The last few years (even the last few months) have seen a surge in organisations exciting us about what farmers will do with sensors and mobile devices. We’ll be able to collect data with drones flying above our fields and tiny devices scattered across our farms. We’ll collect data with smart ear-tags and boluses in livestock, and we’ll crunch it all with powerful cloud-based analytics and smartphones will be a convenient way to tap into these analytics so that we can make decisions on the move.
That’s the promise.
Given time, much of this is achievable, though being savvy business people, farmers will only adopt technologies that deliver value or reduce risk. Even if the technology does provide significant value the may also be a challenge in connectivity; how do all these devices connect together and to the cloud?
I was thinking about this just last week, when I was demonstrating livestock management software to a group of farmers. It was a wintery day with snow on the hills, and we were standing out in the sheep yards, protected from the sun (but not the wind) by the steel roof.
We had just demonstrated how trivially easy it was to capture information about animals and synchronise it onto a smartphone, and now we were going to show the same information synchronised to the cloud, through a web browser and a mobile data connection.
We pressed “refresh” and we waited – and waited.
Did I mention that I had plenty of time to think?
Of course, eventually the page loaded and the demonstration carried on, reasonably successfully too. The farmers of course took the opportunity to point out that their remote yards would have no coverage at all!
We face a variety of different connectivity challenges in rural environments:
- Low bandwidth and high latency connections;
- Connectivity black-spots and intermittent coverage; and
- Areas with no connection at all.
Low bandwidth and high latency
In many cases, rural networks (fixed or mobile) provide significantly lower bandwidth than those in the cities, or they may have significantly higher latency (time to respond). For some applications, this doesn’t matter at all, but for others it may be a show stopper.
I watched a drone software manufacturer demonstrate the features of a very smart unmanned aerial vehicle (UAV). On-board software used the GPS to navigate the device over a flight plan we outlined using Google Maps. A multi-spectral camera captured high resolution images at 1cm per pixel, and the device transmitted the images to the cloud while it was still flying. That’s pretty incredible!
Transmitting images as they are captured is a great innovation. The intensive processing power required to stitch images together and analyse them is provided by a remote data centre and applied to data arriving from drones around the world. It also frees the operator from messing about with memory cards or thumb drives and processing software.
This works just great on the 4G mobile networks in the Salinas Valley, California. Not so much in rural Wales where the mobile networks drop back to GSM or EDGE, with significantly lower throughput. Still, there may be opportunities to address this. If the drone has sufficient memory, it could cache images until bandwidth improves, or when it returns to a connected base. Alternatively, flying at 400ft above the terrain may provide better coverage than we experience on the ground.
High latency rural connections provide a different challenge. We typically experience an increase in latency with satellite connections, as signal travels through a far-distant satellite, to and from ground stations. Satellite connections can still be very high bandwidth (high data throughput), but have a measurable delay, as you’ll notice if you use a satellite based phone service – “over”.
This latency won’t affect most applications to any noticeable extent, but consider the case of a livestock auction happening at a remote location. Smart auction software allows internet bids to be synchronised with the bids happening on site, right down to subscribers hearing and seeing the auction in real time. A delay of one or two seconds in this case becomes significant, so auction planning and operation needs to take this into account.
Mobility “black spots”
Much of the rural countryside lacks mobile coverage. Hills, trees, and buildings can all affect local reception, especially as distance from a cell tower increases.
This limits our dependence on these networks for “life and death” situations. There’s a reason why Search and Rescue services encourage hikers and hunters to use Personal Locator Beacons (PLBs) rather than rely on mobile phones when in remote areas, and the same should apply to tools that are triggered when an ATV rolls on the farm – the vehicle could well be in a gully outside of cellular network coverage.
We should also plan on intermittent connectivity when developing apps for the farming sector. Requiring an internet connection to start or use an application will limit its use to favourable environments close to home where that coverage exists. An answer of course is smart synchronisation of relevant data, when devices come back into network coverage.
I was quite confident demonstrating our livestock management application remotely for example, as the list of animals had already been stored by the mobile device and most importantly the application was designed to work disconnected and sync later. If there was no coverage, the new data I had captured would be sent to the server when a connection became available.
Bringing your own network
There are always spots with no connectivity: sheep yards in a valley between steep hills; installing nitrate sensors in a creek in a ravine; even equipment in the shadow of a large steel building. What options are there if we must install our technologies where there is no connection?
It turns out there are a range of options, but the need to be technology savvy and the cost of equipment rises in these cases.
- Consumer networking such as Wi-Fi and Zigbee mesh networks operate at 2.4GHz, providing line-of-sight connectivity over relatively short distances. Trees, hills, and even buildings can block these signals. Wi-Fi’s real strength is that it is a relatively low-cost and easily deployed option. It’s less suitable where extreme battery life and low power is a requirement.
- There are also specialist wireless networks. These often use lower frequency radio transmissions, operating at 433, 868, or 900-921 MHz. They require a larger antenna, but these frequencies can travel longer distances without increasing the power requirement, and may be less prone to blockage by trees. Some of these networks transmit point-to-point between sender and receiver, and others form a mesh communicating between multiple devices.
- New standards are evolving for low-power, long-range wireless networks such as the LoRaWAN and SigFox networks. Telecommunications providers are starting to adopt these standards to provide specialist connectivity for “Internet of Things” devices and remote locations. Where these are used at low frequencies in rural networks, they may well provide better coverage for monitoring devices in previously poor locations.
- Finally, we might consider using mobile phones to “collect and deliver” data from monitoring devices in places where there are no coverage, but where coverage spots are regularly visited or passed. A remote sensor with some memory and Bluetooth Low Energy capability might deliver summary information to your smartphone when you pass by. This is the same approach that is used by activity sensors you wear on your arm – data is captured, and delivered to your phone at intervals.
So there are possible solutions even to those areas with currently poor or non-existent coverage – but many of these require some skills, choices or trade-offs, and often a higher investment cost, and this makes adoption of technology more challenging. Companies developing new products can’t cater for every possible network technology, and may end up integrating a just subset of connectivity options.
Our team at Rezare Systems often discusses these challenges and the evolving set of solutions, because these will enable some of the software solutions we “brain-storm”. We’re not telecommunication providers or hardware developers, but we are very interested in how connectivity influences adoption of new technology and our ability to capture and leverage data.
Have you had experience with connectivity issues in rural applications? Can you share advice for farmers and rural professionals about some of the suggestions above and their alternatives? Let us know your thoughts.