Designing Energy Harvesting-Low Power Wide Area Networks; A Feasibility Analysis

In Layman’s term ”wireless communication” often refers to either Wireless Fidelity (Wi-Fi) or cellular data connections we have these days. Each of them can fit in for a variety of use-cases based on what we need. For example, Wi-Fi enables its users to receive a large amount of data like watching a film from the comforts of their home, within a local area (small vicinity) and for a relatively low cost that is spent as a monthly fee or modem charges. Similarly, with the advancement in cellular infrastructures, cellular connections (such as, 3G or 4G/LTE) can almost do the same for a relatively broader geographical area but for the higher costs in terms of monthly data bundle being subscribed on the top of the mobile price. With the revolution of Internet of Things (IoT) paradigm, the transition of conventional cities into smart cities is in- evitable where the technological shift has to lead the transformation of communication phenomenon from personal to Machine-to-Machine (M2M) communication where billions of interconnected objects would be communicating with each other to contribute in an overall smart environment. Let’s assume a smart environment where a user reaches back home, and his car automatically starts communicating with the garage for opening up the door. Once inside, the temperature conditions are already adjusted according to his preference, and a lower intensity light is on in his chosen color for the relaxation, looking at the pacemaker data that indicates the day has been stressful. Acquiring this kind of services requires technologies different than Wi-Fi or cellular networks, that leads to the inception of Low Power-Wide Area Network (LP-WAN). LP-WAN are targeting the niche in a broad range of IoT applications where long radio coverage is required on the cost of extremely low bit rate support in an energy constrained environment where the message frequency is not reasonably higher. A plenty of contesting candidates have already been marked their entry into the market with some of them floating since 2009. At that time, the number of connected objects was relatively small. The Wireless Local Area Network (WLAN) (e.g., Wi-Fi) and Personal Area Network (PAN) (such as, Bluetooth) technologies had created their hype around IoT so it took some time for LP-WAN candidates to flourish enough to come up with mature solutions. Almost all the LP-WAN solutions are well capable to address the most critical issues like cost, coverage, current, and capacity of these networks. They can be seen in two different categories; Proprietary and standard LP-WAN solutions. Both of them have their own pros and cons. The standard LP-WAN solutions do not require a new infrastructure from the scratch, instead, what they only need is a software update to the existing infrastructure. On the top of it, they are very much available in almost all the countries with almost 100% area already covered by multiple operators in each country and they are very much capable to provide services not only locally but beyond the borders by reaching out different roaming agreements. On the other hand, proprietary solutions have to undergo deployment from the scratch though faster enough because of their operation in unlicensed Industrial, Scientific, and Medical (ISM) spectrum. They do not suffer from interference and resilient enough towards multi-path fading due to employing asynchronous communication protocols. However, they can offer limited Quality of Service (QoS), scalability and global reach as compared to standard solutions. When comes to cost and energy consumption, they are good enough in the sense that they neither need a separate subscription to start providing services nor they cause fast battery drainage to make the lifetime as longer as possible. With this enormous growth in the years to come, the challenges faced by LP-WAN paradigm are also gigantic. One of the premier among them is energy utilization and, consequently, battery life belonging to these low cost modules. Although low energy consumption is one of the peculiarities of these networks, but energy exhaustive operation can still leave a serious dent on the hype of LP-WAN. As the radio modules belonging to almost all the technologies come with tiny battery offering a small amount of storage capabilities due to their size limitations, thus they are designed to support most optimized operation to prolong the battery lives as maximum as possible. Such energy limitations might declare LP-WAN candidates unfit for several IoT use- cases which demand for the performance optimization (e.g., delay-critical applications). As performance optimization requires higher resources and it is not possible to achieve energy and performance optimization at the same time being two different objectives. Then, there are several energy exhaustive IoT use-case who require a relatively higher message frequency causing the motes depleting their batteries too earlier than their expectancy life of 8-10 years. On the top of it, it is generally undesired to replenish the batteries of these low cost modules because of their replacement cost (including higher labor cost needed to replace the batteries installed on harsh environment). Hence, it is utmost important to investigate the energy harvesting potential in the realm of LP-WAN to arrange for a separate but parallel source of energy to make them suitable for a wide variety of IoT use-cases and to enable them turning them- selves from energy optimized operation to performance optimized operation wherever needed. There is a wide range of energy harvesting techniques (including solar, wind, vibrational, thermoelectric, RF energy, to name a few) with their own pros and cons, with some of them to fit well in the LP-WAN paradigm. To this end, we started off with reviewing the energy harvesting MAC protocols already available in the literature for low powered sensor nodes. The MAC operation also plays an important part as it is responsible for all the channel sensing, reception and transmission operation. Thus, it is important to understand the interplay between energy harvesting techniques and the MAC protocols towards achieving the joint performance optimization. A variety of energy harvesting MAC protocols are studied with a special focus on the fundamental techniques, evaluation approaches, and key performance indicators along with presenting the pros and cons of each protocol discussed which provides an insight of the various research challenges and design guidelines for the research community working in this area. Moreover, the feasibility of a completely cable-less IoT deployments is investigated where dual radio LoRaWAN gateways are equipped with wireless backhaul and fed by a photovoltaic plant. To accomplish this purpose, the power demands of a dual radio gateway, that serves a mix of M2M applications and leverages different combinations of front-end chipset and backhaul wireless technologies, are evaluated. Then, the photovoltaic plants are properly sized in lined with those requirements and, installation as well as operational costs of this kind of deployments are properly calculated. Finally, cost and carbon footprint savings analysis is presented to demonstrate the economic viability and environment friendliness for this kind of approaches to demonstrate its benefits for Mobile Network Operators (MNOs). Furthermore, another most energy exhaustive use-case i.e., industrial monitoring is investigated employing LoRa monitoring nodes in the industry 4.0. Although duty- cycled operation can play its role to prolong the battery life of LoRa modules employed in a smart industry, however, it would introduce long communication delays causing a bulk of damaged products (i.e., higher damage penalty) on the production line in case of an anomaly. In an attempt to reduce damage penalty, battery replacement cost tends to go higher due to higher communication frequency (i.e., short sensing interval). The work presents a model to analyze this cost trade-off. We first analyze LoRaWAN performance in plain industrial environment and then highlight the benefits of energy harvesting potential available within an industry 4.0 in terms of prolonging the battery life of LoRa monitoring nodes and the flexibility of shortening the sensing interval as per the application requirement.