Energy Neutral Operations


Can we autonomously manage nodes in wireless sensor networks (WSN) to optimise and extend the utility of their harvested energy taking into consideration variability in the environment and battery storage degradation?



Imperial College London


We designed and deployed a 20 node network of microclimate sensors in Queen Elizabeth Olympic Park and tested the proposed algorithm for two months.


We developed a lightweight battery degradation model to monitor battery health in a constrained compute environment.


We developed a novel lightweight sensor management scheme to extend the lifetime of the deployment of an arbitrary IoT application while guaranteeing energy neutral operation.


The results show that an increase of deployment lifetime of 307% can be achieved without a reduction in average system performance.

The Internet of Things (IoT) has rapidly matured in recent years and is becoming a viable solution for real world deployments. The major barriers to the practical adoption of IoT systems have been the limited deployment opportunities if the devices run on mains power and limited operational lifetime if the devices run on batteries. Added costs include the need for human intervention when batteries eventually need to be replaced, the environmental concerns of disposal of batteries, and the introduction of errors when those batteries near end of life. Energy harvesting from sources, such as solar, wind, thermoelectric, and vibration have been put forth as potential solutions to this problem when paired with a rechargeable battery. The introduction of energy harvesting to IoT brings improvements for system performance, however, gains from energy harvesting come with additional complexities, i.e. management of the energy budget given dynamic demands on the network, potentially unpredictable energy generation and energy storage life-time management.

Energy Neutral Operation (ENO) is a mode of operation of an IoT object where the energy


consumption is always less or equal to the energy harvested from the environment. However, existing ENO approaches do not take into account the degradation of the capacity of the battery. Specifically, the capacity of the battery is assumed to be fixed throughout the lifetime of the deployment. In fact, battery capacity degrades over time and can only undergo a limited number of charge/discharge cycles before it fails. Through our research, we aim to address the issue of how to incorporate awareness of this degradation into ENO optimisation algorithms so that both the battery capacity and duty cycle of each sensor node in an arbitrary IoT application are maximised. As a result, the deployment time of an application can be extended, which results in reducing the associated costs of replacement and in field maintenance.

Beyond WSN, the optimisation of rechargeable battery health to promote longevity has applications in electric vehicles, home battery backup systems, healthcare, and consumer electronics such as wearables, smartphones, and beyond.