Interpretable Rule-Based System for Radar-Based Gesture Sensing: Enhancing Transparency and Personalization in AI

TL;DR

MIRA is introduced, a transparent and interpretable multi-class rule-based algorithm tailored for radar-based gesture detection that underscores MIRA’s ability to deliver both high interpretability and performance and emphasizes the potential for broader adoption of interpretable AI in safety-critical applications.

Abstract

The increasing demand in artificial intelligence (AI) for models that are both effective and explainable is critical in domains where safety and trust are paramount. In this study, we introduce MIRA, a transparent and interpretable multi-class rule-based algorithm tailored for radar-based gesture detection. Addressing the critical need for understandable AI, MIRA enhances user trust by providing insight into its decision-making process. We showcase the system's adaptability through personalized rule sets that calibrate to individual user behavior, offering a user-centric AI experience. Alongside presenting a novel multi-class classification architecture, we share an extensive frequency-modulated continuous wave radar gesture dataset and evidence of the superior interpretability of our system through comparative analyses. Our research underscores MIRA's ability to deliver both high interpretability and performance and emphasizes the potential for broader adoption of interpretable AI in safety-critical applications.

Figures (2)

• Figure 1: Algorithm's progression from data capture to gesture recognition.
• Figure 2: Comparing model interpretability: A) MIRA generates transparent rules. After removing the last else-default rule, personalized rules are appended. B) XAI methods provide insights into the model, highlighting feature influences. Full transparency is not achieved.