Robotic Anthropology – Roombas

I know full well I can just read the manual or Google the questions I have when it comes to how 2 Roombas will interact in a small environment, but I have so much more fun black box testing and theorising how something works.  Besides, the most exciting phrase in science isn’t “Eureka!”, it’s “Now isn’t that interesting”.

All that said, into the badlands I go to see how they interact.  Keep in mind I’m deliberately not Googling or RTFM’ing.

IR Towers - The Borders of the Village
IR Towers – The Borders of the Village

So the Roombas can be prevented from wandering too far by erecting one of these towers.  How they work is pretty straight forward I guess, it’s invisible and directional, so probably not radio based.  More than likely infrared.  Now while my arduino is out of commission I won’t be capturing any IR anytime soon, but the day is saved by my cell phone.

For some reason (which I’m just going to put down to a need to operate in low light condition) the IR that is normally deliberately filtered out or just isn’t picked up by normal cameras and the human eye is quite easily picked up by the humble cell phone.

I see you there, tower
I see you there, tower

So that little purple light there is blinking away at me.  Looking at the casing of these towers, it looks like I’m right.

Instruction for Tower

 

Actually, if I record the video then apply a nifty little algorithm (which for some reason I can use but can’t remember the bloody name) I could record or decode it then save it as an audio file replay it through my my phones headphone port into an IR LED to recreate it.  Light to sound to light  More on that later.

Back to the cultural workings of our robotic tribe.  The villagers have boundaries and respect them.  Let’s take a look at their home.  While I have 2 charging stations/docks/huts for these guys, I’m only going to work with 1 for now.  Again looking at our cell phone you can see the IR blinking away.

In fact, you can see 2 of them pointing off on 2 different angles.  So it looks like each of these base stations have a directional beam at 30 to 45 degree offsets to create a V shaped homing beacon.  Now the strange part is that in both situations, I didn’t really expect the IR to be coming from these places.

What the hell is this then?
What the hell is this then?

So now I find myself wondering what the hell is this thing?  There’s one on each device and the Roombas too.  It’s a reflective inverted cone in a clear housing.  So it’s either for focusing any light in a 360 degree view into a single receiver or sending a light from a single point to 360 degree spray.  We’ll come back to that in another post.

So, these Roombas have homes but how do they feel about sharing?  The easiest way to find out is my favourite part.  Experimentation and observation.  This is where it gets into the anthropology a little more.

I know from an earlier experiment that Roombas are a bit slutty.  Any port in a storm.  They don’t care whose dock they use.  But when 2 Roombas vie for a place to call their own, it appears second place is a very bad thing.  As soon as one Roomba occupies the dock, it covers the docking IR and leaves the other Roomba out in the cold.  Out of sight, out of mind the lonely Roomba wanders off looking for a home.  The loser in this game of musical chairs is unfortunately sentenced to death as it wanders into the wilderness looking for the home it will never find.  Slowly it loses charge and dies.

This is a horrible thought really and I’d feel super bad about the whole thing (in fact I did feel bad and retrieved the poor lost little guy) but then I took a closer look at our noble worker.

That evil little bastard has been watching me.  Both of them have.  This entire time.  LOOK AT IT!!! THOSE SUNKEN DEMON  EYES!!!  Now I don’t feel so bad at all.  Yes I know that those are IR emitters, but I’m going to assume there’s a receiver there to pick up reflections so it can anticipate walls (which it then slows down before bumping).  Technophobia aside, with a fresh fill of Robo-hatred I’m good to subject it to some more torture.

Roombas are, in fact, inconsiderate assholes.  I mean, we’ve established that they can both see each other and yet neither one wants to concede and allow the other clear passage.  Like Robin Hood’s first meeting with Little John, it’s going mean a butting of heads.  So now that I understand just how bad these demonic asshole sluts suck, I don’t feel anywhere near as bad when I put one in a tiny room with no doors (in my opinion the most understated and epic threat ever delivered considering it was death himself who spoke it).

While my disdain for these horrid robotic imps has garnered the application of torture a little easier than I would like to admit it appears that, when faced with a small room and no way out, lady luck smiles on them still.  The corner swiffer along with their nature, either through great programming or dumb luck, lead the Roomba to gently nudge the walls of it’s invisible prison until it Shawshanks itself out of purgatory and off to fornicate in the warmth of the closest charging station.  Yours, mine, Roomba 2’s, or the neighbours.

In conclusion, I’d say the Roomba are a robust tribe of workers who display a blatant disregard for their own kind.  A selfish robot with loose morals and little intelligence.  Screw these little assholes.

 

Next post:  Building the arena…

 

5 comments on “Robotic Anthropology – RoombasAdd yours →

  1. I… I have a Samsung Navibot… I’m not even kidding when I say whenever I turn it on, it runs out of the lounge, round the corner, down the hall, round 2 more corners, and hides in the laundry where it then gets lost and shuts down. It’s like playing hide and seek with a very tiny alzheimer’s patient.

  2. Hmmm…and here I was thinking there was going to be some really interesting doscussion about Artificial Intelligence and debates about “Do Robots Dream” and what donthey think about…well working with the material presented I would say in yje case of the Rumbas you’re projecting jthe current neoliberal hyper competitive hyper individuated social world you live in and obviously use as the source of most of your thinking onto a vacuum cleaner and the psychological reductionist approach to the Samsung suffers from the all too popular practice of contemporary indistrialised society to medicalise things. Btw I think when robots can/do think and dream they’re goi g to look at this and be somewhat disheartened regarding the inverse question “can humans think?”. 😉

    1. Valid points and a lot more thinking than most. I make no attempt at covering the deeper intricacies of robotic intelligence as this has been well covered before by those more studied than I. What I would suggest is that a basic few rules create behaviours that can approximate actual anthropological qualities. Something more akin to the paper “Elephants Don’t Play Chess” written by the professor at MIT (Rodney A. Brooks) who has been pioneering a new approach to AI and learning through the concept of a limited set of basic rules allowed to proceed and develop on a sped up timeline. I’m not doing it justice by describing it here and I think it would best be understood if read by yourself. http://people.csail.mit.edu/brooks/papers/elephants.ps.Z

      One thing I would like to make clear is that this is by no means the extent of my understanding of robotics, but a whimsical moment at play to help steady my mind. Do I truly believe the roombas have feelings. No. Do I anthropomorphise them as a bit of a giggle? Yes. But I do believe there is substantial cause to re-examine robot and artificial intelligence in line with Brooks theories.

      1. I appreciate the giggle factor and actually enjoyed the anthropomorphism a lot. I was on a bus with my iPad warning me of low power when I “penned” my hasty response and I started off trying to be finny but wanted to appear smart to and looks like I failed at both (my post also reads like I am typing with a hammer).

        Reading through Professor Brooks paper I am struck by what a profoundly functionalist approach it seems Robot AI developers are taking. If the efforts outlined in the paper were intellectual/theoretical undertakings they would be called out for the seeming crass positivist epistemology they are being guided by. I found it very interesting as an insight into where Robotcs is at currently but I think the work suffers a bit from no apparent engagement with social or cultural anthropology and the interdisciplinary work these fields are doing in collaboration with neuroscience (the absence of any sense of this work is quite striking comsidering that Artificail Intelligence is what’s at stake.). The cursory dismissal of discussions of intelligence that Brooks makes in his paper is unfortunate and I was initially going to reccommend you to the work of Ingold on anthropology and animals when I read the title of the paper mentioning elephants but I think the mechanical emphasis of Brooks et al probably indicates I’d be “speaking at cross purposes”. On an interesting tangent its worth noting that where Brooks speaks of debates about intelligence being mainly concerned with finding a satisfactory definition there are those of us who are interested in how can the artificial version of what ever it is decided intelligence is, be taught to be inculcated with a matrix of perception that matches that which we call culture. Understanding that the capacity to differentiate light wave length variations (colour) is a cultural one and that is why colours do not often match cross culturally, culture will have a very important place in AI development. The fact that there seems little (possibly no) engagement thus far to this ends might be part of yhe problem confronting the people developing it. I’m thinking here also that the most “effective” forms of intelligence are social and socially gained.

        1. A fair call, I do most of my editing on my phone while in the elevator or train. I wouldn’t say you failed, a certain number of mistakes and the like are generally accepted on the net and it would be to invite the same scrutiny on my own posts to pick you up on any of it. Half the reason why I’ll avoid getting too in depth about AI conversation here as I feel I would not do it the justice it deserves in this forum. The other half is that it’s supposed to be fun here 😛 . I will admit Brooks does appear to have a point to make in his paper which derails it some. One thing I enjoy about the simpler stuff is that through no great complication the hardware develops a personality of its own. I won’t even go into the reason I still put electronics in the fridge as a last ditch effort to fix it. Sometimes I start with that. And no it has nothing to do with the hard drive thing 😛

Leave a Reply

Your email address will not be published. Required fields are marked *