WiFi sensing technology has emerged as a promising technology for activity recognition, leveraging the Channel State Information (CSI) to capture fine-grained movement details. However, the difficulty and scarcity in collecting the required training CSI data with variations hinder the development and deployment of practical WiFi sensing systems in different settings. This paper presents MockiFi, a novel system that learns WiFi CSI data transformations across different activities of known individuals and generates CSI data for the activities of new individuals using their base activity CSI data and by mimicking transformations learned from known individuals. Our approach employs a Conditional Neural Process (CNP) to synthesize realistic activity patterns through the learning of pattern transitions using a cosine similarity-based loss function. The effectiveness of our approach is validated through extensive experiments, achieving a high cosine similarity score between the generated and real activity data, indicating the precision and reliability of the generated action fingerprints. We also show that a classifier trained on synthetic data of a new person can successfully recognize their actual activities, demonstrating zero-shot learning capabilities. These results show that MockiFi can help develop customized WiFi sensing systems without the need for collecting excessive new training data, and thus can facilitate their practical usage.

Physical therapy exercises are critically important for the rehabilitation of patients with motor deficits. While these exercises can be most effective when performed properly under the supervision of a physical therapist, it may not be a viable option for all patients. Thus, there is a growing trend towards at-home physical rehabilitation tracking systems as they can be more accessible and flexible for patients. However, existing systems mostly depend on camera and wearable based solutions, which can be costly and limited. To this end, we propose a low-cost and non-intrusive end-to-end solution using IoT-based wireless sensing devices. Our solution, Wi-PT-Hand , leverages Channel State Information (CSI) captured from ambient WiFi signals and uses Bayesian optimizers and a hierarchical deep learning model trained to recognize the prescribed hand exercises. The proposed system includes (i) segmentation of the therapy time into activity and non-activity durations, (ii) recognition of the exercise performed in an activity segment, and (iii) counting of the number of repetitions of the exercise performed within that segment. Extensive experimental results show that the proposed system is robust and performs well in various real life scenarios, and thanks to the lightweight design it can work on low-resource edge devices properly.

Maintaining independence in daily activities and mobility is critical for healthy aging. Older adults who are losing the ability to care for themselves or ambulate are at a high risk of adverse health outcomes and decreased quality of life. It is essential to monitor daily activities and mobility routinely and capture early decline before a clinical symptom arises. Existing solutions use self-reports, or technology-based solutions that depend on cameras or wearables to track daily activities; however, these solutions have different issues (e.g., bias, privacy, burden to carry/recharge them) and do not fit well for seniors. In this study, we discuss a non-invasive, and low-cost wireless sensing-based solution to track the daily activities of low-income older adults. The proposed sensing solution relies on a deep learning-based fine-grained analysis of ambient WiFi signals and it is non-invasive compared to video or wearable-based existing solutions. We deployed this system in real senior housing settings for a week and evaluated its performance. Our initial results show that we can detect a variety of daily activities of the participants with this low-cost system with an accuracy of up to 76.90%.

Since the rapid growth of e-commerce, the number of package deliveries to residential doorsteps has significantly increased. However, the convenience of online shopping has also led to a rise in package theft, resulting in frustration and financial loss for both consumers and companies. While various commercial solutions are available to address package theft, they often have considerable drawbacks, such as high ongoing costs and privacy concerns. In response to these issues, we introduce Wi-Alert, a budget-friendly WiFi sensing-based package detection system. Our solution analyzes Channel State Information (CSI) acquired from ambient WiFi signals and employs deep learning models trained to identify movements at the front door. The system accurately distinguishes between various actions, such as knocking, lingering visitors, package deliveries, and package theft. These actions trigger real-time alerts, enabling users to monitor their front door activity and swiftly respond to security threats. Through real-world experiments, we demonstrate the versatility and practical application of the system in diverse residential settings, including houses and apartment buildings. Our solution offers a convenient and economical approach to enhancing package security, providing peace of mind to individuals receiving deliveries.

Physical therapy (PT) exercises are critically important for the rehabilitation of patients with motor deficits. While rehabilitation exercises can be most effective when performed properly under the supervision of a physical therapist, it can be costly in terms of several aspects and may not be a viable option for all patients. At-home systems offer more accessible and less costly solutions to patients while also providing flexibility in scheduling prescribed exercises. However, current systems mostly depend on camera based solutions that have limitations (i.e., deployment cost, requiring patients to be in the sight of camera, potential privacy violations) or wearable solutions that are cumbersome and intrusive. To this end, in this paper, our goal is to leverage the WiFi infrastructure available in most indoor locations (i.e., homes, apartments, nursing homes, etc.) for tracking the exercises prescribed to patients during their rehabilitation. Our solution, Wi-PT, is based on the analysis of Channel State Information (CSI) captured from ambient WiFi signals, and uses deep learning models trained to recognize the prescribed physical therapy exercises. Through our experiments, we show that the proposed solution can successfully recognize different types of physical therapy exercises such as hand and finger movements, limb movements and movements performed with exercise equipment. Moreover, we show that our system can recognize the person performing different activities and can identify when they are at rest or actively performing an exercise.