LittleRock is a novel low-power sensing architecture for cell phones where sampling and low-level processing of sensor data is offloaded to a dedicated processor for enabling energy-efficient continuous sensing-based applications.
Modern cell phones are equipped with multiple sensors such as accelerometers, proximity sensors, compasses etc. Since a cell phone is carried by its user most of the time, valuable user context information can be obtained by continuously monitoring the activity on and around a phone. For example, the accelerometer readings on a phone gives a accurate measure of user activity, while audio signals captured on the phone gives a valuable information about the user’s surrounding. However, the current cell phone sensing architecture, where the phone's main processor has to actively probe a sensor to collect data, is not suitable for such continuous sampling of sensors. This is because the phone's main processor consumes several orders of magnitude more energy than a typical sensor.
LittleRock sensing architecture uses a dedicated low-power processor with limited processing capability for sampling and low-level processing of sensor data. The low-power processor notifies the main processor when an interesting event that requires further attention is detected. This enables the phone's main processor to be in a low-power sleep state while the low-power processor is actively sampling sensor data. The LittleRock project is aimed at developing both the hardware sensing architecture and software APIs for accessing the sensors through the low-power processor.
Primary contact : Bodhi Priyantha (bodhip)
LittleRock hardware prototype
LittleRock prototype block diagram