CARE: Contrastive Alignment for ADL Recognition from Event-Triggered Sensor Streams

TL;DR

CARE tackles ADL recognition from event-triggered ambient sensors by bridging sequence- and image-based representations through supervised cross-view contrastive alignment. It introduces time-aware sequence encoding, spatially-informed image representations, and the Sequence–Image Contrastive Alignment (SICA) objective, all trained in an end-to-end CARE framework that jointly optimizes representation learning and classification. The method achieves state-of-the-art results on CASAS Milan, Cairo, and Kyoto7 datasets and demonstrates robustness to sensor malfunctions and layout changes, highlighting its practicality for real-world smart-home deployments. By aligning temporal and spatial cues in a shared latent space, CARE delivers noise-resilient, discriminative embeddings and a deployment-friendly one-stage training pipeline. The approach suggests broader applicability of cross-view contrastive alignment to multimodal sensing domains beyond ADL recognition.

Abstract

The recognition of Activities of Daily Living (ADLs) from event-triggered ambient sensors is an essential task in Ambient Assisted Living, yet existing methods remain constrained by representation-level limitations. Sequence-based approaches preserve temporal order of sensor activations but are sensitive to noise and lack spatial awareness, while image-based approaches capture global patterns and implicit spatial correlations but compress fine-grained temporal dynamics and distort sensor layouts. Naive fusion (e.g., feature concatenation) fail to enforce alignment between sequence- and image-based representation views, underutilizing their complementary strengths. We propose Contrastive Alignment for ADL Recognition from Event-Triggered Sensor Streams (CARE), an end-to-end framework that jointly optimizes representation learning via Sequence-Image Contrastive Alignment (SICA) and classification via cross-entropy, ensuring both cross-representation alignment and task-specific discriminability. CARE integrates (i) time-aware, noise-resilient sequence encoding with (ii) spatially-informed and frequency-sensitive image representations, and employs (iii) a joint contrastive-classification objective for end-to-end learning of aligned and discriminative embeddings. Evaluated on three CASAS datasets, CARE achieves state-of-the-art performance (89.8% on Milan, 88.9% on Cairo, and 73.3% on Kyoto7) and demonstrates robustness to sensor malfunctions and layout variability, highlighting its potential for reliable ADL recognition in smart homes.

Figures (10)

• Figure 1: Contrastive Alignment for ADL Recognition from Event-Triggered Sensor Streams (CARE) Framework.
• Figure 2: Frequency-based Event Filtering on Cairo
• Figure 3: Sequence Representation.
• Figure 4: Image Representation.
• Figure 5: Heatmap of BiLSTM and ResNet18 on Cairo.