Tagnoo: Enabling Smart Room-Scale Environments with RFID-Augmented Plywood

Document Title

Tagnoo: Enabling Room-Scale Smart Environments with RFID-Augmented Plywood

Document Information

• Domain: Human-Computer Interaction, Computational Materials, Smart Environments
• Keywords: Smart environments, computational materials, RFID, sensing technology, machine learning, user activity detection, furniture manufacturing, technology democratization, sustainability

Research Background and Problem Statement

• Identified Problems and Challenges:
  • Most current embedded computational materials require batteries and wired connections, making them unsuitable for power-free and wireless communication environments.
  • Materials fail to maintain functional integrity during traditional operations such as cutting and assembly.
  • Existing methods often demand technical expertise, making them inaccessible to non-specialized users like carpenters.
• Importance of the Problem:
  • As computational technology becomes more prevalent, integrating sensing capabilities into everyday materials (e.g., wood) can facilitate the realization of smart environments and enhance convenience in daily life.
  • Such integration must consider cost, operability, and compatibility with existing workflows.
• Research Motivation and Related Work:
  • Inspired by the vision of constructing room-scale smart environments, leveraging smart materials (e.g., sensor-embedded wood) to improve efficiency in living and working spaces.
  • Prior studies like "Capacitivo" have demonstrated the potential of interactive materials but face limitations in applicability and cost.
  • Proposing a novel method to integrate RFID into plywood, enabling the development of more versatile and practical computational materials.

Solution

• Proposed Approach:
  • Tagnoo: A smart plywood embedded with RFID tags capable of detecting objects and user activities. This material is battery-free, low-cost, and suitable for furniture and infrastructure applications.
  • Material design prioritizes compatibility with traditional workflows and resilience to common processing operations such as cutting and painting.
• Innovations:
  • Introducing a battery-free, low-cost smart material solution compatible with traditional workflows.
  • Utilizing a 2D grid of embedded RFID tags for object recognition and user activity tracking.
  • Optimizing the integration of RFID tags with wood to enhance durability and sensitivity.
• Implementation Steps:
  • Designing and experimenting with RFID tag types, grid density, embedding depth, and adhesive strategies.
  • Creating various Tagnoo materials with different densities for diverse scenarios (e.g., furniture surfaces, flooring).
  • Building a smart office environment comprising tables, chairs, bookshelves, and flooring.
  • Generating low-resolution heatmaps from sensor data and applying machine learning algorithms for activity classification.

Research Outcomes

• Specific Results:
  • Tagnoo retained sensing capabilities after cutting, painting, and assembly.
  • Experimental validation demonstrated its ability to accurately detect 18 types of everyday objects and user activities, achieving over 90% recognition accuracy.
• Advantages Over Existing Methods:
  • Enhanced the likelihood of materials retaining functionality during processing.
  • Eliminated the need for batteries and complex wired connections, lowering the entry barrier for ordinary carpenters.
  • No additional technical configuration required; users do not need specialized electrical engineering knowledge.
• Experimental Results:
  • General model accuracy: 93.9% (SD=5.9), with some misclassification in book position detection and floor activity recognition.
  • Individual furniture model accuracy: table 95.3%, chair 94.2%, bookshelf 91.2%, floor 91.6%.
• Limitations and Future Directions:
  • Limited to static scenarios, unable to handle dynamic furniture layouts or simultaneous detection of multiple objects.
  • Current system sampling rate is low, unable to capture fine-grained user activities (e.g., writing).
  • Exploring Tagnoo's recycling mechanisms to address potential environmental impacts.
  • Future research will consider using deep learning to expand system capabilities, achieve real-time multi-object detection, and optimize sensing range.