Body and Emotional State Inference

Healthcare Robot

• Main Goals

To develop a common robot S/W middleware for target healthcare robots (chair-bot, ride-bot, and lifecare-bot).

To design and implement an inference engine for body and emotional states based on several bio-signals.

To develop a communication framework to interface with a variety of bio-signal sensors such as ECG, temperature sensor, and blood pressure sensor, etc.

• System Overview

• Applied Bio-signals

Electrocardiogram (ECG), electroencephalogram (EEG)

Heart rate variability (HRV), galvanic skin response (GSR), photoplethysmogram (PPG)

Blood pressure, temperature, body fat

• Bio-signal Analysis

Time-domain analysis

Time series analysis

Frequency-domain analysis

Spectrum analysis

Coherence analysis

Sensor fusion

• Classification of Emotion States

Mehrabian and Russell’s PAD (pleasure, arousal, dominance) model

• Body and Emotion Inference

Emotion inference procedure

Learning models and algorithms for emotional state classification

Artificial neural networks (ANN)

Support vector machine (SVM)

• S/W Component Configuration