01-21-2021, 07:43 PM
I still have serious questions about the overall purpose of this device.
Is it a testbed for post-Internet decentralized wide-area networking? That's what it sounds like. If that's the case, then we're really bikeshedding this by worrying about irrelevant details like how large the screen should be and whether or not it needs a front-facing camera. If we're going to make any progress on the extremely involved technical problem of self-configuring mobile mesh networking - a problem that no one has yet solved to the degree needed to make it scale to continent-spanning or even city-spanning networks - then we need to focus on that and let go of the rest. If this is what the PineCom is for, and it needs to be low-cost, then it shouldn't bother having any cameras, extra fiddly bits, and maybe not even Wi-Fi; all of the extra BOM budget should go into putting a decent software-defined radio transceiver in it, ideally backed by an FPGA for DSP tasks. Yes, that's expensive territory, but that's the territory we're in if we're trying to push the limits of mobile wireless networking. My first choice of hardware for this sort of project would be the Fairwaves XTRX, though since that costs about an order of magnitude more on its own than the target retail price for the entire PineCom, we'll obviously have to start smaller. The AT86RF215 chip, or whatever similar part happens to be available, might be a good start.
LoRaWAN is not going to be an effective solution, because it's designed for polling sensor networks over relatively small areas (a few kilometers, not tens or hundreds) and generally relies on internet-connected gateways as well. Additionally, LoRa is proprietary and its chips are expensive, which adds pointless complications to this already-very-complicated project. LoRa's link budgets are impressive, but an FPGA-backed SDR could relatively easily outmatch those numbers with the right spread-spectrum techniques, and with the specifics of those techniques published openly, it will actually be possible for other types of devices to interoperate with this one, which will be crucial.
Regulatory issues are going to be another challenge that I'm not really an expert on. I don't expect this will be able to get certified as an FCC-compliant consumer wireless handset, but it might not need to be if it, like other SDRs, gets sold as test equipment (which it will be).
Is it a testbed for post-Internet decentralized wide-area networking? That's what it sounds like. If that's the case, then we're really bikeshedding this by worrying about irrelevant details like how large the screen should be and whether or not it needs a front-facing camera. If we're going to make any progress on the extremely involved technical problem of self-configuring mobile mesh networking - a problem that no one has yet solved to the degree needed to make it scale to continent-spanning or even city-spanning networks - then we need to focus on that and let go of the rest. If this is what the PineCom is for, and it needs to be low-cost, then it shouldn't bother having any cameras, extra fiddly bits, and maybe not even Wi-Fi; all of the extra BOM budget should go into putting a decent software-defined radio transceiver in it, ideally backed by an FPGA for DSP tasks. Yes, that's expensive territory, but that's the territory we're in if we're trying to push the limits of mobile wireless networking. My first choice of hardware for this sort of project would be the Fairwaves XTRX, though since that costs about an order of magnitude more on its own than the target retail price for the entire PineCom, we'll obviously have to start smaller. The AT86RF215 chip, or whatever similar part happens to be available, might be a good start.
LoRaWAN is not going to be an effective solution, because it's designed for polling sensor networks over relatively small areas (a few kilometers, not tens or hundreds) and generally relies on internet-connected gateways as well. Additionally, LoRa is proprietary and its chips are expensive, which adds pointless complications to this already-very-complicated project. LoRa's link budgets are impressive, but an FPGA-backed SDR could relatively easily outmatch those numbers with the right spread-spectrum techniques, and with the specifics of those techniques published openly, it will actually be possible for other types of devices to interoperate with this one, which will be crucial.
Regulatory issues are going to be another challenge that I'm not really an expert on. I don't expect this will be able to get certified as an FCC-compliant consumer wireless handset, but it might not need to be if it, like other SDRs, gets sold as test equipment (which it will be).
