In this paper, Gaussian process regression (GPR) for fingerprinting based localization is presented. In contrast to general regression techniques, the GPR not only infers the posterior received signal strength (RSS) mean but also the variance at each fingerprint location. The GPR does take into account the variance of input i.e., noisy RSS data. The hyperparameters of GPR are estimated using trust-region-reflective algorithm. The Cramér-Rao bound is analysed to highlight the performance of the parameter estimator. The posterior mean and variance of RSS data is utilized in fingerprinting based localization. The principal component analysis is employed to choose the k strongest wi-fi access points (APs). The performance of the proposed algorithm is validated using using real field deployments. Accuracy improvements of 10% and 30% are observed in two sites compared to the Horus fingerprinting approach.

No abstract

Crowd detection and monitoring is an active area of research because of its significance in many areas, especially law enforcement. Various sensor modalities, such as infrared imaging, video feed, received signal strength indicator, Radio-frequency identification, GPS signals, and audio tones through mobiles have been used in earlier work. In this paper, a method that uses passive acoustic sensors in a multimodal sensor network for crowd monitoring is described. This multimodal system uses three modalities, namely, carbon dioxide level, sound intensity level, and received signal strength for crowd detection and monitoring. The first two modalities are sensed using passive sensors, whereas the last one is an active sensor. This combination makes the proposed algorithms energy efficient and computationally less complex. The proposed multimodal crowd monitoring algorithm requires an effective clustering method. Hence, three clustering algorithms that utilize temporal, spatial, and spatio-temporal information are also proposed. Subsequently, an algorithm that fuses the information in different modalities is also proposed for efficient crowd monitoring. Additional contributions of this paper are the development of attenuation, reverberation, and additivity models, using real sensor deployments. Both simulation and real field experiments are conducted to evaluate the performance of the proposed algorithms in indoor and outdoor spaces. The results of crowd detection and monitoring obtained from these methods are motivating enough to use the proposed method in real small-scale deployment scenarios.Index Terms-Attenuation modeling, spatial-temporal clustering, crowd monitoring, multi-modal sensor network. 1530-437X

No abstract

No abstract

No abstract

No abstract

No abstract