Smart Cow Collars • Hacker News Board

Re: Smart Cow Collars

#41

Post by davisoneee » Thu, Sep 24, 2020, 10:53 PM UTC

I'm in the process of writing up my thesis, focused exactly on this domain; collars for detection of eating, rumination, and oestrus behaviour (welfare and fertility), from Afimilk Silent Herdsman (collar system was originally a spinout from my University), combined with Automatic Milking Systems (robots) from Fullwood Packo (the Merlin).

In general, the cows are more driven to the robotic milkers by the feed they recieve while in the machine, rather than the feeling of relief (these systems generally provide an alert of cows being lazy...ones that haven't been milked in 10+ hours and need a shove), however these systems _do_ provide both improved health and improved productivity.

Even in experienced dairy herds, handling the animals and shoving them all to the robot at the same time causes stress. Having these machines running constantly (1 machine can handle a herd of ~50) lets cows regulate their own pattern, which can benefit herd dynamics as well (more timid animals can self regulate to a quieter period away from pushier animals). Farms can easily see 10-20% improvement on milk yield.

It also has the benefit that animals can go more or less often...you may have some animals who are comfortable with 3 visits a day, some comfortable at 4 (or even 5 if they have really high yield).

Re: Smart Cow Collars

#42

Post by BenjiWiebe » Thu, Sep 24, 2020, 11:51 PM UTC

floren wrote: post #32

iamwpj wrote: post #31

So I grew up on a dairy farm and we had electronic id badges since the mid-90's (actually there was a Surge system before that, but the main one was a DeLavel in 1997). We had a variety of issues that these were good for, but the gist is that the cow's identity is scanned and matched with statistics regarding her milking. Everytime she entered the parlor she was id-ed and then the system could be configured to separa…

The bands also look poorly-suited for interfacing with the standard lockups used for artificial insemination and preg-checking.

Our dairy cows have activity tracking (heat detection/poor health detection) collars, and they work fine with lockups. The cows do turn their head a bit differently to make sure the collar clears, but they learn quickly and we've had no deaths yet (200ish cows, 5ish years of using the collars)

Re: Smart Cow Collars

#43

Post by zzarzzur » Fri, Sep 25, 2020, 12:27 AM UTC

kabouseng wrote: post #24

Reminds me of a tale here in South Africa. An ostrich farmer contracted an electronics company to develop a product to determine which of his ostriches were sick. The company realised that a sick ostrich do not eat, and as such developed a neck band with a tilt sensor, to detect which ostriches were not eating. Great, problem solved. Except now you have to seperate the sick ostriches from the healthy ones. And the fa…

Sounds to me like the farmer paid for research on animal behavior.

Re: Smart Cow Collars

#44

Post by 4043D » Fri, Sep 25, 2020, 1:04 AM UTC

iamwpj wrote: post #31

So I grew up on a dairy farm and we had electronic id badges since the mid-90's (actually there was a Surge system before that, but the main one was a DeLavel in 1997). We had a variety of issues that these were good for, but the gist is that the cow's identity is scanned and matched with statistics regarding her milking. Everytime she entered the parlor she was id-ed and then the system could be configured to separa…

i currently own a small cattle farm in rural australia and you are correct collars really are not a good solution not only do they rub on things but they love chewing on things especially rubber or plastic wires and electronics they would 100% chew on each others collars. We already use rfid ear tags which are small and unobtrusive and they still manage to occasionally rip those out, rubbing on things. If you were mustering a herd on a large holding say a few thousand hectares this could possibly be of some value at least in initially finding the herd however at least in my country if you are on a multiple k hectare property there is a good chance you wont have any sort of mobile coverage, also cows are creature of habit they have there daily grazing route they eat one place rest in another depending on time of day and weather conditions farmers typically know this. I will also add that the value proposition I have seen with other smart tags that are actual ear tags that do temp / heart rate etc etc does not add up at least with my small holding. the price per month of $5 per cow would eat into the already somewhat slim margins, the data is great but it isn't anything i would not notice by visually checking stock daily. im very skeptical on the faas model in general. Farmers at least in my country would balk at spending that sort of money a month that doesn't directly add very much in the way of weight / condition per month, at least from a beef cattle stand point. (perhaps on a prize bull you could justify the monthly fee)

Re: Smart Cow Collars

#45

Post by 4043D » Fri, Sep 25, 2020, 1:13 AM UTC

LeifCarrotson wrote: post #18

tantalor wrote: post #12

Earlier quoted context omitted.

"Pulse" is extremely vague. A pulse of what? Electric shock?

Yes, the sound of a pulse of electricity similar to the fence they're already familiar with. Cows are typically kept in pens with high-voltage electric fencing. The energizer fires a 3,000V pulse, lasting about 300 microseconds, every second or two. This 1Hz impulse is often audible as a repetitive hum if you stand near the wire. By replicating the sound I expect you could suggest the cattle move somewhere else.

my cows are too smart for electric fences. i set one up to separate 2 herds with a bull in each herd. the bulls will fight if the ladies are in season. they watched me set up the fence worked out that I placed the control box up at the same time the the fence got all zappy they pulled all the wires out of the control box. Cows are a lot smarter than people give them credit for.

Re: Smart Cow Collars

#46

Post by tyingq » Fri, Sep 25, 2020, 1:47 AM UTC

On my grandfather's ranch, he had those rows of pipes laid in the ground with spaces between them at the road entrance. The cows couldn't walk across them, but your car would do just fine.

When he made a back entrance, he just painted white lines instead of the more expensive pipe setup. He'd learned from neighboring ranches that once the cows understand the pipes, they won't walk across the painted lines. And, I don't quite know how, but new cows learn from the existing herd not to try the white lines.

Re: Smart Cow Collars

#47

Post by chrisco255 » Fri, Sep 25, 2020, 1:55 AM UTC

tyingq wrote: post #46

On my grandfather's ranch, he had those rows of pipes laid in the ground with spaces between them at the road entrance. The cows couldn't walk across them, but your car would do just fine. When he made a back entrance, he just painted white lines instead of the more expensive pipe setup. He'd learned from neighboring ranches that once the cows understand the pipes, they won't walk across the painted lines. And, I don…

Yeah, and there's lots of analogues in human behavior, to be sure.

Re: Smart Cow Collars

#48

Post by rapjr9 » Fri, Sep 25, 2020, 2:23 AM UTC

I spent a year or two doing research on electronically herding cows. Dean Anderson at the USDA holds patents on it (which are probably about to expire, or may have already). We worked with him some and with CSIRO in Australia who I think also holds patents (maybe Halter has licensed them and collaborates with them?) The main reason there is interest in this technology is that fencing is expensive. Think many kilometers of fencing, the cost adds up, and they have to be maintained as well. Also areas next to water are difficult to keep cows away from and fences don't work as well there and the cows tend to damage the soft land. We tried using only sound to guide the cows, with no Pavlovian conditioning and found it didn't work very well. CSIRO used electric shock and found it worked very well, but also upset the cows. Dean and CSIRO then tried electric shock for Pavlovian conditioning, then switched to just sound for earlier warnings before finally shock was applied and found that the cows would eventually respond to just the sound and got less distressed. It's not perfect though, some cows are strong willed and will ignore sounds and shocks and just keep going. If the rest of the herd moves away, they will tend to rejoin it though.

There are some problems with the technology. Temple Grandin visited us and said she didn't like the idea of people actually visiting the cows less. Collar worn sensors can't tell if a cow has a rock in its hoof or detect all diseases. With less human observation the cows may suffer more. I've seen one cow start chewing on another cows collar to help it try to get it off. GPS antenna have to stay facing skyward and can be blocked by trees. Some of the later research switched to ear tags instead of collars because they are less of an impedance and less likely to get damaged by the cows. It's more difficult to get everything you need into an ear tag though. There have been several attempts to start companies based on the technology, all of which failed so far. Halter appears to be the latest.

We stopped doing research when a prison company approached us with an interest in putting the collars on prisoners.

Re: Smart Cow Collars

#49

Post by codezero » Fri, Sep 25, 2020, 3:15 AM UTC

I loathe to see us getting more and more experienced at killing and controlling animals - go check out that thread on crow intelligence - we should stop and think for a while before we laude our superiority over the world.

Re: Smart Cow Collars

#50

Post by jessaustin » Fri, Sep 25, 2020, 3:51 AM UTC

ogre_codes wrote: post #14

fiftyfifty wrote: post #6

Earlier quoted context omitted.

I would love to see this kind of technology take off. In the western US there is barbed wire everywhere, even on public land.

Yes, drives me insane when I'm out in the middle of nowhere enjoying a view only to stumble across a stretch of barbed wire. I'm skeptical public grazing is a great use of public lands regardless, but now that we have this technology it should be mandatory that anyone who grazes on public lands use this instead of adding more fencing.

Cattle shouldn't be grazing on public land in Western USA, period. Even if this technology worked over such vast spaces (I'm sure it doesn't), it would be a mistake to let it become another excuse for people like Cliven Bundy to break the law.