Source-linked AI summary
A Wearable Sensor Array on Conductive Textiles for Body Surface Temperature Mapping
Amiri Onuki, Akihito Noda
TL;DR
Distributed garment temperature sensing is difficult because arrays need power and data connections across clothing without compromising wearability. This paper uses conductive textiles as a shared power-and-data bus and demonstrates an 18-node garment system with real-time mapping. The prototype achieved a 99.2 Hz refresh rate, while remaining limited in node size, power consumption, and mechanical robustness.
Problem
Garment-integrated distributed temperature arrays require power delivery and data collection across multiple locations, while wiring can complicate systems, restrict sensor placement, and reduce wearability.
Source: §1 · p. 1Method
The system uses conductive textiles as a shared bus for DC power and data, with an 18-node garment prototype and real-time temperature mapping viewer.
Source: §1 · p. 2Results
Takeaways & Limitations
The prototype demonstrates the feasibility of high-rate distributed temperature sensing on clothing and provides a basis for future garment-integrated monitoring.
Source: §7 · p. 6The current prototype is not fully optimized for node size, power consumption, or mechanical robustness, and daily-life use requires improved noise tolerance under movement and textile deformation.
Source: §6.2 · p. 5, §7 · p. 6
Abstract
from arXiv · showhide
Continuous monitoring of body-surface temperature distributions can provide useful information for wearable health and thermal-condition monitoring. However, practical daily-use wearable systems for distributed temperature sensing remain difficult to implement because multiple sensor nodes require power delivery and data collection across a garment. Conventional wiring increases system complexity, reduces placement flexibility, and can compromise the wearability of garment-integrated sensor arrays. To address this issue, we present a conductive-textile-based wearable temperature sensing system in which the textile itself serves as a shared bus for both DC power delivery and data communication. Each sensor node is implemented as a compact, battery-free circular board with a diameter of 21 mm and a weight of 0.9 g, integrating a commercial temperature sensor and custom circuitry for UART/DC superposition. The shared textile bus enables flexible node placement without individual wiring while supplying power to each node. Address-based half-duplex UART with master polling enables collision-free multi-node communication. To demonstrate the system-level capability, we implemented an 18-node sensor array on a garment worn by a mannequin and developed a real-time temperature mapping viewer. In a thermal scanning demonstration using warm air from a heat gun with a set temperature of approximately 50 {\deg}C, the 18-node array captured spatiotemporal temperature variations over the garment surface at a refresh rate of 99.2 Hz. These results show that the proposed conductive-textile bus can support real-time, high-refresh-rate distributed temperature mapping, providing a practical basis for garment-integrated body-surface monitoring and future multi-sensor wearable systems.
1. INTRODUCTION
Distributed body-surface temperature sensing can reveal localized thermal variation, but garment-integrated arrays must deliver power and collect data across deformable fabric without undermining wearability. The paper proposes conductive textiles as a shared bus for power and data, then demonstrates an 18-node garment system with real-time mapping.
1. INTRODUCTION
- Infrared thermography visualizes wide-area temperature distributions but requires an external camera, while wrist-worn devices provide practical long-term point measurements.Source: §1 · p. 1
- Skin-attachable sensing can support distributed mapping, but a cited NFC approach relied on an external antenna and scanned 65 sensors in approximately 3 s.Source: §1 · p. 1
- Garment arrays need power delivery and data collection across deformable fabric; individual wiring adds complexity, restricts placement, and can reduce connection reliability during movement.Source: §1 · p. 1
- The conductive-textile bus carries power and data together, eliminating individual wiring, on-node batteries, and wireless antennas while allowing flexible sensor placement.Source: §1 · p. 2
- An 18-node garment-integrated prototype with a real-time mapping viewer achieved a refresh rate of 99.2 Hz.Source: §1 · p. 2
2. SYSTEM ARCHITECTURE AND PRINCIPLES
The system uses conductive textiles as a shared transmission medium for both power and communication. A master coordinates sensor data collection and displays the measurements as a real-time heatmap.
2. SYSTEM ARCHITECTURE AND PRINCIPLES
- The master coordinates network communication, aggregates measurements from the sensor nodes, and visualizes them as a real-time heatmap on a PC.Source: §2 · p. 2, Figure 1 · p. 2
- Two conductive textile layers separated by insulation form a shared transmission line that connects sensor nodes electrically in parallel and supplies common power and signal paths.Source: §2 · p. 2
- UART digital signals are superimposed as voltage variations on the DC power line and recovered at the receiver by evaluating them against voltage thresholds.Source: §2 · p. 2
- The system uses baseband UART and half-duplex transmission, avoiding a carrier wave while dynamically switching transmit and receive drivers via an MCU control signal.Source: §2 · pp. 2–3
3. INTEGRATED TEMPERATURE SENSOR NODE
The integrated node combines temperature measurement with circuitry for communication over the textile bus. Its compact board also addresses self-heating by placing the temperature sensor at the board edge.
3.1. Temperature Measurement and Control Unit
- The node integrates a temperature sensor and MCU, with the MCU acquiring temperature data over I2C for transmission.Source: §3.1 · p. 3, §3 · p. 3
3.2. Communication and Signal Processing Unit
- An LC filter separates DC power from UART signals, an analog switch supports half-duplex bus access, and a hysteresis comparator converts textile voltage variations into digital pulses.Source: §3.2 · p. 3
- The sensor node’s MCU acquires temperature data via I2C and transmits it to the textile bus through the UART/DC interface circuit.Source: Figure 2 · p. 3
3.3. PCB Implementation
- The integrated circular PCB has a diameter of 21 mm and weighs 0.9 g.Source: §3.3 · p. 3
- The sensor is placed at the PCB edge to limit measurement effects from board heat, but active communication consumes approximately 110 mW and can cause self-heating that affects sensor readings.Source: §3.3 · p. 3
4. COMMUNICATION PROTOCOL
The communication protocol uses address-based master polling to coordinate half-duplex exchanges on the shared bus. Fixed-length frames and short polling intervals support high-rate array updates, though the measurement frequency can be reduced for physiological monitoring.
4.1. Addressing and Collision Avoidance
- Address IDs and master-initiated sequential polling let only the addressed sensor respond, preventing data collisions on the shared textile bus.Source: §4.1 · p. 3
4.2. Communication Flow
- During each exchange, the master sends an addressed request, the matching sensor responds, and the master collects the data before polling the next ID.Source: §4.2 · p. 4
4.3. Data Format
- Each transaction uses a 3-byte master request and a 4-byte sensor response, for a total payload of 7 bytes.Source: §4.3 · p. 4
4.4. Timing Analysis and Bus Occupancy
- At 1 Mbps, a 7-byte polling cycle takes approximately 70 µs theoretically; polling each node at approximately 560 µs yields an 18-node array refresh rate of 99.2 Hz.Source: §4.4 · p. 4, Figure 5 · p. 4
- Body-surface temperature is described as slow-varying, and a 1 Hz measurement frequency is generally sufficient to capture physiological thermal responses.Source: §4.4 · p. 4
5. SYSTEM EVALUATION
The evaluation assessed polling scalability and sensor thermal stability. Polling time increased linearly with node count, while sustained high-rate operation produced self-heating that can be reduced through lower-power operation.
5.1. Polling Rate and Scalability
- 560 µs × N polling time yielded approximately 5 ms for N = 9 and exactly 10 ms for N = 18, demonstrating linear scalability.The evaluation tested scalability before applying the network to physiological measurements.Source: §5.1 · p. 4, §5 · p. 4
5.2. Self-Heating Evaluation and Limitations
- At 99.2 samples/s, each 110 mW node rose from 24 °C to 27 °C within 30 s, 30 °C after 2 min, and nearly 32 °C after 10 min.The measurement used the highest-power operating condition, with the MCU continuously active.Source: §5.2 · p. 4
- Lowering the sampling rate, sleeping the MCU between measurements, or using lower-power MCU and op-amp models can reduce self-heating.Source: §5.2 · p. 4
6. DEMONSTRATION OF TEMPERATURE MAPPING
The 18-node garment array visualized temperature changes during a warm-air sweep, while the current prototype retains limits in scalability, wearability, and measurement accuracy.
6.1. Real-time visualization of temperature distribution
- The viewer estimates temperature at arbitrary display points by smoothly interpolating sensor readings with distance-based weights.The weights emphasize a sensor near its location and combine multiple sensor values between sensors.Source: §6.1 · pp. 4–5
- The 18-node array tracked dynamic garment-surface temperature variations during a warm-air sweep, demonstrating its temporal responsiveness rather than biological measurement accuracy.The readings included self-heating offset, and the experiment did not evaluate precision or accuracy for biological applications.Source: §6.1 · p. 5, Figure 6 · p. 5
6.2. Limitations of the Current Prototype and Future Improvements
- Scaling beyond the tested 18 nodes requires communication-protocol optimization, while standalone use requires replacing the wired USB master.Daily-life deployment also needs improved wearability and noise tolerance because movement, textile deformation, and contact can destabilize communication waveforms.Source: §6.2 · p. 5
- The prototype’s relatively large thermal mass limits transient response, and MCU self-heating degrades sensing-point temperature accuracy.The node was initially designed to ease assembly and debugging rather than prioritize miniaturization and thermal optimization.Source: §6.2 · pp. 5–6
- The authors characterize the thermal and accuracy issues as implementation artifacts, not fundamental drawbacks of the sensing concept.They propose a lower-thermal-mass sensor package, lower-power MCU, and PCB redesign to reduce thermal conduction and self-heating.Source: §6.2 · p. 6
7. CONCLUSION
The conclusion presents conductive textiles as a shared power-and-UART bus for garment-integrated temperature sensing. An 18-node prototype visualized temperature in real time at 99.2 Hz, with communication headroom for future expansion.
7. CONCLUSION
- An 18-node garment prototype demonstrated real-time heatmap visualization at 99.2 Hz, establishing feasibility of high-rate distributed sensing on clothing.The textile supplies DC power and UART communication while eliminating individual wiring and on-board batteries.Source: §7 · p. 6
- The prototype is not yet optimized for node size, power consumption, or mechanical robustness, and its communication headroom supports future expansion.The authors identify miniaturization, lower-power circuits, and more nodes or sensing modalities as future directions.Source: §7 · p. 6
- The authors envision garment-integrated networks supporting higher-resolution, more accurate, and more practical whole-body physiological monitoring.This is a future direction, not a demonstrated outcome of the current prototype.Source: §7 · p. 6