Abstract: Techniques and apparatuses are described that implement a smart device with an integrated radar system. The radar integrated circuit is positioned towards an upper-middle portion of a smart device to facilitate gesture recognition and reduce a false-alarm rate associated with other non-gesture related motions of a user. The radar integrated circuit is also positioned away from Global Navigation Satellite System (GNSS) antennas and a wireless charging receiver coil to reduce interference. The radar system operates in a low-power mode to reduce power consumption and facilitate mobile operation of the smart device. By limiting a footprint and power consumption of the radar system, the smart device can include other desirable features in a space-limited package (e.g., a camera, a fingerprint sensor, a display, and so forth).

Abstract: Techniques and apparatuses are described that implement a smart-device-based radar system capable of performing gesture recognition using a space time neural network. The space time neural network employs machine learning to recognize a user's gesture based on complex radar data. The space time neural network is implemented using a multi-stage machine-learning architecture, which enables the radar system to conserve power and recognize the user's gesture in real time (e.g., as the gesture is performed). The space time neural network is also adaptable and can be expanded to recognize multiple types of gestures, such as a swipe gesture and a reach gesture, without significantly increasing size, computational requirements, or latency.

Abstract: Techniques and apparatuses are described that implement a smart device with an integrated radar system. The radar integrated circuit is positioned towards an upper-middle portion of a smart device to facilitate gesture recognition and reduce a false-alarm rate associated with other non-gesture related motions of a user. The radar integrated circuit is also positioned away from Global Navigation Satellite System (GNSS) antennas and a wireless charging receiver coil to reduce interference. The radar system operates in a low-power mode to reduce power consumption and facilitate mobile operation of the smart device. By limiting a footprint and power consumption of the radar system, the smart device can include other desirable features in a space-limited package (e.g., a camera, a fingerprint sensor, a display, and so forth).

Abstract: Techniques and apparatuses are described that implement smartphone-based power-efficient radar processing and memory provisioning for detecting gestures. The described techniques map different situations that occur with a user to different memory states. A radar processing system's memory management module allocates at least one memory pool according to an active memory state. As the radar system detects different situations that occur with the user, the memory management module reallocates the memory pool for the appropriate memory state, which can adjust a power mode of the radar system. In some cases, physically separate memories may be allocated for different memory states, which respectively include one or more sequences that are executed by different processors. The memory management module enables efficient use of available power and available memory for radar processing such that a total amount of power and a total size of memory used may be significantly reduced for detecting gestures.

Abstract: Techniques and apparatuses are described that implement smartphone-based power-efficient radar processing and memory provisioning for detecting gestures. The described techniques map different situations that occur with a user to different memory states. A radar processing system's memory management module allocates at least one memory pool according to an active memory state. As the radar system detects different situations that occur with the user, the memory management module reallocates the memory pool for the appropriate memory state, which can adjust a power mode of the radar system. In some cases, physically separate memories may be allocated for different memory states, which respectively include one or more sequences that are executed by different processors. The memory management module enables efficient use of available power and available memory for radar processing such that a total amount of power and a total size of memory used may be significantly reduced for detecting gestures.