Falling events are a global health concern with short- and long-term physical and psychological implications, especially for the elderly population. This work aims to monitor human activity in an indoor environment and recognize falling events without requiring users to carry a device or sensor on their bodies. A sensing platform based on the transmission of a continuous wave (CW) radio-frequency (RF) probe signal was developed using general-purpose equipment. The CW probe signal is similar to the pilot subcarriers transmitted by commercial off-the-shelf WiFi devices. As a result, our methodology can easily be integrated into a joint radio sensing and communication scheme. The sensing process is carried out by analyzing the changes in phase, amplitude, and frequency that the probe signal suffers when it is reflected or scattered by static and moving bodies. These features are commonly extracted from the channel state information (CSI) of WiFi signals. However, CSI relies on complex data acquisition and channel estimation processes. Doppler radars have also been used to monitor human activity. While effective, a radar-based fall detection system requires dedicated hardware. In this paper, we follow an alternative method to characterize falling events on the basis of the Doppler signatures imprinted on the CW probe signal by a falling person. A multi-class deep learning framework for classification was conceived to differentiate falling events from other activities that can be performed in indoor environments. Two neural network models were implemented. The first is based on a long-short-term memory network (LSTM) and the second on a convolutional neural network (CNN). A series of experiments comprising 11 subjects were conducted to collect empirical data and test the system’s performance. Falls were detected with an accuracy of 92.1% for the LSTM case, while for the CNN, an accuracy rate of 92.1% was obtained. The results demonstrate the viability of human fall detection based on a radio sensing system such as the one described in this paper.

The growing elderly population living independently demands remote systems for health monitoring. Falls are considered recurring fatal events and therefore have become a global health problem. Fall detection systems based on WiFi radio frequency signals still have limitations due to the difficulty of differentiating the features of a fall from other similar activities. Additionally, the antenna orientation has not been taking into account as an influencing factor of classification performance. Therefore, we present in this paper an analysis of the classification performance in relation to the antenna orientation and the effects related to polarization and radiation pattern. Furthermore, the implementation of a device-free fall detection platform to collect empirical data on falls is shown. The platform measures the Doppler spectrum of a probe signal to extract the Doppler signatures generated by human movement and whose features can be used to identify falling events. The system explores two antenna polarization: horizontal and vertical. The accuracy reached by horizontal polarization is 92% with a false negative rate of 8%. Vertical polarization achieved 50% accuracy and false negatives rate.

Human fall detection has become one of the principal applications of smart healthcare. Several proposals based on sensors integrated into consumer electronic devices have been developed for fall detection. One of the most recent proposals is to employ radio signals of commercial WiFi devices as sensing probes. This radio sensing (RS) approach does not require wearing sensors and is viable for quick deployment due to the ubiquity of communication systems. However, the accuracy rate of WiFi-based fall detection systems is affected by several factors, such as the effects of thermal noise. In this paper, we present the idea of a frequency-diversity enhanced RS system for human fall detection based on the analysis of Doppler signatures. We present an experimental proof of concept to validate the efficiency of this technique. For this, we use a deep learning model based on long-short-term memory networks. Our results have a significant improvement in the precision rate obtained with strategies that do not consider frequency diversity, increasing from 92.10% to 98.10%.

An integral image is a vital pre-processing step for several detection algorithms. Object detection algorithm requires real-time operation. Therefore, a strong restriction of time is imposed over the integral image processing. However, the integral image calculation speed is proportional to the image resolution. Hardware implementation provides the necessary speed to achieve a real-time object detection. Previous approaches have relied on an image that has been acquired and stored beforehand to start processing it. However, the performance of the system must take into consideration the acquisition and storing of the image. This paper introduces an integral image architecture that integrates the acquisition and processing stage. Therefore, the proposed architecture leaves out the need for external memory and eliminates the need to wait for an image to be acquired and stored.