ABSTRACT
This chapter introduces the activity recognition and localization using array antenna, referred to as an array sensor. The array sensor exploits an antenna array on the receiver side and decomposes the received signals into eigenvectors and eigenvalues. The array sensor can realize passive localization using support vector machine (SVM) based on those features in a fingerprinting manner. The chapter shows an example of an array sensor. When a pair of transmitter and receiver is fixed, the signal subspace spanned by eigenvector changes when the indoor environment of interest changes. For detecting simple events, such as intrusion, a simple threshold detection based on the change of the first eigenvector is used. For detecting and classifying more complex states and activities, such as sitting in a bathtub and falling in a bathroom, machine learning, such as SVM, is used. The chapter explains how these methods are implemented in the array sensor.
