👻 CASPER

CASPER (Conversational Assistant for Smart home Problem Explanation and Resolution) is an intelligent system designed to support users in creating, understanding, and improving smart home automations through both conversational and visual interaction.

The system combines conversational agents, automation analysis, and goal-oriented reasoning to detect problems in smart environments and propose personalized solutions focused on user well-being and efficiency.

🔗 Live Website: https://rulebot.isti.cnr.it/casper/

Overview

CASPER enables users to:

• Create smart home automations through a conversational agent
• Analyze existing automations to detect conflicts and undesired behaviors
• Identify problems related to high-level user goals, such as: comfort, well-being, energy saving, health
• Receive personalized suggestions to resolve detected issues
• Explore problems and solutions through both:
  • a visual interface
  • a conversational interface

The main goal of CASPER is to make smart home automation more understandable, reliable, and user-centered, especially for non-expert users.

Key Features

💬 Conversational Automation Creation

• Users can define automations using natural language
• A conversational agent guides the user during automation creation
• Lowers the technical barrier for configuring smart environments

🔍 Conflict & Activation Chain Detection

CASPER automatically analyzes the automation ecosystem to detect:

• Conflicts between automations e.g., contradictory actions triggered by the same event
• Chains of activations: sequences of automations triggering each other, directly and indirectly

🎯 Goal-Oriented Problem Detection

Instead of focusing only on low-level rule errors, CASPER reasons about high-level user goals, including:

• Comfort
• Well-being
• Energy efficiency
• Health

This allows the system to identify subtle or hidden problems that may not appear as explicit conflicts.

🧩 Personalized Solutions

For every detected problem, CASPER:

• Generates customized solution proposals
• Adapts suggestions to the user’s specific environment and goals
• Helps users make informed decisions rather than applying automatic fixes

🖥 Dual Interaction Modality

Problems and solutions are presented through:

• a visual interface, offering structured insights
• a conversational interface, providing explanations and guidance in natural language

This dual approach improves understanding and reduces cognitive load.

Design Philosophy

CASPER is built around the idea that smart environments should adapt to humans, not the other way around.

Key principles include:

• Explainability over automation opacity
• Goal-driven reasoning instead of rule-only validation
• Human-centered interaction through dialogue and visualization

209ca026456dfcfa23de8a24119a4462d09b88d4