Idea: SCRAPi - Social CellPhone Robot Avatar

This page represents an idea by Andrew Noske. Date of idea: ...........9/July/2011 Date added here: ....22/Feb/2012 Status: ..................Made contact with a few university groups (which was perhaps futile of me)... and also put the idea onto Quirky, but no luck. In the meantime, it turns out a bright girl called Claire has developed something very similar on her own! Check out: Botiful - released on kickstarted July 2012 and looks like it's shaping up to be a very successful project! :-) Ideas I've publicly posted here I'd love to implement myself, but I know I (realistically) don't have enough time and/or resources. While most people keep their ideas secret, I'd prefer someone else implement and benefit from this idea rather than it fade in my head and never happen! If you like this idea or know of a similar one please e-mail me at . :-) For more info and a list of my public ideas visit: Ideas.

Conceptual picture.

• 
• 
• 
• 
• 
• 
• 
• 

Executive Summary:

Modern smart phones, like the iPhone and Google Android phones, have been used for all manner of exciting new applications. Inside their tiny frame, most smart phones have not only a high definition video camera, but an accelerometer, compass, and a global positioning system (GPS) - meaning almost every smart phone and tablet on Earth not only has the potential to see, but orient and position itself on a map. Several exciting accessories and uses for smart phones have already been discovered, but (to the best of my knowledge) there is no commercially available kit for turning the iPhone into a mobile avatar. Here I propose the design and construction of an affordable robot-like remote controlled vehicle which has a special area to mount a smart phone, and using special software, the user can use either a second phone or wireless internet to see through the vehicle mounted smart phone and use the screen to project their image and interact with people. This would be largely a novelty item, something parents might buy their children for Christmas, but from a social perspective such a robot represents a very interesting study into how people will talk to and interact with this little avatar. The proposed name for such a project is Social Cellphone, Robot Avatar Project - and the robot itself will be called "SCRAPi" for the rest of this document.

Fig 1: Illustration of what unit might look like
CG illustration showing what the remote control avatar may look like (right) and what it may look like to control the unit via another smart phone (left) - although this could just as easily be controlled from any computer with internet and a webcam.

Section 1: Inspiration

Fig 2: SCRAPi version 0.01
This crude little robot like device was constructed for under twenty dollars US, using the cheapest remote controlled car I could find, and the cardboard which came with the pack. Although the power and range of this RC vehicle was very limited, my first "preliminary test" on my housemate and friends proved to me its value as a novelty item. Even this cheap-as-possible version is something I was able to drive out of my room and into the living room to have a conversation and many laughs. Future iterations will involve a more reliable car and be tested outside for human interaction with strangers.

It was in 2010 I finally gave into peer pressure and forked out the cash for an iPhone. Like all phone contracts it seemed ridiculously expensive but, like all first-time smart phone users, I was excited by all the things it could do - most notably its ability to use Google maps and the use of Skype to call my family in Australia. I realized straight away the GPS meant it almost had a big brother quality, and sure enough, soon afterwards a range of applications to track location, including the powerful Google latitude, were released. Other applications, released around the same time, were more specifically targeted to help locate and/or lock down a missing or stolen phone. It was then I thought... wouldn't it be fun if a phone could drive itself to safety! I didn't give this idea any more thought until mid 2011, when my new housemate and I were discussing work, and I was complaining about how easily I was taken advantage of whenever I'm at work. My desk is in a very open area of the lab, so whenever I'm there I'm asked for help by students and fellow scientists.... and wouldn't it be great if I had a way to keep track of what was going on at my work without actually being there. What if there was some way I could have an avatar which I could drive around, to see what was going on.... and yet be able to retreat home where I was less likely to be disturbed and could get more work done!

Section 2: How it works

This idea isn't yet fully matured, however what I've written below are several ways SCRAPi could be made to function, to interact with people and to become a popular toy.

2.1. A Cheap RC Kit Model Car Plus Skype

Fig 3: Bottom range RC car
A photo of the fifteen dollar New Bright RC car I bought from Target, but here I want you to imagine the product instead being a 4-wheeled "robot" like that in Fig 1 and advertising that it's no ordinary RC car - it becomes a social robot once you mount your phone on top.

This represents something along the lines of what I've already done when "building" SCRAPi v0.01 (Fig 2), whereby I bought a $15 remote control car from Target (Fig 3), stripped off the top, and using cardboard from the box it came in, it took me about five minutes to construct a platform with robot-like arms and a "head" onto which I mounted my iPhone. To drive the car around I used the car's controller and projected my face and voice via Skype on my computer - and found using FaceTime from a second iPhone worked just as well. While I do fancy the idea of selling "build-it-yourself instructions" online, to make something a little more professional you'd probably want to approach a RC car company like New Bright nd hopefully they would give some reward for the idea (if it made money). What they would package and sell is a RC vehicle or kit which already includes the arms and a perfect little dock to fit an iPhone/Android. Like any other remote controlled car, this would come with a standard RC controller to drive it, although a nice feature would be if the controller could grip onto the bottom of a second phone using a spring loaded catch (see Fig 4) and/or rubber to secure it in place. I found trying to hold both the remote control (for driving) and iPhone (for FaceTime/Skype) at once was awkward - really you need both hands to drive the vehicle. In this form the unit would come with no software, it's pretty much just a remote control car with some instructions on how to use Skype.

Fig 4: SCRAPi controller with iPhone
Just like SCRAPi himself, the controller can have a slot which locks a second phone into place.

2.1.1 Making it look human and/or comical:

As a side note, I think a key to success of these avatar units is to make them look fun and to give them human-like features. If sold to a manufacturer, it's obviously up to them to make any final design decision, but personally, I think arms are very important - even if they never move - as they instantly make the avatar look more like a person/robot.... if it's just a car with a iPhone sticking out of the top it doesn't have nearly the same comic and human appeal to it. In addition to recognizable human features, making it look "cute" (and/or comical) will help further maximize the degree to which people relate and interact with these social robots. Part of the appeal of SCRAPi v0.01 is that he is cardboard and looks fairly silly - thus was very easy to get a laugh out of people as he drove around. One thing definitely worth testing is to make a version of SCRAPi which is constructed onto a teddy bear - making it something that people may feel surprisingly compelled to hug. Other ideas along these lines is to mount an iPhone over a big "Lego-like" man or even Darth Vader.... all in the name of comic effect. Another option would be to sell the final unit with an interchangeable head and/or arms so the users could customize its look or even play "dress up" to appeal more to a fashion conscious audience.

2.2. A More Sophisticated RC Avatar with Turning Head

Fig 5: Movement
CG illustration showing how the phone could tilt up/down and head could spin

This route is really just an upgrade on the first option. Whereas the first option could be sold at under $30, this car would be fancier and probably somewhere between $50 and $100. For one thing I image that, instead of wheels you would use treads - allowing for a zero degree turning circle to turn around and face people more easily (see also: Mecanum wheels - wheels that can go sideways as demoed in this video). Additional to this, I imagine the actual "head" of the robot - the part where the phone mounts - would be able to turn up and down, and hopefully in a circle too (see Fig 5). This means the remote control wouldn't just have the standard forward/back and left/right controls, but separate controls for turning the head up/down and potentially left/right to face people. Although not essential, it's interesting to think about incorporating moving arms, and perhaps even the ability to pick up and move things. Taking this further you could add sensors and a great example of this is the Wall-E robot (here). This combination of treads, and a turning head would make the social robot almost like a tank - and indeed perhaps the next prototype (SCRAPi v 0.2) be built on top of a RC tank to test the idea of the phone tilting up and down (like a gun turret would). Most importantly however, with a more expensive unit comes a greater range and more powerful motor / "engine". This more powerful motor means the unit would be powerful enough to carry a tablet - meaning you could project your face at almost life size and people wouldn't have to squint or bend down to see and talk to you. As a note on speed - a robot avatar vehicle shouldn't be fast for two reasons: (1) if it's too fast it might flip and destroy your phone and (2) if it's any faster than a human can walk, people are more likely to recognize it as just a remote control car, instead of a human's avatar. I haven't run tests yet, but my guess is that you can maximize interaction by approaching a person at about the same speed as another person might approach someone to ask a question, and to make sure people see him I think you'd want to have a flag which you could attach to the back.

2.3. Adding Phone Software to Drive the Car

So far, in section 2.1 and 2.2 we've just been talking about using 3rd party video call software like Skype to project your image on screen, and using the RC controller which comes in the box to drive the car itself. To make SCRAPi really cool however, it would be nice to actually use another phone and/or computer software program to actually drive SCRAPi. Ideally SCRAPi would work the phone on top of the car being plugged into into the car itself via a USB or other connection, with the hope that the smart phone can pickup movement instructions (left/right, forward/back) sent from the controlling phone (or computer).... and then translate these signals to move the car. While most RC cars are limited to drive within only a few hundred feet of the controller, by interfacing with the phone the desire is you could drive the phone anywhere that has phone reception and still see where it's going via the GPS and the video camera. The idea of controlling a remote control car and other devices (even real cars) from an iPhone has already been done - in fact one of these is about to be sold commercially (see here) - the iKon RC which uses an Apple-approved dongle that plugs into the bottom of the iPhone and uses infra-red, and they are hoping to make a version controlled via Android phones too. An even more recent example is the Rover Spy Tank. Generally speaking, however, using a smart phone to emit the blue tooth, USB, or WiFi and/or infrared signals to control a car typically requires hacking or jailbreaking the phone, so that's not desirable for the average smart phone user. I still haven't done a tonne of research, but it doesn't appear easy to just design an application for either an iPhone or Android which can emit signals to the USB.... it just doesn't work that way. So if you can't use software to access USB output how else could this work? One clever method people have investigated is using the 3.5mm headphone jacks from iPhones and various Android phones to send control signals - using certain frequencies or patterns which can be interpreted into a digital message. This is a nice idea, but in SCRAPi's case we want the use of sound output to talk to people via the phone's speakers, which are tiny but surprising loud these days. Using sound to transmit left/right, forward/back signals on top of a person speaking would likely create too much interference. That's when I had my own idea - why not use part of the LCD display to transmit movement signals?!

Fig 6: Using video output and photoreceptors to control movement
CG illustration showing how the controlling device (left) could contain basic controls to move the unit and show the perspective of the car... while the car itself could relay the control signals to the car itself by showing colored squares to inexpensive photoreceptors at the bottom of the LCD screen.

2.3.1 Controlling movement using photoreceptors over the LCD display:

The idea here would be that when you sell the unit, the package includes instructions on how to install a free application called "SCRAPi controller" to your phone. Once started, you can select either: (a) "use this phone on top of SCRAPi" or (b) "use this phone to control SCRAPi". On the phone used to control/drive SCRAPi most of the screen shows what SCRAPi is looking at, but the bottom of the phone has four or so buttons for left/right, forward/back (Fig 6, left). On the phone mounted on SCRAPi (Fig 6, right), most of the screen shows the person, but the very bottom is a black strip where colored boxes illuminate when the user pushes the buttons. Small and inexpensive photoresistors near the base of the phone (where it sits) could then pick up these signals, and be used to drive the car. The advantage of using the LCD display is that most smart devices - iPhone/iPad, HTC, Motorola, Samsung, Nokia phones etc could be quite easily calibrated to project the little colored boxes in the right places and/or adjust the photoreceptors in the right positions (Fig 7). If SCRAPi has extra moving parts, I'm sure these could be added as extra little colored squares, and on the controller side it might be intuitive if the user could swipe his finger on the video display area to pitch and rotate the head (up/down, left/right).

Fig 7: Using different sized smart devices (phones vs. tablets) in the same unit
CG illustration showing the size difference between an iPhone (left) and iPad (right). Although Apple products were used here, most other smart phones and tablets will be of similar size and should fit the unit just as well. Note that a mounting tray could easily be designed such that any of these devices could fit in/rest in the same unit, with the main difference being that a tablet allows the face to be seen from a greater distance.

2.3.2 An advanced "SCRAPi avatar controller" program:

In addition to installing a control application onto your mobile phone(s) devices, what would make SCRAPi even more fun is if you could go to the SCRAPi website and download a more full featured "SCRAPi avatar controller" program. I imagine this program would work best installed on a computer/laptop where the comparatively large screen (large compared to phone screens) could be used to show you not only the "SCRAPi cam" but also the location of SCRAPi on Google maps. Suddenly it would become possible to drive SCRAPi down to the local corner store - so long as you have video reception! To make the SCRAPi even more fun for kids, the controller program would also have an area of other control widgets/options, featuring a variety of amusing noise and/or images / overlays you could project. Some of the most successful mobile phone apps are those which do simple distortion effects on the image.... and another application which made a ridiculous sum of money did nothing but emit farting noises. If these things are easy to add (and I'm sure they would be), then it would be smart to add these features, and kids would certainly get a lot of amusement by driving up behind someone and farting. Taking it even further you could project a little 3D avatar face - although to my mind the whole appeal of this is that it is your own face you are driving around and reacting in real time.

2.3.3 Extra SCRAPi modes for independent exploration:

Up until this section we've assumed there is always a person controlling SCRAPi and projecting their own face as the avatar. To make SCRAPi ever more versatility however, it would be nice if he wasn't completely dependent on people to move and interact with the world around him. For this reason, I think SCRAPi should have several modes:

1. Control mode - the default mode, whereby SCRAPi picks up signals from a calling device and this drives him.
2. Explore mode - where SCRAPi randomly drives around the room, using sensors and/or the phone's video feed to help minimize collisions and drive more effectively through doorways and halls. One ambitious idea is that he could use the camera and direction to paint a picture and/or map out of the room/around him and *possibly* even set way-points that he could find his way back to.
3. Follow mode - where SCRAPi attempts to follows and drive anything that moves (human or animal) or (if no movement is detected via the camera) which makes noise.
4. Reaction mode - where SCRAPi reacts with simulated emotions such as happiness, playfulness, sadness, fear and fright depending on what he hears and sees. As a few examples, SCRAPi could display a frightened face and roll back when he hears a loud noise (light a clap), SCRAPi could laugh and say "I love you too" if picked up (sensed by the gyroscope), or cry out "why won't anyone play with me" when left alone in a quiet, still room.
5. Dance mode - where SCRAPi wiggles his arms to music and displays a dancing CG avatar (or multiple avatars) on the LCD screen. The music could be either music heard through the speaker or played on the phone itself via a separate application.
6. Parrot mode - where SCRAPi listens to people and will randomly repeat things, possibly in an amusing high-pitch... not unlike the highly successful "Talking Carl" phone application.
7. Comedy mode - where SCRAPi has a long list of jokes (especially robot-related jokes) stored as text, and he turns to face people and delivers these jokes. By storing jokes as text, and letting the phone generate a computer voice you could have a potentially huge database of jokes and dialog for SCRAPi - many days worth in fact.
8. Radom mode - where SCARPY randomly switches between all the modes above.
9. Robot-robot interaction mode - a mode only available if you have multiple SCRAPi devices and they detect each other via a WiFi or blue tooth connection. Once connected, the robots could find each other and coordinate scripted interactions - slightly similar to the InterAction Wall-E and Eve toys (see video) but using more dialog, since this could be generated via the phone.
10. Programmable mode - where the user could line up a series of instructions for SCRAPi, telling him where to drive and perhaps also what to say, or what mode to switch to - and how long to spend doing each action. The idea of using video and compass to automatically find a way-point might be little ambitious (even via the fastest and smartest phone), but certainly the user could program (or rather "queue") direction and turns. A fun game for kids would be to setup an obstacle course of little domino-like colored cardboard cards (possibly which come in the pack or cut out cereal box-style) and program SCRAPi to drive from one side to the other without knocking over any of the cards.

By implementing some or all of these mode, SCRAPi suddenly become more than robot avatar extension of the driver - when not controlled by another phone or computer he could have his own little personality/personalities and modes to provide hours more entertainment. I should point out that several of the modes above (follow, reaction and dance) are very similar to the U-Command Wall-E toy, but in the case of SCRAPi, you can do much, much more than any custom build sensor vehicle because you'd have access to the impressive functionality and computational power of the smart phone itself. By having a programmable mode and putting the specs of SCRAPi online I believe a lot of electronically minded people will become interested in using SCRAPi as a "hacking platform" and could easily release their own phone applications capable of driving SCRAPi and adding their own extra functionality. Short of installing a huge bunch of expensive sensors, additional electronics and extra video feed into SCRAPi, implementation of the modes above - particularly the advanced modes like storing and/or adding hours of dialog as text - requires the use of "SCRAPi avatar controller" software and thus whoever buys SCRAPi would need at least one compatible smart phone. Most people these days own, or know someone who owns a smart phone, however it would still be sensible if SCRAPi can do at least a couple of things without needing a smart device mounted on top and/or another device controlling him. Since it won't cost much extra to manufacture, it would be wise to still include a normal RC controller inside the pack.... allowing the person to drive the unit around if he doesn't have a smart phone and/or if the movement signals from the phone stop working for whatever reason. In some cases it might be fun for kids to share by letting one kid drive (using the RC controller) and another kid to be the face. With ideas like this, and a nice number of modes to choose from, SCRAPi could become a huge family pleaser.

2.4. Turning it into a Social Project and YouTube video

Above I've talked a bit about how you could build SCRAPi and how he could be packaged and sold commercially. What interests me even more in this idea, however, is the human aspect. How will people - especially strangers - react to a little "robot avatar" when it drives up to them and ask them a question. I've run preliminary tests on friends, and that was great fun - but the real challenge is strangers! In this little subsection I talk about how you could turn this into an interesting university project. In my case, I work in computer science and cell biology (science with medical applications) so I don't have much time to dabble in "human interaction with computers", but this would definitely be a fun project to do on the side with the right people. These studies could also be turned into an awesome YouTube video to get people interested in SCRAPi and the idea of social robots. My plan was to run these tests in the social setting of my university campus at UCSD near the food court where there are some nice pathways and a lot of students hangout.... and a few of the many "social tests" to run are:

1. SCRAPi drives around the main strip - to see how many people stare.
2. SCRAPi waits for people to walk nearby and waves hello (without speaking) - to see if they wave back.
3. SCRAPi drives up to under-grads and asks directions.
4. SCRAPi sings to people - to see if they laugh.
5. SCRAPi is put on his side and calls out for help - to see if people put him upright.
6. SCRAPi drives up to students and asks for money - to see how many people might actually give money..... potentially enough to pay for a new version, or at least enough for the bus.
7. SCRAPi asks people for their life story - human interest and to see if they do.
8. SCRAPi does a survey and asks people if they think he's attractive - comic effect.
9. SCRAPi asks for a hug - might be funny, although you wouldn't describe SCRAPi as very huggable.
10. SCRAPi asks for advice about love/relationships - more comic effect especially if he claims to be in love with a small dust bin.
11. SCRAPi asks girls for their numbers - a challenge to see how many phone numbers he can get in just 5 minutes using a little back compartment with pen and paper.
12. SCRAPi chases a dog - hopefully not resulting in the end of SCRAPi and (more importantly) my phone!

In all of these tests you would have at least one student filming, and another student discretely driving SCRAPi and the person actually doing the voice/video of SCRAPi would obviously need to be hidden. I feel such a video would not only measure people's reaction, but, with the right personality controlling SCRAPi would make a hilarious video. To make it even funnier you'd build a few of these (even if it's the $15 version) and have the robots talk to each other.... or drive around like a little robot gang. I imagine such a video would have a pretty high chance of going viral and, if it does, all of the viewers would be referred to a website/mailing about SCRAPi to help build a fan base of people who might like to buy or build their own social robot. If SCRAPi got famous enough, he could even appear on TV or a stand-up comedy gig. Obviously you could have a lot of fun with this, but taking it one step further.... if a reliable unit was built, I think you could easily ask for sponsorship money to drive SCRAPi across the country. Since SCRAPi would be slow, what's more likely is that SCRAPi would have to hitch-hike - to see if people with the goodness of their hearts can take him all the way from west coast to east coast - say from UCSD to New York or Washington - or if he'll simply get stolen or stuck! In many cases SCRAPi would be out of phone reception, so might need to rely on people to watch a pre-recorded message or recharge him... but this type of little quest could certainly get some good feature on TV - especially if the little scamp actually makes it, and especially if you watch his progress live on a website using a Video Feed plus Google maps! In my mind a long distance journey and these kind of "human kindness" test is what the project is all about. That and engaging the public!

2.5.1 Seeking TV Publicity

Depending on the success of any YouTube videos, it would also be very intelligent to seek publicity in the form of television. Already I've noticed an idea similar to SCRAPi has appeared on the "Big Bang Theory" in season 2, episode 4 where Sheldon, the geekiest of the characters, creates a full height robot wearing a shirt and projecting his face from his room. Sitcom shows like Big Bang Theory and others could have a field day if you even just sent them a device like this as a freebie. For an ever bigger audience, I'm sure there are lots of quirky TV talk show hosts (Conan, Jimmy Kimmel, and many other late night show host) which would love to drive around in another city and chat to people via a cute little version of themselves - it's got good entertainment and good publicity written all over it!

2.5. Turning it into a University Sponsored Interaction Area

One final method to make the SCRAPi project popular is to have a live area on the website where people can setup an account agree not to use any bad language or rude signs and then, on a first-come-first served basis, they could play 10 minutes each on a small number of SCRAPi devices in an enclosed area. Such systems where users can play with real gadgets via a video field have become very popular in certain universities and inevitably they get far more people wanting to play with these toys than they can cater for at once. In the case of SCRAPi a dozen or so robots could be put in an enclosed glass area somewhere like the front of the CalIt2 building at UCSD where live webcams could watch the robots interact and talk to each other, and also wave to the people outside their glass enclosure. Obviously the units themselves would have to have low enough power not to damage themselves or the glass, as users are always going to want to test their luck by seeing if they can be destructive. Via the web interface it would also be interesting to post up challenges, such as moving objects or trying to get someone to wave. What would be most interesting however, is that if some of the physical challenges, such as getting a ball out of the corner (assuming the corners are square and maybe this is a bad idea and corners should be round), require more than one SCRAPi unit and thus people must talk to each other via their robot avatars to help problem solve, strategize and coordinate their efforts. Better yet, it would be great to have players log their success or even use the cameras to record success rates so that you might publish results-and-finding on the website and/or to a journal.

Fig 8: Proposed display cabinet where public can login and use SCRAPi
CG illustration how members of the public could interact with each other (A) inside a display cabinet, and possibly even work together to help solve puzzles by moving around various objects in the box. Players could also try and interact with people outside the box and website visitors not playing could use live webcams to watch the interaction between the SCRAPi units (B). I propose a great place to put such a cabinet to attract attention but still be secure is the southern arm of the CalIt2 building at UCSD where the cabinet could be placed up against the glass wall so both people inside and outside the building walking back from coffee can see what's going on (C).

2.5.1 Recharging the Device via a Magnetic Docking Station

The only real problem I foresee with the display cabinet above is that you'd need a person to maintain it by recharging the devices when they run out of juice! For this reason I think battery life is a major consideration when choosing what smart phones to mount, but you'd also want to put a large battery pack inside the device itself which is cable of keeping both the phone and the moving parts running many hours. That said, at the end of each day I imagine a recharge will be needed, so the top of the display cabinet would unlock and include a "recharge station" where each unit can be plugged in. Getting a little fancier, it would be very cool if each of the little SCRAPi units above were designed with a magnetic plug - possibly sticking out the front or one of the hands... and when this got closed to the charging station it would attract and automatically lock into the charger. Designing such a recharge mechanism would be a great challenge in itself and very useful for a university to investigate since such a mechanism could be patented and used in any future smart-phone driven security devices - devices which could be remotely driven around someone's home when their away, but need to be able to return to some kind of recharge base in order for the phone to remain on many days. In the context of the display cabinet above, it means that instead of needing to open the cabinet every night, the web service would tell players to drive the vehicles back to the recharge station at a certain time and/or when the batteries neared depletion. Five minutes later the web service would shut off all users and a volunteer (read: PhD student) could drive any undocked SCRAPi units back to their dock. A few hours later, after charging had finished, the website would reopen, and people who logged in would be able to back the device away from the magnetic docking station and recommence interaction.

Section 3: Uniqueness

As mentioned before, there are a lot of people who have used iPhone and other smart phones to control vehicles, but I was quite surprised that, looking over the web, I couldn't really find any examples of people mounting their devices onto the front of a remote control device with a view of turning it into a social entity - a "little avatar" or "miniature version of themselves". Furthermore, I'd like to think (but haven't yet checked) that a few of my other ideas, such as using the LCD screen to control the car's drive mechanisms are original too. Of course, there are a tonne of robots on the market, and many try to simulate a person's behavior and/or emotions, but this is unique because you're talking to a real person and, unlike a normal video chat session, that person is able to move around the room like a real person might... only much smaller!

Section 4: Target Market

The obvious target market for this is people who like gadgets and like to tinker with electronics, and if you were marketing to these people exclusively you'd probably want to sell it as a kit that requires assembly. However, I believe the target market for this is much, much bigger. Almost everyone (first world) these days has some kind of smart phone, and so when these people see a YouTube video or even news piece about "social robots" I guarantee they'll all briefly fancy the idea of playing with one. Some people will be interested in the idea of having a virtual presence, and others will be more entertained by the idea of sneaking around the house to spy on and/or creep up on their friends and family. The marketing and price of these "social robots" is key. If it's cheap enough and sold in supermarkets I'm sure a lot of people will think "instead of getting just a RC car, why not get an RC car that doubles as a social robot.... AND it can be driven around via video feed without the price tag of buying any toy with a video camera". Many kids already have RC cars, but SCRAPi offers much more in terms of interaction and the fun of using the vehicle to explore a room you're not in. Another target market may surprise you - and that's for security minded people and pet owners. Often when people are away on holidays they get paranoid and wish they could take a peek at their house. Furthermore, there are a lot of people with pets who, even when at work, would love the opportunity to just check up on their pets. Short of installing a video camera in every room, you could see a version of this unit that works with even a generic cheap China smart phone - a fraction of the cost, but can still hook in with WiFi. If a house has WiFi, and you work out how to dock the unit into power or long term battery, people can use the "SCRAPi avatar controller" on their own phones to dial home, and then drive their SCRAPi around to check the house is in good order, the dog is okay, and the nanny is not misbehaving. It has got a real big-brotherness to it, but hey - it's their house and I'm actually really curious to see if a family pet would recognize a person's face on the screen...... maybe even feel like they are in the same room. Pet lovers are a huge market. One final target market shows that SCRAPi can be more than just a toy. Throughout America are thousands of people in hospitals or even in their own homes who are rendered immobile through poor health or injury. In some tragic cases, these people's lack of mobility may be permanent, and suddenly they become unable to interact with their families the way they used to. The simple ability to walk around the house and talk to family members whenever they want has been taken away from them. SCRAPi could bring a life back to these people by allowing them to once again roam their own house and interact with family, keep a watchful eye on their children, play and perhaps even see a toddler's first steps. In this situation SCRAPi could be setup with a magnetic charge station on the ground, and an application to receive calls permanently on. By calling with another mobile device, the immobilized patient could then drive through the house (assuming doors are not shut) and talk to family members via their tiny avatar. This could even include sitting on the dinner table and asking questions about everyone's day. This represents more than just a phone call - this allows the patient to move and interact, in some small capacity, with their loved ones. If they became popular, such units would probably be modified to be much larger, less "toy like". Such adjustments could include a head which could adjust up to eye level, and possibly even small arms capable of opening doors. Such a mobile unit wouldn't realistically be capable of cooking dinner (not without incredibly fancy/expensive robotics), but would at least make the patient feel more connected with their family, and like they still have a presence outside their own bed. When they see their power run down the patient would then drive back to the magnetic driving station... perhaps even using the rear facing camera to reverse back into the magnetic dock, and disconnect the call, knowing that the unit is charging up and will remain ready to (automatically) receive the next call.

Section 5: Budget

I think the budget would be quite reasonable. The very first prototype (SCRAPi v0.01) cost a grand total of $15, but for a better version with better range I estimate $100-200. What may be more expensive is actually developing the software and/or working out a way to make the idea of "social robots" something people will want to buy for their kids. Also, the "Social Cellphone Robot Avatar Project" itself would take a small team of people to put together, including some reasonable video cameras, and so for this you'd probably need a few thousand and would want to approach a phone manufacturer for sponsorship money - and in return you would use their phone of choice.

Section 6: Potential Challenges

Aside from the technical challenges of getting SCRAPi up and running the biggest challenges SCRAPi might face include:

• (a) losing video phone reception (which is only a problem if your phone is miles away)
• (b) getting stepped on accidentally
• (d) getting stolen!

Since smart phones are quite valuable there's definitely a possibly someone can simply steal the phone out of its mount or even just take the whole unit and run. In most cases I think a person would be stupid doing this because: (a) the person controlling SCRAPi is (most likely) hiding just around the corner and (b) when a smart phone is stolen the plan can be cancelled quickly and with the right security app installed the phone can be locked down (rending it useless) and often tracked and/or take photos of the perpetrator. Still, there are definitely some stupid people around, and who knows when one may get the impulse to snatch and run. Now in most cases social robots would be used around family and friends, so no real danger of theft, but if deployed in a public area it stands to reason that you have someone nearby keeping an eye on him - just in case! Losing phone reception is something we simply can't help and it's only a big problem if you've driven SCRAPi a mile down the road to ask the neighbors for milk. For people who live in remote areas, SCRAPi can hopefully be designed to work via WiFi (think Skype) or Bluetooth rather than needing Cellphone reception (like FaceTime). And so, really the biggest concern above is the idea that mum could accidentally step on your expensive iPhone. Most smart phones are pretty robust, but not smart enough to survive mum's killer heels, which is why the unit would come with a flag to draw attention. Again however - common sense prevails, so if you drive up behind grandpa in stealth mode you're asking for trouble and really don't deserve to own the iPhone daddy bought you. One final challenge worth mentioning comes to mind in the form of liability issues - being able to use the device to spy on people and since it is a device that's almost ground level (less than a foot above the ground) with a camera angled upwards it could easily be used to look up women's skirts. I'm no legal expert, but I think a disclaimer could get around that, and it obviously hasn't been an issue for the Rover Spy Tank" which is not only on the market, but actually advertises itself as a spy device!

Section 7: Branding

I haven't thought of too many clever names, I imagine many people would be tempted to call it an iAvatar or something starting with i, but I see this unit as working with pretty much any smart phone and I really like the phrase "social robot". Sure, underneath the "robot arms" it's a remote controlled vehicle, but to make it really stand out I think a funny name like "SCRAP" (Social Cellphone, Robot Avatar Project), "SCRAPi" (Social Cellphone, Robot Avatar for Personal Interaction) or "SCRAPPIE" (Social Cellphone, Robot Avatar Project for Personal Interaction and Exploration) would work well and I've designed a little logo below (Fig 9). During live social experiments, this logo is something I would suggest mounting on the front of the car plus the flag on top so people are likely to read it, see the thing as being "official" and thus be more friendly and open. Another possible name I like (perhaps even better) is "SCAMP" or "SCAMPi" (Social Cellphone Avatar Machine for Personal Interaction). For commercialization, it could probably be any name (it doesn't have to be an acronym) and I'm sure better names exist. As a general name to describe any such machines, I think "social robots" or "phone avatars" are both nice short terms. Social robots don't necessarily need to use cell phones - they can have tablets or even come with their own LCD screen and video camera to eliminate the need for the user to provide his own.

Fig 9: Suggested logo for the project.
This logo would feature on the website, YouTube video and also be put on the front of the car so people would see it as being more "official" (not just a fun project) and thus I would assume they'd be more likely to all stop and interact with our adorable little robot.

Section 8: Work to Date

So far it's just me designing the crude version you see above.... I've been meaning to field test it on campus with a few of my lunch friends, but the range and power of this unit is so bad I haven't gotten around to it. I haven't made any website yet - the only other thing I've done is write this page up.... it's a good test to see if the format I propose helps people develop and further think about how their ideas will work... and I have to say that - even though I'm rushing to finish this - the format works great. Already while typing this I've put in a lot of extra thought and came up with the idea of using the bottom of the LCD screen to control the vehicle.

Most recently: I submitted this idea to Quiry (www.quirky.com) on the 20th Jan 2012.

Section 9: Taking it to the Next Level

To take this further I need time and the help of a few fellow UCSD employees. The project has only just started, so what I really want is to film SCRAPi in the field, interacting with students and then from this I can make a YouTube video and then get a few friends to help me build a better version of SCAPY. Some of the appeal is that he will look a bit silly, but certainly would be nice to build something with a plastic frame (not just cardboard) and that would lead to another YouTube video and website.

Section 10: Statement of Intent

Since I'm already flat out with other projects, I think at that strange I'd be happy to hand this over to a different group - maybe even the arts department or some engineering students for them to take it further and hopefully generate some good publicity for the university. For me, it's always nice to have a non-work project on the side, so I'd like to always stay involved in some capacity and acknowledged for having these ideas. If it happened to make some money, I think I'd definitely appreciate I got a small cut, but I don't hold out huge expectations - I'd be just happy to see someone turn this into something fun. Probably I'll send this to a few friends and then find a few people high up in CalIt2 who might value it's potential - people who might have or know students in need of a new project.

Conclusion

Plenty of interesting applications have been written for smart phones which use their GPS and camera - but here I've proposed a project and possibly marketable product where smart devices can be turned into little avatars which can drive around and talk to people. The words cell phone and "mobile device" have become almost synonymous, but really these phones are mobile because we carry them - here the phone could drive itself. More than just a hilarious YouTube video, this could capture people's imagination about how we interact with each other, and how people react to something mobile which represents a real person, yet isn't a real person. Just as interesting is how the person controlling the device may behave - whether or not they lose their inhibitions as they aren't in the same room! Although it's very early days for this idea, the potential for this device goes beyond just talking to people - can spawn into more and more sophisticated devices which allow mobile and virtual presence - allowing you to have a physical presence in two places at once. This idea not only has huge potential not just as a novelty device, but also in home security. Of all my ideas I'm pretty sure this is the most likely to happen in the near future - if not by us then almost certainly by some other group. Smart phone applications and tablets are still a hot topic and every day we are seeing new devices that interact with iPhones... it's a matter of time before this idea becomes real - I just hope it is people from our university that can take the credit!

Links - Similar Products

Below are a list of products which at *most* similar to this idea. Some of these devices are pretty cool, but you'll notice that none of them represent avatars.

• Rover Spy Tank .......... (controlled from iPad)
• Rovio ......... (wifi enabled mobile webcam from WowWee with great mecanum wheels and auto docking station with beacon lights)
• iKon RC ...................... (similar to above, but controlled by iPhone)
• U-Command Wall-E ... (fun robot with multiple functions)
• Spykee ....................... (an attractive and functional WiFi driven security camera robot)
• Roomba ..................... (very cool robotic vacuum cleaner, and popular "hacking platform")
• SmartPet - toy iPhone robot pet ....... (my friend Raj found this, this robot dog with a comical iPhone face came out June 2012, won an award in Toyko and is getting closer to my idea)
• AVA 500 ..................... (teleprescence device which is human size, screen on wheels, done by CISCO and iRobot)
• Double ............. (beautiful product for iPad) great telepresence robot from Y Combinator using segway technology, just two wheels and a long pole - currently $2000, but hopefully will come down.
• JIBO - family robot...... (intriguing robot which announces emails and is supposed to act like one of the family).
• RoboMe - WowWee..... (very versatile telepresence robot for $20-120 depending where you buy it.... only works with iPhone, but has some capability even without any phone - see these demo videos part1 and part2 - videos on official website are awful)

Update: July 2012 and Social CellPhone Robots Almost Here - Some Botiful News

• Botiful ........ (a brand new update on July 2012, released the first proper social cell phone avatar)

As it turns out, French inventor/creator Claire Delaunay from Palo Alto thought up this idea and commenced work probably years before this page! I pre-ordered it on kickstarter immediately to show my support, and I'm sure if you search "Botiful" you can order your own soon too! I also contacted Claire and she commented how similar the designs are, although I still think her's is far cooler - and it is, for me, inspiring to see that she was able to take it so far. As it happens I moved to Mountain View to joined Google and only in 2014 discovered Claire had started working at Google X, so we met up for a fantastic chat. Very wonderful lady. :)

Update: Feb 2013 Social Robots Are Available

My Botiful arrived and I noticed some new ones have too:

• Romo - plays more with the robot idea, with a cute face and self navigation.
• Helios - another telepresence device for mobile phones and uses my idea of using photoreceptors.

Acknowledgements: Raj Singh from UCSD for contributing ideas (in fact he had very similar idea to this one, but more on the security side of things) and interest. The wonderful Heather Klat for her support and starring in the main pictures. :) Also thanks to my roommate Felix Grobler and co-worker Rick Guily for their input too.