Over-the-Air Firmware Updates for Constrained NB-IoT Devices.

The Internet of Things (IoT) is being deployed to provide smart solutions for buildings, logistics, hospitals, and many more. It is growing with billions of connected devices. However, with such tremendous growth, maintenance and support are the hidden burdens. The devices deployed for IoT generally have a light microcontroller, low-power, low memory, and lightweight software. The software, which includes firmware and applications, can be managed remotely via a wireless connection. This improves flexibility, installation time, accessibility, effectiveness, and cost. The firmware can be updated constantly to remove known bugs and improve the functionality of the device. This work presents an approach to update firmware over-the-air (OTA) for constrained IoT devices. We used Narrowband IoT (NB-IoT) as the wireless communication standard to communicate between the managing server and devices. NB-IoT is one of the most promising low power wide area (LPWA) network protocols that supports more than 50k devices within a cell using a licensed spectrum. This work is a proof of concept demonstrating the usage of NB-IoT to update firmware for constrained devices. We also calculated the overall power consumption and latency for different sizes of the firmware.

Downlink Performance Modeling and Evaluation of Batteryless Low Power BLE Node.

Deploying low maintenance and long-life systems is an important requirement of emerging commercial Internet of Things (IoT) solutions. Such systems can be envisioned in which the connected devices are powered by energy harvested from ambient sources and stored in long-lifetime capacitors rather than short-lived and polluting batteries. However, due to the unpredictable nature of ambient energy harvesting, such batteryless IoT devices might not always have enough energy to initiate communication. The Bluetooth Low Energy (BLE) specification defines support for Low Power Nodes (LPNs) using the friendship feature, where the LPN is associated with a neighbouring friend node (FN). The LPN can receive downlink (DL) data and remain connected to the network via the FN that buffers the LPN's incoming packets while allowing the LPN to save energy by sleeping or turning itself off. This novel BLE feature makes the LPN highly suitable to support the connection of batteryless ambiently-powered IoT devices. While the LPN can decide when to transmit uplink (UL) packets and does not depend on the FN to receive downlink (DL) data, the LPN needs to poll its FN to receive the buffered packets. However, the DL packet latency increases with this process due to the buffering time at the FN. Therefore, in this work, we present an analytical model to characterize the performance as a function of DL data latency and packet delivery ratio (PDR) of a batteryless LPN powered by different harvesting powers and capacitor sizes. This would help to optimally choose the correct configuration of the batteryless LPN for its network deployment. We also compare the analytical model and simulation results, showing consistency with an average error of 2.23% for DL data latency and 0.09% for the PDR.

Enabling Low-Latency Bluetooth Low Energy on Energy Harvesting Batteryless Devices Using Wake-Up Radios.

With the growth of the number of IoT devices, the need for changing batteries is becoming cumbersome and has a significant environmental impact. Therefore, batteryless and maintenance-free IoT solutions have emerged, where energy is harvested from the ambient environment. Energy harvesting is relevant mainly for the devices that have a low energy consumption in the range of thousands of micro-watts. Bluetooth Low Energy (BLE) is one of the most popular technologies and is highly suitable for such batteryless energy harvesting devices. Specifically, the BLE friendship feature allows a Low Power Node (LPN) to sleep most of the time. An associated friend node (FN) temporarily stores the LPN's incoming data packets. The LPN wakes up and polls periodically to its FN retrieving the stored data. Unfortunately, the LPNs typically experience high downlink (DL) latency. To resolve the latency issue, we propose combining the batteryless LPN with a secondary ultra-low-power wake-up radio (WuR), which enables it to always listen for an incoming wake-up signal (WuS). The WuR allows the FN to notify the LPN when new DL data is available by sending a WuS. This removes the need for frequent polling by the LPN, and thus saves the little valuable energy available to the batteryless LPN. In this article, we compare the standard BLE duty-cycle based polling and WuR-based data communication between an FN and a batteryless energy-harvesting LPN. This study allows optimising the LPN configuration (such as capacitor size, polling interval) based on the packet arrival rate, desired packet delivery ratio and DL latency at different harvesting powers. The result shows that WuR-based communication performs best for high harvesting power (400 µW and above) and supports Poisson packet arrival rates as low as 1 s with maximum PDR using a capacitor of 50 mF or more.