control difficulty using surface electromyography (sEMG) from human upper-limb during physical human-robot interaction (pHRI) task and present a transfer learning approach to transfer a learned model to new subjects.
Task Difficulty Prediction in Physical Human Interaction using EEG and Deep Learning
In this study we are using the EEG recording to predict the robot instability during physical human robot interaction.
SHASTA A Simulator for Human and Swarm Team Applications
In this paper, we present SHASTA, an open-source simulation platform to study human-swarm applications
Selective Eye-gaze Augmentation To Enhance Imitation Learning In Atari Games
We propose a selective eye-gaze augmentation (SEA) network that learns when to use the eye-gaze information to improve the reinforcement learning agent’s performance.
Extracting Interpretable EEG Features from a Deep Learning Model to Assess the Quality of Human-Robot Co-manipulation
we propose a novel computational approach to increase the interpretability of results from deep learning algorithm using two popular saliency detection algorithms: integrated gradients and ablation attribution method.
Classification of Reaction Time in Simulated Multiple Robot Tele-Exploration
In this study, the physiological features of the operator are used to classify the reaction time as fast, normal and slow corresponding different levels of task difficulty. The physiological features are extracted from the eye (though eye tracking) and brain (through Electroencephalogram) from the operator performing teleoperation using two drones. Among the calculated features glance ratio and mental workload resulted in maximum classification accuracy when task type information is included.