Integrated sensing and communication (ISAC) represent a critical technology for future wireless systems, enabling simultaneous environmental sensing and data transmission. In this paper, we experimentally demonstrate a novel integrated sensing and backscatter communication (ISABC) framework that uses a monostatic backscatter system for device-free indoor user localization. In contrast to most existing ISAC-related works in the literature, we consider energy harvesting backscattering tags as communication users, and our sensing application is non-intrusive, where we localise the target without tagging it. Specifically, we utilize Wireless Identification and Sensing Platform (WISP) tags and a machine learning model based on Support Vector Machines (SVM) to accurately estimate user positions within a defined sensing region. Through experiments conducted with off-the-shelf hardware, we systematically evaluate the influence of different numbers of tags, transmission power, and spatial resolution on localization accuracy and system through-put. Overall, our novel experimental demonstration indicates that the proposed ISABC framework achieves a localization accuracy that exceeds 90 % without a significant degradation in backscatter communication performance, highlighting its potential as a low-power and sustainable solution for next-generation Internet of Things (IoT) applications.
This paper evaluates the feasibility of powering low-power sensor nodes by conducting an RF power survey for ambient and dedicated emitters. The RF survey for ambient emitters was conducted at four locations of our campus -two indoors and two outdoors- from 150 MHz to 6 GHz, which covers six major frequency bands: DTT, LTE 800, GSM 900 and 1800, and Wi-Fi 2.4 GHz and 5.0 GHz. The results showed that in all cases, the RF power levels were below -30 dBm (the selected limit for powering sensor nodes), except in two cases: the GSM 900 band using a mobile phone as the emitter and the GSM 1800 band. To enhance the RF power level, a dedicated RFID reader was used and positioned at distances of 1 m and 10 m from the measurement setup, with a transmitting power range from 17.5 dBm to 39 dBm. At the standard maximum EIRP value of 35 dBm, received power levels of -14.1 dBm and -18.9 dBm were achieved at distances of 1 m and 10 m, respectively, which would allow powering sensor nodes. Finally, the results are compared with other works.
Self-healing (SH) functional polymers, composites, and blends offer innovative solutions for developing fully soft, elastic electronics, antennas, and sensors with the ability to heal themselves. In this paper, we demonstrate, for the first time, the stretching performance and SH functionality of passive UHF RFID tags fabricated of SH elastomer blends. We analyzed the functionality of the passive UHF RFID tag based on a dipole antenna with different conductor thicknesses after being bisected and stretched. After completely bisected, 82.6% of the tag's read range was regained within 24 hours, and it could be uniaxially stretched up to 75%. Our results provide important proof of concept for the development of stretchable and SH antennas, and wireless sensors. One of the most important aspects for future work is enhancing the performance of RFID tags under stretching by improving the bonding of the IC to the radiator with structural design that can also enable waterproofing, thereby making the tags suitable for wearable applications requiring washability.
Artificial Intelligence (AI) algorithms have emerged as a suitable solution for chipless RFID tag detection. However, key considerations for AI-based detection approaches include the generalization capability of the models and their deployability on computationally limited embedded digital signal processors, for their use in real-world applications. This paper presents a brief review of recently presented AI models in the literature. We evaluate the models' generalization capabilities by training and testing on measurements collected at different times, without guaranteeing the same exact set-up. Deployability is assessed based on two indicators: memory requirements and computational cost. The provided results manifest the necessity of evaluating AI models on entirely distinct datasets, with observed differences in accuracy of up to 0.2319 between testing on a subset of the training dataset and testing on a distinct dataset. The memory requirements of all the presented models indicate that they could be stored on simple embedded digital signal processors, with the largest model requiring less than 35 MB of memory. However, most of the models are computationally expensive, with the most demanding model requiring more than 825 million floating-point operations to perform an inference.
This paper investigates the robustness and performance of chipless RFID technology for identification and tracking, emphasizing its localization capabilities in dynamic and challenging warehouse environments. We examine how chipless tags placed on the ground can facilitate and enhance the real-time localization of autonomous guided vehicles. An evaluation of chipless RFID tags under harsh electromagnetic conditions, including moisture, dust, and physical obstructions, demonstrates their reliability in challenging environments. A co-polarized bistatic antenna pair, mounted on a tilted support, enhances the signal-to-noise ratio by reducing specular reflections, thereby improving detection performance. The impact of the tag's tilt orientation and translational movement on location accuracy is also analyzed. Dynamic measurements are conducted for two tag orientations: inclined placement (phi = 30 degrees) and flat placement (phi = 0 degrees). These measurements assess detection performance under motion tests, showing that the system can track tags at realistic warehouse speeds without any degradation in location performance.
