Abstract: In a general aspect, a set of observed frequency-domain channel responses is filtered to remove noise or distortions that are not related to changes in the physical environment. In some aspects, for each frequency-domain channel response, a time-domain channel response is generated based on the frequency-domain channel response; and a filtered time-domain channel response is generated based on a constraint applied to the time-domain channel response. Additionally, a reconstructed frequency-domain channel response is generated based on the filtered time-domain channel response. An error signal is also generated, and a determination is made as to whether the error signal satisfies a criterion. The error signal can be indicative of a difference between the frequency-domain channel response and the reconstructed frequency-domain channel response. In response to each of the error signals satisfying the criterion, motion of an object in a space is detected based on the set of frequency-domain channel responses.

Abstract: In a general aspect, a set of observed frequency-domain channel responses is filtered to remove noise or distortions that are not related to changes in the physical environment. In some aspects, for each frequency-domain channel response, a time-domain channel response is generated based on the frequency-domain channel response; and a filtered time-domain channel response is generated based on a constraint applied to the time-domain channel response. Additionally, a reconstructed frequency-domain channel response is generated based on the filtered time-domain channel response. An error signal is also generated, and a determination is made as to whether the error signal satisfies a criterion. The error signal can be indicative of a difference between the frequency-domain channel response and the reconstructed frequency-domain channel response. In response to each of the error signals satisfying the criterion, motion of an object in a space is detected based on the set of frequency-domain channel responses.

Abstract: In a general aspect, a wireless communication device operating as a client in a wireless communication network receives wireless signals transmitted from an access point of the network. The device generates channel information from the wireless signals and processes the channel information to identify a degree of motion and average breathing rate of a person. Upon determining that the degree of motion and average breathing rate are below respective thresholds, the device begins sleep monitoring. Sleep monitoring includes generating additional channel information that is processed to identify a category of sleep.

Abstract: In a general aspect, a set of observed frequency-domain channel responses is filtered to remove noise or distortions that are not related to changes in the physical environment. In some aspects, for each frequency-domain channel response, a time-domain channel response is generated based on the frequency-domain channel response; and a filtered time-domain channel response is generated based on a constraint applied to the time-domain channel response. Additionally, a reconstructed frequency-domain channel response is generated based on the filtered time-domain channel response. An error signal is also generated, and a determination is made as to whether the error signal satisfies a criterion. The error signal can be indicative of a difference between the frequency-domain channel response and the reconstructed frequency-domain channel response. In response to each of the error signals satisfying the criterion, motion of an object in a space is detected based on the set of frequency-domain channel responses.

Abstract: In a general aspect, motion in a space can be detected based on machine learning of wireless signal properties. In some aspects, sets of tagged neural network input data are obtained at a neural network training system. Each set of tagged neural network input data is based on a statistical analysis of a series of wireless signals transmitted through a space over a respective time period, and each set of the tagged neural network input data includes a tag indicating whether motion occurred in the space over the respective time period. The sets of tagged neural network input data are processed by the neural network training system to parameterize nodes of a neural network system. Parameterizing the nodes configures the neural network system to detect motion based on untagged neural network input data.

Abstract: In a general aspect, motion of an object is detected based on wireless signals. In some aspects, a receiver in a space receives wireless signals; the wireless signals are based on transmissions of a reference signal by a transmitter. Each of the received wireless signals is based on a respective transmission of the reference signal at a distinct time. Channel responses are determined based on the received wireless signals and the reference signal. Each channel response is determined based on a respective one of the received wireless signals. Motion of an object in the space is detected based on the channel responses.