Human activity recognition based on channel state information (CSI) using commercial WiFi devices plays an increasingly important role in many applications, such as smart home and interactive games. In this paper, we propose a WiFi CSI based human activity recognition approach using deep recurrent neural network (HARNN). HARNN mainly exploits four key techniques to recognize different human activities. HARNN firstly constructs a novel two-level decision tree for using two environment variation statistics efficiently. Meanwhile, a linear regression method is also introduced to seek for the optimal parameter for the designed decision tree. Depending on this, the decision tree is used to sense indoor environment variation, and then detect whether there is any human activity occurring in a target area. In addition, a noise removal mechanism is devised to eliminate the influence of random noise derived from indoor environments. Then, to characterize various human activities, two representative features are extracted from different statistical profiles, including channel power variation (CPV) and time-frequency analysis (TFA). Finally, a recurrent neural network (RNN) model is utilized to recognize different human activities by leveraging the extracted representative features above. According to the above steps, the proposed HARNN could establish a robust relationship between human activities and WiFi CSI compared with most of the existing WiFi CSI based approaches. The proof-of-concept prototype of HARNN is implemented on a set of commercial WiFi devices, and its overall performance is evaluated in several typical indoor environments. The experimental results demonstrate that HARNN can achieve better recognition performance compared with some benchmark approaches.

With the prevalence of commodity WiFi devices and development of the Internet of Things (IoT), the usage of WiFi has been extended from communication to context aware. Based on this, indoor localization has attracted increasing attention in the academic community, without the need for additional sensors and any active engagement from the users. However, the localization performance is vulnerable to the background noises due to relying on signals changes. To address this issue, in this article, we propose a passive 3-D indoor localization with a radio map for the mobile target by exploiting channel state information (CSI) of WiFi signals, realizing human–computer interaction (HCI). To this end, 3-D space is first divided into multiple independent regions and we construct a spatial radio map by traversing all the subspaces using CSI measurements. Next, to fully characterize the profiles of the locations of the mobile target, we reconstruct CSI time series to form a CSI tensor through integrating WiFi transmission links and a CANDECAMP/PARAFAC (CP) decomposition method is applied to this tensor for obtaining representative features. Then, the features-location data set is optimized by combining $t$ -distributed stochastic neighbor embedding (t-SNE) and information theory method to reconstruct a fine-grained fingerprint map for improving system performance. Finally, a recurrent neural network (RNN) model is introduced to learn the features data set optimized and then build a nonlinear correlation between input and output for realizing the purpose of accurate indoor localization. The proposed scheme is implemented on a set of commodity WiFi devices and evaluated in indoor scenarios. Based on real-world CSI data, our experimental results confirm the effectiveness of the proposed scheme in terms of localization accuracy and robustness against the noises.

Human identity identification based on channel state information (CSI) using commercial WiFi devices has drawn increasingly attention, and it can be used in many applications such as smart home, intrusion detection, building monitoring, activity recognition, etc. However, most of the existing identity identification approaches are sensitive to the influence of random noise derived from indoor environments, and thus their identification accuracies are far from satisfactory. In the present paper, a device-free CSI based human identity identification approach using deep learning (Wihi) is proposed. Wihi mainly utilizes three key techniques to identify different people. Firstly, to eliminate the influence of the random noise, discrete wavelet transform (DWT) strategy is introduced to denoise raw CSI data by leveraging signal decomposition. Secondly, in order to characterize human’s gaits profoundly, several representative features are exploited from different statistical profiles, including channel power distribution in time domain (CPD), time-frequency analysis (TFA), and energy distribution in different frequency bands (ED). Thirdly, a recurrent neural network (RNN) model with long short-term memory (LSTM) blocks is employed to learn the representative gait features extracted above and encode temporal information for realizing human identity identification. The proof-of-concept prototype of the proposed Wihi approach is implemented on a set of commercial WiFi devices, and multiple comprehensive experiments have been carried out to evaluate the performance of identity identification. The experimental results confirm that the proposed Wihi can achieve a satisfactory performance compared with some state-of-the-art approaches.

With the ubiquity of commodity WiFi devices and the rapid development of Internet of Things (IoT), there are increasingly intelligent sensing applications emerging by utilizing fine-grained channel state information (CSI) from WiFi signals, which can realize contactless human-computer interaction (HCI) for smart home. However, most of CSI-based activity recognition approaches are vulnerable to random noises derived from indoor environments. In this paper, we present a multimodal fusion-AdaBoost based human activity recognition scheme on WiFi platform. For this purpose, we develop two theoretical underpinnings, including a sensing model and a recognition model. The sensing model is firstly established to investigate the impact of human activities on propagation properties of WiFi signals. To this end, Fresnel zone model is used to quantify the correlation between CSI dynamics and human activities. Moreover, the recognition model is used to exploit physical activity-induced signal changes to infer potential activity information. In the process, we firstly construct a CSI tensor through integrating all CSI information at each WiFi receiver. Then, a CANDECAMP/PARAFAC (CP) decomposition method is applied to this CSI tensor for obtaining representative features. Finally, based on the features extracted, human activities are able to be recognized by combining multimodal fusion and AdaBoost. We implement the proposed scheme on a set of WiFi devices and evaluate it in both laboratory and corridor environments. The experimental results confirm that the proposed scheme can achieve average recognition accuracies of 96% and 95% in two indoor scenarios, respectively.

With the increasing demand for location-based service (LBS), WiFi-based localization has attracted considerable attentions due to the wide deployment and low cost of WiFi devices. However, most of existing approaches need to work within line-of-sight (LoS) range or require users to take dedicated devices. Thus, a device-free and passive indoor localization scheme is still desired. In this paper, a multi-view features fusion and AdaBoost-based indoor localization scheme by exploiting channel state information (CSI) of WiFi signals is proposed. To this end, we firstly formulate a shared consistent representation by using CSI measurements from all WiFi receivers, modeling the common properties among all views. Based on this, a features extraction is applied to the consistent representation for charactering channel properties. Meanwhile, these features extracted are optimized via a t-distributed stochastic neighbor embedding (t-SNE) algorithm for selecting the effective features, contributing to system performance. AdaBoost then builds a non-linear mapping between the CSI features and locations, realizing the purpose of localizing a user relying solely on WiFi signals with a high level of granularity, without any active engagement from the users. We implement a prototype on commodity WiFi devices and conduct comprehensive experiments in indoor scenarios. Based on real-world CSI measurements, the results confirm that the proposed scheme can achieve average localization errors of 0.8 m and 1.1 m in two indoor scenarios.