Simone Di Domenico

Trained-once device-free crowd counting and occupancy estimation using WiFi: A Doppler spectrum based approach

This paper presents a WiFi-based device-free crowd counting and occupancy estimation system that can be used in rooms/environments different from the ones in which the training process has been performed. Therefore, crowd counting is achieved without requiring another training phase in each new environment. The proposed approach analyzes the shape of the Doppler spectrum of the received signal which is correlated to the number of people moving in the monitored environment. Unlike a radar-like approach, the use of a reference signal is not required. Experimental results are presented for two different rooms/environments without any constraint on the movements of the volunteers.

Radios as Sensors

Radio receivers, besides acting as wireless network nodes participating to the Internet of Things (IoT) communication task, may act as opportunistic sensors participating to the IoT sensing task. In particular, a radio receiver is intrinsically an electronic sensor which may be used for device-free human activity recognition. In this paper, we analyze recent results on how the identification of the human body presence and movement can be carried out analyzing the RF signals transmitted by sources of opportunity. The impact of channel bandwidth, transmission mode, carrier frequency, and signal descriptors on the recognition performance is discussed. Moreover, we present a novel crowd counting system and assess the performance considering two different types of signal descriptors. Results prove the effectiveness of the presented crowd counting system and allow to get more insights into the relation among the specific sensed environment, chosen signal descriptors, and classification accuracy.

A Trained-once Crowd Counting Method Using Differential WiFi Channel State Information

This paper focuses on the problem of providing a rough count of the number of people in a room using passive WiFi Channel State Information (CSI) measurements taken by a single commodity receiver. The feature which mainly distinguishes our work from others is the attempt to emerge with an approach which does not require any dedicated training inside the specific environment where the system is deployed. Our proposal stems from the intuitive observation that features which account for em variations of CSI are expected to be less sensitive to the surrounding environment as opposed to features which account for absolute CSI measurements. We turn such intuition into a concrete proposal, by suitably identifying a set of differential CSI feature candidates, and by selecting the (two) most effective ones via minimization of the summation of the Davies-Bouldin indexes. We preliminary assess the effectiveness of the proposed approach by training once for all the system in a room, and testing the system in two em different rooms having different size and furniture, and involving people freely moving in the rooms with no a-priori movement constraints.

LTE signal fingerprinting localization based on CSI

This paper investigates the possibility to use Channel State Information (CSI) extracted from Long Term Evolution (LTE) signals for signal fingerprinting localization. Being the first work in this direction, several types of signal fingerprinting-based approaches have been compared (e.g., CSI-based vs RSSI-based, statistic vs deterministic matching rule). In particular, the paper proposes a novel CSI-based signal fingerprinting that uses as fingerprint not directly the vector of channel gains per subcarrier, but rather some features extracted from these vectors. This method would greatly reduce the memory requirement of the database as well as the computational complexity of the matching phase. Experimental results, shown for both indoor and outdoor environments, confirm the effectiveness of the proposed method and also provide interesting insights on the use of LTE signal fingerprinting based on CSI.

PCA-based artifact removal algorithm for stroke detection using UWB radar imaging.

Stroke patients should be dispatched at the highest level of care available in the shortest time. In this context, a transportable system in specialized ambulances, able to evaluate the presence of an acute brain lesion in a short time interval (i.e., few minutes), could shorten delay of treatment. UWB radar imaging is an emerging diagnostic branch that has great potential for the implementation of a transportable and low-cost device. Transportability, low cost and short response time pose challenges to the signal processing algorithms of the backscattered signals as they should guarantee good performance with a reasonably low number of antennas and low computational complexity, tightly related to the response time of the device. The paper shows that a PCA-based preprocessing algorithm can: (1) achieve good performance already with a computationally simple beamforming algorithm; (2) outperform state-of-the-art preprocessing algorithms; (3) enable a further improvement in the performance (and/or decrease in the number of antennas) by using a multistatic approach with just a modest increase in computational complexity. This is an important result toward the implementation of such a diagnostic device that could play an important role in emergency scenario.

Exploring Training Options for RF Sensing Using CSI

This work analyzes human behavior recognition approaches using WiFi channel state information from the perhaps less usual point of view of training and calibration needs. With the help of selected literature examples, as well as with more detailed experimental insights on our own Doppler spectrum-based approach for physical motion/presence/cardinality detection, we first classify the diverse forms of training so far employed into three main categories (trained, trained-once, and training-free). We further discuss under which conditions it is possible to move toward lighter forms of calibration or even succeed in devising fully untrained model-based solutions. Our take home messages are mainly two. First, reduced training might not necessarily kill performance (although, of course, trade-offs will emerge). Second, reduced training must come along with a careful customization of the technical detection approach to the specificities of the behavior recognition application targeted, as it seems very hard to find a one-size-fits-all solution without relying on extensive training.

LTE-based passive device-free crowd density estimation

In the last few years, there has been a growing interest in developing sensing systems using RF signals of opportunity, especially exploiting radar-based techniques. Long Term Evolution (LTE) signals are excellent candidates as signals of opportunity thanks to their wide availability and penetration in indoor environments. This is the first work investigating the possibility to use LTE signals for crowd density estimation. The proposed approach is not radar-like but it exploits the correlation between the variations of the received LTE signals (in particular, of the Reference Signal Received Power) and the number of people. An experimental evaluation of the performance is carried out in a indoor environment testing three different positions of the LTE receiver. Achieved results in terms of classification accuracy are very promising.