To protect sensitive information on smartphones, state-of-the-art (SoA) studies exploit the built-in camera to capture PPG signals from fingertips as a hard-to-forge biometric. However, those studies do not provide a comprehensive analysis to optimize the camera parameters and finger pressure, leading to distorted and unstable PPG signals that degrade the authentication performance. To overcome these limitations, we propose the CamPressID framework. First, we analyze various camera parameters and optimize their configuration to obtain PPG signals with a high signal-to-noise ratio. Second, we investigate different finger pressures to identify the best pressure for every subject, in order to avoid signal distortion. To evaluate the performance of CamPressID, we collect a diverse dataset with 58 subjects. Our evaluation results show that CamPressID can improve the average balanced accuracy (BAC) by 10%. Moreover, the BAC reaches 90%, which is similar to the accuracy reported in the SoA using a dedicated PPG sensor for authentication.@en

Cardiac patterns are being used to obtain hard-to-forge biometric signatures and have led to high accuracy in state-of-the-art (SoA) identification applications. However, this performance is obtained under controlled scenarios where cardiac signals maintain a relatively uniform pattern, facilitating the identification process. In this work, we analyze cardiac signals collected in more realistic (uncontrolled) scenarios and show that their high signal variability (i.e., ir-regularity) makes it harder to obtain stable and distinct user features. Furthermore, SoA usually fails to identify specific groups of users, rendering existing identification methods futile in uncontrolled scenarios. To solve these problems, we propose a framework with three novel properties. First, we design an adaptive method that achieves stable and distinct features by tailoring the filtering spectrum to each user. Second, we show that users can have multiple cardiac morpholo-gies, offering us a much bigger pool of cardiac signals and users compared to SoA. Third, we overcome other distortion effects present in authentication applications with a multi-cluster approach and the Mahalanobis distance. Our evaluation shows that the average balanced accuracy (BAC) of SoA drops from above 90% in controlled scenarios to 75% in uncontrolled ones, while our method maintains an average BAC above 90% in uncontrolled scenarios.@en

Sensing people with mmWave radars is gaining significant attention. This growing interest is due to two factors: radar monitoring provides more privacy than camera-based alternatives, and radio waves are not as easily blocked as light waves. Most mmWave studies, however, have three common characteristics. They are done indoors, without protecting the sensor (no casing), and the evaluation is performed for short periods of time. To assess the suitability of mmWave sensing in realistic outdoor scenarios, we deploy two nodes to track the flow of pedestrians over a period of three months. This longterm deployment provides three main contributions. First, we follow a detailed process to design a casing that can protect the sensors from harsh environmental conditions. Second, we install our nodes close to a set of cameras that were already deployed in the area. To compare the performance of both types of sensors, we propose a framework that considers the different coverage patterns of cameras and radars. Third, the time frame of our evaluation considers various types of weather, from sunny days to rainy and windy. Our results indicate that mmWave sensors need to be explored further outside the comfort zone of indoor spaces. To the best of our knowledge, this is the first long-term study assessing the reliability of radar sensors in the 60 GHz ISM band.@en

Cardiac patterns are being used to provide hard-to-forge biometric signatures in identification applications. However, this performance is obtained under controlled scenarios where cardiac signals maintain a relatively uniform pattern, facilitating the identification process. In this work, we analyze cardiac signals collected in more realistic (uncontrolled) scenarios and show that their high signal variability makes them harder to obtain stable and distinct features. When faced with these irregular signals, the state-of-the-art (SOTA) reduces its performance significantly. To solve these problems, we propose the CardioID framework1 with two novel properties. First, we design an adaptive method that achieves stable and distinct features by tailoring the filtering process according to each user’s heart rate. Second, we show that users can have multiple cardiac morphologies, offering us a bigger pool of cardiac signals compared to the SOTA. Considering three uncontrolled datasets, our evaluation shows two main insights. First, while using a PPG sensor with healthy individuals, the SOTA’s balanced accuracy (BAC) reduces from 90–95% to 75–80%, while our method maintains a BAC above 90%. Second, under more challenging conditions (using smartphone cameras or monitoring unhealthy individuals), the SOTA’s BAC reduces to values between 65–75%, and our method increases the BAC to values between 75–85%.@en

Positioning based on visible light is gaining significant attention. But most existing studies rely on a key requirement: The object of interest needs to carry an optical receiver (camera or photodiode). We remove this requirement and investigate the possibility of achieving accurate positioning in a passive manner—that is, without requiring objects to carry any optical receiver. To achieve this goal, we propose PassiveVLP, in which we exploit the reflective surfaces of objects and the unique propagation properties of LED luminaires. We present geometric models, a testbed implementation, and empirical evaluations to showcase the opportunities and challenges posed by this new type of passive positioning. Overall, we show that our PassiveVLP can track with high accuracy (a few centimeters) a subset of an object’s trajectory, and it can also identify passively the object’s ID.@en

Considering that the IoT is giving rise to new sensing infrastructures with stringent requirements, we are posed with new trade-offs. On one hand, there is a need to develop systems that have a low cost, complexity, and ecological footprint, while being privacy-aware. On the other hand, despite these constraints, there is a need to still maintain a high accuracy. Since deploying a large number of precise but expensive sensors is not an option, the research community is investigating alternatives that require either deploying low-cost sensors or re-purposing existing sensors for other applications. For example, instead of using cameras for indoor monitoring, which infringes privacy, a new generation of low-cost mmWave radars are used for that purpose; and instead of deploying hospitals in remote areas, researchers are re-purposing smartphone cameras to perform health checks such as blood pressure. The main challenge solved by these studies is achieving good accuracy with low-cost or re-purposed sensors. This thesis follows that same line of research: expanding the pervasiveness of reflective sensing systems in the Visible Light, Infrared, and Microwave spectra. To contribute to tackling that challenge, we need to answer the following research question: What design alternatives are available to approach the performance of high-end sensors with either low-cost or re-purposed sensors? This thesis argues that the above research question can be explored through different options. In conventional approaches, the gap is mainly filled by high-end sensors purposely designed for the required task and some methods on top of the sensor to perform data processing. If, on the other hand, the system relies on low-cost or re-purposed sensors, the sensing gap is exposed. The sensing gap can be investigated at two levels. The first level is to solely quantify the gap, without trying to bridge it. From a research perspective, such an approach allows exposing the magnitude of the problem raised by using low-power or re-purposed sensors. The second level is to bridge the gap in Chapter 2, 3, and 4. For this approach, we identify three alternatives. The first alternative is to enhance the methodology in Chapter 2. For scenarios where the sensor is re-purposed, an enhanced methodology could overcome the low quality of the received signals. Advanced signal processing and machine learning techniques are particularly valuable for this alternative. The second alternative is to enhance the re-purposed sensor in Chapter 3. The best example of this alternative is the use of smartphones for health-related applications. Several studies show that optimizing different camera parameters or performing minor physical modifications to the microphone can allow checking auditory or cardiovascular issues. The third alternative is to enhance the object in Chapter 4. Given that reflective sensing is fundamentally determined by the properties of the object's external surface, for some applications, it may be possible to perform minor modifications to the object's surface to facilitate sensing. @en