Comment by bennyg | Hacker News Reader

Comment by bennyg | Hacker News Reader

Parent Full thread

As an iOS developer who has spent the last 9 months making beacon enabled things for my employer, I personally think beacons are shit. Some of my findings over the last 9 months:

- Estimated battery life was estimated very poorly.

- Nobody likes things they didn't ask for getting sent to their phone.

- Indoor navigation that relies on iBeacons only is going to shoot you in the foot. I recommend www.indoo.rs as the only solution that got close to what we needed (though it's still pretty darn immature).

- Speaking of indoor navigation, you need a ton of points to do it well (whether that's beacons or wifi routers, you decide).

- Using beacons for granular location-aware uses might as well go out the window if you're looking for accurate and precise readings around ~1-3ft.

- Did I mention power yet? All of the beacons we started with around 9 months ago have dead batteries. Even with conservative power modes set. The only beacons that don't are RadBeacons USB sticks that plug into wall outlets. And that's only because they don't have batteries.

I really, really want them to be good. But they die too quickly, and don't provide granular signal data. Hold the phone between you and the beacon and get a decent signal, then turn 180 degrees so that you are between the phone and the beacon. Apparently you've moved 45ft.

(Disclaimer: I work for Estimote) Indoor location works much better if you utilize data from other sensors in the phone (like accelerometer etc) and apply machine learning to make better sense out of all the signals. I'd suggest you take a look at our SDK that works surprisingly well even with just 4-5 beacons. You can also map your room with our iOS app by just walking around it. You can find it here if you're interested: http://estimote.com/indoor/

We've also implemented some adaptive algorithms to optimize the battery life in our beacons, so with conservative settings they should work for 5+ years now.

The beacons my company received from you guys must have been out of the norm then. All 3 Estimote beacons are dead, and they all died around the same time (one in super conservative mode, one in what should be a normal use case, and then one in max power mode).

I was excited when I saw your indoor SDK, but it's too simplistic and not realistic enough for the production apps we have and are building (walking around and mapping every single room in a hospital is untenable).

I don't mean to bash your company at all here, so I'm sorry if this comes off that way. I also found that almost all clients scoff at ~$33/beacon when you're pitching something in the 10,000 beacon range. Even then, for $30/beacon I can get ones that run off the wall circuit and last until that's off (for all intents and purposes, longer than 5 years).

Obviously at 10,000 beacons we're talking enterprise pricing, but even if you were to get that to around $5 that's still too much for what's essentially a CR2032 battery and a tiny radio chip.

I'm excited to see where Estimote takes beacons though, since they seem to be doing the most R&D type pushes with this technology compared to the other manufacturers.

Thanks for your feedback! Sorry to hear about faulty beacons, what happened definitely doesn't sound normal. I'd suggest to contact our community team at contact@estimote.com and they're probably able to replace these beacons with new ones.

Indoor SDK is still in very early phase, but we're pushing very hard to roll out a series of updates with new features like multiple rooms support, cloud integration and even better accuracy, so you can expect more comprehensive solution in the near future.

And if you're interested in bigger deployments we also offer beacons in bulk quantities with much better pricing.

We're treating R&D very seriously as we're thinking of beacons as much more than simply hardware. There's still a lot of ideas to explore and we try to support the community as much as we can, so I hope you'll give it another chance.

Thanks again for sharing your thoughts. That's extremely important to us and makes it possible to address issues people have with beacon technology ASAP

Yeah no problem, I'll talk to my director at work to reach out about the beacons since he's been the point of contact with Estimote so far. Thanks for following up and fielding my concerns/criticism. Like I said earlier, I honestly can't wait to see what you guys come out with next in advancing this technology.

(Disclaimer: Work at TAH) I totally agree to what you are saying. We have had a similar experience working with our devices and SDK.

http://tah.io/get

I absolutely agree with this. I pushed beacons really hard internally, and got a couple projects with them. Unfortunately, they're an absolute pain in the ass as far as reliability goes. One iPhone 5 detects the beacon immediately, the other has to sit around it for 3 minutes before it's detected, one can't detect it at all. Makes life pretty difficult when my only answer is "Is your bluetooth on?".

I hope that some OS level improvements give us some free reliability and performance increases. Supposedly Android recently got some updates re: beacons.

We at our company also tried to develop solutions based on beacons for our customers. The major hindrance to acceptance of beacons is people here almost always switch off bluetooth in their phones. I think phone designers should give 2 different switches for switching bluetooth on/off in the UI. One for traditional bluetooth and another for BLE. That way customers can be pursuaded to keep ble on.

Ahh yes, I totally forgot about the phones maybe, sometimes, probably could this time work and range the beacons. Half the time my suggested method for getting the ranging to work is to delete app, restart phone, turn bluetooth on/off, redownload app, pray, sleep, buy new phone... It's unbelievable how ridiculously inaccurate the system software is using CoreLocation out of the box. I've written a wrapper for it (https://github.com/Intermark/Buoy/) to make internal dev a lot easier, and even then it's a pain in the ass to find why this prototype isn't working.

I wonder whether the fault is at the iBeacon API/CoreLocation framework or the hardware beacons used? From your description above it sounds to me the hardware beacons are of questionable quality (low quality batteries and bad antenna maybe?)

I thought this too. But even with the plugged in RadBeacon (http://www.radiusnetworks.com/ibeacon/radbeacon/) and no EM interference between it and the phone, I was still getting extremely spotty results.

It's most probably a combination of all random variables (EM interference, beacon hardware, software, phone BLE antennae, etc) that makes it so frustrating and practically non-deterministic.

Have you tried emulating iBeacon with iDevice or a Mac? I'd consider those to have high quality Bluetooth chips and great antennae. If it is still not ideal, the blame could be on the software stack or maybe the whole idea is flawed :|

> ... then turn 180 degrees so that you are between the phone and the beacon. Apparently you've moved 45ft.

Why don't you fit the beacons on the ceiling?

You're right - antenna placement is super important with any real-world practical implementation. However, placing them on the ceiling/floor is not always an option.

Is there a specific reason for the coin cell batteries?

I would be much happier paying the same amount for something that looks like a AA battery and will last a few years.

I got some demo units from these guys[1] that use AA batteries for that very reason.

no affiliation. haven't tested the units yet. [1] http://www.bluecats.com/

I think it's size. Most beacon companies are doing what the phone companies are doing - slimming everything down.

Batteries last 5 years in very dry packages under 90 degrees Fahrenheit, but this might be a nice place to mention new coats for batteries; PD&D Magazine mentions them as cutting the electrical damage bit out of the swallowed batteries risk. Then I would ask about the naivety of supercaps, now with catastrophic conversion -over- 110 degrees Celsius (or when ingested, I suppose.)

Reply on news.ycombinator.com