By fu3l1css

New hardware design trends for constrained NBIoT battery operated sensors

Athens, 28/09/2018

Since the first pilot projects around the world proving the connectivity and performance merits of NB-IoT networks, our view over the emerging sensor design trends has dramatically changed.

At the beginning, let us understand what is going to change with the introduction of NBIoT networks.

-These networks have been designed to broadcast on existing ubiquitous frequencies that we use since the 90’s. Mobile network companies have been continuously investing in the infrastructure upgrade since then. Nowadays, these networks, have been upgraded to receive very small (constraint) messages over a new protocol that is called NarrowBand. Sensors nowadays attach to the wireless infrastructure your mobile phone is using. Nothing more, nothing less. The networks are there and they have been designed to facilitate more that 50.000 sensors per cell per base station. Network operators should thing big, our innovation team has a shifted mindset towards massively deployed sensors. 

-The protocol opts for constraint devices powered by batteries. What we practiced so far, is that any type of sensor used, should be either connected to the power grid (under a AC to DC transformer) or being powered by batteries (e.g the vehicle battery). This requires the end mile integrator to include in his budget the installation cost. That includes extra copper deployed, extra labor and higher maintenance OPEX. NBIoT networks simple render this practice obsolete, because:

  1. All the sensors are battery operated; therefore there is no need for powering them up, either from the grid or any type of bulky battery configuration. This accounts for immediate reduction of CAPEX cost.
  2. All sensors can operate in very low signal thresholds. With the current NBIoT version released, we have attached to a base station 2,5 Km away, with a signal level of -131 dBm. For the RF guys around there, this is simply a magnificent breakthrough!  Never before a city network can operate in such low sensitivity levels and next releases claim that network connection can be even possible at lower signal levels.  How does this translates to cost reduction? First of all, no external antennae have to be used. Just a F-inverted meander monopole printed on the sensor as a radiating element and you can successfully transmit and receive. Second of all, since NBIoT protocol can establish connection between the sensor and the base station even in underground installation schemes, therefore zeroing any need of deploying extra power or data networks.
  3. Constraint devices (like NBioT sensors) need to bear ingenuity with respect to data payload and power preservation. This is where innovation kicks in. Ingenuity of a sensor is the way we make the sensor operate under a minimum data transferring payload, without sacrificing data value. Why we need such a performance rationale; simply because our mindset must take account that we are considering tenths of thousand of sensors per base station. Now, consider the deployment of wireless networks in a city level. There are thousands of base stations deployed. Now, multiply the capacity per base station with the number of possible sensors per base station. The number becomes massive with a lot of digits. Now consider operating millions of sensors deployed in an area to measure any type of physical value or human activity. If the sensors are dummy, then a database will be gathering data junk along with valued data, therefore high costs of preserving and mitigating towards meta-data. Now imagine a sensor capable of preserving the lowest OPEX in IT infrastructure possible, by sending only data when the physical value measured changes. This is possible on the firmware level, where algorithms or machine learning code can govern the way sensors react to their measurement. Adding the need for battery constrained devices to operate for a decade, the firmware code must be advanced enough for the sensor to ‘wake up”, “sample” and “sleep”, under an orchestrated scenario.

In Fuelics, we do not just create sensors, but we create tailor-made sensors for the end user application. This is not an easy thing to do and requires experimental data and practical knowledge in order to reverse engineer the field installation specifics into firmware code. The results are magnificent by the way. We have sensors piloted in the open field that create 1/20th to 1/50th or less of the custom data payload of non-constraint sensors (e.g Power grid GPRS operating sensors). This straightforwardly minimizes OPEX and the northbound application framework can produce services by using only thin data payloads.

-Constraint devices for massive deployments must be cheap. Everybody asks for cheap sensors, but actually nobody knows why! We believe that for the first time, since we are opting for battery operated sensors that will be operating for at least 10 years, we need to change the narrative. Sensors cannot be considered as passive devices. Even if installation costs are minimum, a cheap sensor is a sensor that simultaneously is CAPEX and OPEX cheap. But, since we are discussing about a decade of operating life, what is more important to a business model is the lowest possible OPEX cost.

Summing up, in order to set things in new perspective, we need to be able to design and develop sensors that will be able to facilitate new types of business models that take under consideration the revolutionary aspects of the NBIoT protocol, the new type of specifications for constraint devices, the new rationale on minimizing data throughput and lastly but not least, the nearly zero $ yearly data tariff.

NBIoT networks change the narrative in legacy IoT integrated systems, by transforming the operational and manufacturing guidelines to support the transition from dummy to clever sensors, from CAPEX intensive projects to long term thin OPEX spending, from sensor agnostic IT applications to sensor specific applications.

Dr. Evangelos S. Angelopoulos is an industrial adjunct research in NCSR Demokritos developing exponential fuel and smart city sensors.

By fu3l1css

Massive Deployment of Exponential IoT sensors. Is this the time for it?

Exponential networks like NBIoT and LoRa have emerged with the aim to facilitate the deployment of cheap, abundant and intrinsically smart battery operated sensors that will be commissioned on the fly and will be working for 10 years. This is the new era. In Fuelics we try to follow these design principles and produce sensors that will be easily commissioned (actually we want the sensor to be installed by agnostic users), will produce minimum data (as a way to minimize OPEX) and will be able to measure accurately any operational value.

But, is the market ready for massive deployments of sensors? Well, yes and no.

In order to be able to deploy battery operated sensors first of all you need to be able to transfer the measured data over the cloud. In NBIoT the data are being broadcasted over the already installed LTE cellular network (which is a rather dense infrastructure in the city environment), whereas in LoRa the data require concentrators that have to be privately deployed in unlicensed chunks of spectrum. So, in both cases you need the infrastructure and somebody to turn the engine on. With NBIoT, the aforementioned ignition is a software upgrade (well it looks like that from a macroscopic view). With LoRa you actually need to install antennae that will concentrate signals over a mesh network. Fueling the ignition process requires ONE THING. Investments!

And now let us consider that investments are done and Athens, Greece (the city I work and live) has a fully fledged exponential infrastructure. Will we be able to massively deploy sensors now? Well yes technically, but NO financially or ecosystem-wise. Once the networks operate and the 125KHz of NBIoT spectrum (it is obvious that in Fuelics we love this technology) is available everywhere, the market has been weaponized with the infrastructure, but we lack of the end-mile services. During our involvement with NBIoT development we have recognized that weakness. What will be the services fueled by exponential sensors that will be massively rolled out?

I have made a short list of possible services based on massively deployed sensors.

A) Tank Monitoring. Tanks come in millions. Fuel tanks, water tanks, garbage tanks (bins = a form of tank), water gathering tanks (cisterns). In Fuelics we have been developing NBIoT sensors for all the above markets.

B) Water and Gas Metering. Most of the sensors already deployed do have a dry contact (pulse) output, therefore by attaching a exponential module instantly billing or value added services can be offered. In Fuelics we have developed a multimodal NBIoT telecommunication extender capable of receiving information from water or gas sensors.

C) Facility Management. This is a market segment with vast interest, since this is an old school business that is being gradually transformed into an IoT business. In Fuelics apart from water, gas, or fuel management of buildings, we consider investing in the development of exponential grade sensors towards facilitating the holistic building management. Burners and elevators will be our primary focus.

D) Electricity Management. Countless technologies can be mentioned here. Metering, sub metering, energy harvesting, energy actuation and so one. From heating to lighting, numerous IoT sensors can be deployed and integrated with legacy SCADA or BMS systems. This market segment is huge. At Fuelics we are investing in designing a remote energy management unit with power switching and net metering capabilities. 

E) Parking Management. The market is huge once again. How many parking slots are available within a smart city infrastructure? Hundred thousand to million parking slots, I roughly calculate. We have piloted our NBIoT parking sensor successfully.

F) Are there more? Let’s discuss it and see if we can find synergies.

Overall, in my opinion, we are in a chicken and egg situation. Everybody wants to transform towards IoT rationale, but there are no integrated services yet or fully fledged networks to support them. The remedy for our ecosystem is to create synergies between sensor manufacturers and middleware IoT platform providers with the aim to launch cheap and innovative services for the end mile.

Dr. Evangelos Angelopoulos is a serial entrepreneur and industrial Adjunct Researcher in NSCR Demokritos Greece, doing research on the quality and quantity assessment of diesel fuel, while trying to identify markets for massive deployment of exponential sensors and simultaneously facilitates the transformation of basic research towards new innovative industrial grade products.

By fu3l1css

What is Narrow Band IoT (NBIoT)?

Well, in @Fuelics we believe that it is the technology that will really deliver true #IoT Massive deployments. I will try to analyse that in brief:

1. NBIoT uses a small definite part of the #800MHz #spectrum that has been traditionally being used since the first day of cellular networks. The frequency itself is penetrative with respect to its wavelength (in comparison to higher frequencies) and the cellular service at 800MHz is ubiquitous (keep that in mind). 

2. NBIoT comes as a software upgrade of existing base stations. At the end of the day, cellular networks will just upgrade their software. This is really amazing for the IoT ecosystem, since launching does not interrupt current cellular service. 

3. Every #cellular base station once upgraded, will be able to provide connectivity to a vast number of sensors. The numbers are crazy. 50000 to 100000 devices per cell per base station. Let me elaborate. Just raise your eyes in the sky and search for the antennae of base stations. This is the only infrastructure you need to deploy IoT. The infrastructure is there already. 

4. “But 50 thousand (50000) devices per cell is a huge number. How will the network per cell per base station can service so many devices”? “Narrow Band” is the plain answer. Sensors dispersed physically around every cell will communicate with the base station in very short radio burst within just 200KHz of allocated spectrum. It is like having several flocks of birds tweeting at the same time in the same frequency. Small, short but unique tweets. Like a charm! 

5. ” I cannot see why tech evangelists consider NBIoT the best technology for Utility networks? ” some may ask. The answer is plain and simple. “Because it will be cheap and very intrusive with respect to the radio signals that can penetrate even within 2 floors underground”. Oops!!! This is something new for the cellular era. Bad reception in underground spaces was always a vulnerability for cellular networks.

I am very happy to tell you that yesterday (7th March 2018) we tested our revised radio modem and we where able to communicate with the NBIoT enabled base stations at a nominal distance of 2,5 Km, with our sensor being installed in a diesel tank, almost 6 meters below ground. Additionally, we did not have any direct optical connection with any of the NBIoT enabled base stations, thus the building itself was new, therefore was built with reinforced cement, which acts as an insulator for RF spectrum (actually our diesel tank resides more or less within a Faraday cage)

6. “When will NBIoT be ready for the masses”? Good question! Let me elaborate by asking in which part of the world do you happen to be? In Europe, things seem to go on a slow but decisive pace. In Greece by the way, the largest mobile carrier Cosmote worked extensively with Ericsson in order to deliver functional (but not yet fully fledged) NBIoT networks in Northern Greece. We were honored to be cited here and here, by the way. We believe that by the end of 2018 most of the cities around Europe will be NBIoT enabled. After all, it is being done just by a software upgrade!!!

7. Last but not least. Sensors. IoT in the NBIoT era is all about sensors doing measurements and feeding with big data the knowledge spectrum. Battery operated sensors! This is a game changer. We are referring to sensors that are made to last, be installed and be forgotten for more that 10 years. And smart sensors! How smart? Smart enough (with intrinsic intelligence) and the ability to understand when to throughput data. Imagine an IoT deployment with several hundred thousand of sensors. Each sensor is measuring a physical value and sends data to a database server. When we are talking about millions of sensors, the OPEX infrastructure cost will be significantly lower, if sensors just send only relevant information and not raw unfiltered information back to the server, resulting in lower saved data of at least 50 to 90%. Depending on the intrinsic intelligence of the sensor for monitoring assets like diesel tanks, in warm countries like Greece, burners seize to operate in mid Spring until the beginning of autumn, almost for six months. Now compare a smart sensor that understands that there is no need of sending repeatably the same fuel level value with a dummy sensor that is programmed to send the same value for almost six months, every hour (less is worse in this case) or less. There are no comparisons. Actually this is another novelty of our sensors. To be able to act autonomously and adapt (machine learning it is called).

8. “What about the cost for commissioning one sensor to the cloud”? How does 1$ per month per device sounds like? To my ears, as holy music coming from the skies. We are used to data plans ranging from several dollars to several tenths of dollars. This is really something new and innovative. The cost of service becomes very low, which is very good news for the IoT world, the world we make business and operate, while trying to make sensors smarter than before.

Dr. Evangelos S. Angelopoulos is the founder of Fuelics (www.fuelics.com) and an adjunct researcher in the Institute of Nanoscience and Nanotechnology, NSCR Democritos, where he is trying to design and develop off the self diesel sensors for quantitative and qualitative assessment of diesel fuel.

1 2 3
New hardware design trends for constrained NBIoT battery operated sensors
Massive Deployment of Exponential IoT sensors. Is this the time for it?
What is Narrow Band IoT (NBIoT)?