ACISLab | ITSON ← Back to Home

Cognitive-Affective Frameworks for Intelligent Agents

We design and implement extensible software frameworks alongside computational architectures and 3D simulation platforms for emotionally intelligent virtual agents. Below is an overview of our core research projects and open software tools developed at ACISLab.

Research Projects

Foundational and applied projects investigating computational models of emotion, cognitive modulation, and affective sensing.

Core Architecture Appraisal Theory & Virtual Agents

Computational Models of Emotions for Virtual Agents

Translating psychological and neuroscientific frameworks—such as cognitive appraisal theory—into formal computational models that endow autonomous agents with plausible emotional and social behaviors. These models serve both to enhance human-agent interaction and as simulation testbeds to explore cognitive-affective processes.

  • • Design Patterns & Modularity: Reusable architectural blueprints separating perception, appraisal, emotion elicitation, and behavioral expression.
  • • Cognitive-Affective Simulation: Modeling emotional bias and the influence of affective states (such as anxiety) on agent decision-making.
Appraisal Theory Cognitive Architectures Behavioral Plausibility
Computational Model of Emotion Architecture
Figure 1. Modular appraisal and cognitive-affective modulation loop in autonomous agents.
Affective HCI Multimodal Sensing & IoT

Context-Aware Emotion Recognition in Smart Environments

Developing multimodal sensing architectures that combine physiological signals, behavioral cues, and environmental context within Internet of Things (IoT) ecosystems to infer user affective states in real time and adapt system assistance empathetically.

  • • Multimodal Fusion: Integrating wearable biometric sensors, interaction logs, and situational context for robust affect detection.
  • • Vulnerable Populations: Tailoring non-intrusive affective sensing to support older adults, people living with dementia, and caregivers.
Multimodal Sensing Affective IoT Adaptive Interfaces
Context-Aware Emotion Recognition in IoT Environments
Figure 2. Multimodal data fusion pipeline for real-time affective state estimation.

Software Frameworks & Simulation Platforms

Interoperable tools and simulation testbeds built at ACISLab to standardize the development, integration, and evaluation of affective agents.

Software Framework InFra v1.0

InFra: Integrative Framework for Computational Models of Emotion

InFra Integrative Framework Architecture
Figure 3. InFra cross-language interoperability layer for cognitive and affective modules.

An extensible software framework that enables researchers to decouple, connect, and evaluate independent computational models of emotion alongside external cognitive architectures and virtual agent environments across different programming languages.

Interoperability Modular API Agent Architectures
Simulation Testbed 3D SimEnv

3D Virtual Agent Simulation & Evaluation Platform

3D Virtual Agent Simulation Environment
Figure 4. Interactive 3D simulation testbed for validating embodied emotional behaviors.

An interactive 3D simulation environment designed to test and visualize embodied virtual agents in controlled social and decision-making scenarios, capturing behavioral telemetry and supporting empirical human-agent interaction studies.

3D Virtual Agents Behavioral Simulation Empirical Evaluation

Interested in Thesis Projects or Software Collaboration?

We welcome undergraduate, MSc, and PhD students as well as academic collaborators interested in computational emotion models, affective architectures, and agent simulation.