Hi Wildlabs,
My name is Mark Butterworth, and I’m a PhD student at Cardiff University. I am researching methods for remotely monitoring animal traps that are placed in the rainforests of Malaysia. The Malaysian forest is a dense evergreen forest that has extensive canopy cover. Cellular network do not cover the forest and due to thick canopy cover satellite communications are precluded. The traps are placed over a 20-mile radius, and transport between each trap is slow and difficult. Given the remoteness of each trap, an IoT mesh network is impractical as each node would need to be visited regularly for a battery change. I am intending to use low power, low bandwidth, HF signal using digital signal processing to facilitate a low data rate reliable communications medium. Just wondering if anybody has knowledge of anything similar or any other ideas that I should explore?
11 November 2020 8:39pm
Hi there,
My advice is always the same: use LoRaWAN, if there is no infraestructure there (towers, gateways) use just LoRa, it is a radio system under the 1Ghz band, it is great for low power consumption and given the fact that it is under the 1Ghz it won't have trouble with noise. With a good board from Microchip or Semtech you can reach up yo 15KM of signal from antenna to antenna. With a proper Mesh configuration you can cover a lot of terrain with cheap equipment. The RN2903 module from Microchip can be easly integrated to any hardware having TX/RX connections (serie). You can use an ASCII command set both to use the LoRaWAN stack or sending / receiving an arbitrary number of bytes (I use this scheme a lot), even the power used to transmite messages can be specified.
Cheers, I will be posting about this soon under the technology menu and / or the blog.
12 November 2020 12:50am
We had similar issues on a project in the Philippines where we were monitoring water levels in rice fields for the International Rice Research Institute. The rice paddies were dispersed and it was difficult to get to many of them. We implemented a hybrid wireless system where we had a "local" network using 900 MHz wireless sensor nodes to monitor clusters of paddies located within a 1 km radius. We then had a single gateway node located as equidistant (central) to the cluster as possible that had both the 900 MHz radio and a GSM (2G cellular) radio to send data to a server via SMS.
In countries where GSM has been sunset (obsoleted), we use CDMA (3G cellular) which are a bit more expensive than 2G but much more widely available. The central gateway has high power requirements since it needs to be on all the time to receive transmissions from the local sensor nodes.
Depending on your requirements and location, you may want to consider just implementing a bunch of 2G nodes that only transmit when the trap is triggered. Otherwise, you can keep the cellular modem off the rest of the time saving power. 3G is also possible but might get costly. If the nodes are pretty dispersed so you don't have clusters, this might be the simpler option.
Also you'll need to figure out a way to pay and manage the SIM cards. We had a separate application just for this purpose so that we could pay as soon as the SIM cards would run out of money. We mainly used prepaid SIM cards.
Hope that's useful :)
Akiba
18 November 2020 7:03pm
Hi Both,
Thank you very much for your responses. It has certainly given me something to think about. CGM11 I look forward to reading your blog. I will report back on what my chosen method turns out to be.
Akiba, sadly cellular networks are out. But, I like the idea of local 900MHz beacons reporting back to a gateway node. I will keep you all posted.
Thank you
Mark
20 November 2020 12:13pm
I would also recommend a LoRa or LoRaWAN mesh network. It's the best solution for low power scenarios. You will need to test the distance aspect as forests are notorious for blocking signals. The counter measure to that can be to try to use height - get above or near the top of the canopy if possible. So it might mean going trap to canopy and onwards. Good luck with the project.
Eric
www.wyldnetworks.com