Principal Investigator Anantha Chandrakasan
The Internet of Things (IoT) connects together an exponentially growing number of devices with an estimate of more than 70 billion devices in less than ten years from now. Such devices revolutionize the personal heart monitoring, home automation, as well as the industrial monitoring systems. Unfortunately, the wireless IoT nodes consume a huge portion of their en- ergy on communicating with other devices. On the other hand, a longer battery lifetime or even a batteryless energy-harvesting operation requires a sub-microwatt consumption without significant performance deg- radation. In this work, we propose protocol optimizations as well as circuit-level techniques in the design of a -80dBm sensitivity ultra-low power wake-up receiver for on-demand communication with IoT nodes.
Wireless protocols such as Bluetooth low-energy (BLE) are optimized for short-length packets with small preambles and reduced header sizes. However, the power consumption of a low duty-cycled node in the default connected-mode is limited by the periodic beacons dictated by the protocol. Commercial BLE chips are then limited to tens of microwatts even though their standby power is in the nanowatt range. This wake-up receiver exploits the lower limit of the standby power to achieve significant power reduction through a wake-up scheme wrapped around the BLE advertising protocol. The receiver employs such duty-cycled wake-up scheme to mitigate the power/sensitivity trade-off achieving sub-microwatt average power at the required BLE sensitivity. When the receiver decodes its wake-up pattern inside the BLE advertising packet, it generates a wake-up signal then reconfigures its correlator with a new pattern. The power/latency trade-off is shown where a user with a commercial app can use a cellphone to wake any sleeping IoT node up using the BLE standard according to the application at hand.