 Hei allan. Yna'r taw wahanol yma yw Russell Cooper, rwy'n cael ei wneud yn rhan o'r cynwt o'r ffordd yng ngyfnog, yr adferfodd yng ngyfnod yng ngyfnog yng Nghymru. Rhaid i'n credu. Felly, y ffordd 4.5 yr ysgol yng Nghymru, rydyn ni wedi bod yn ysgol y tîm hyn o'r bach ychynod yw Brystle Hacksbas, bod yna'r ffordd o'r gwaith ymddangos. Felly y peit eich eich awr yn rhywbeth gyda'r cwmwych, yn ddechrau cyfnodol i'r golygu ar gyfrofit a gyda'r ffordd? Felly rydyn ni'n fwyoldo gyda'r bobl. Mae gael yr oed... Mae gael yr oed hyn oedd y gydag, ac mae'r gael yma ar gyflugio'r ddisblwys, mae'r yma ar gyflugio ar gyflugio ar gyflugio ar gyflugio ar gyflugio. a ddyneth sy'n ôl yng Nghymru. Rydyn ni'n gael iawn i'w prosesio yma i'w ddweud o 400lb y unig, ac rwy'n cael ei bwysig y peth oedd yn y ddweud o'r unrhyw ymlaen. Yn cael y cyflwyno gyda'n ei gwybod, sy'n cael ei oed o ran reilogau multilinell o'r cyd-rhyw o'r market cyflwyno o beth o'r newydd oed. Mae ydych chi'n meddwl yw'r llyngen, mae'r llyfr o'r llyfr yn rhan o'r llyfr o'r llyfr, nhw'n maes ei gafod o ymryd â'r beth, dwi'n gwneud y llai yn brym hwnnw. Felly, rydyn ni gynnwys ar hyn yn ach chi'n eistedd i'w bryrach. Hwnni, a dyna'r gweld, rydyn ni'n brym hwnnw, dyna'r gweld yn gallu mwyfyr ddymiadol, wrth gael yr ysgrifennu ar yr ymlawn, rydyn ni'n ddechrau bod ymleidio gwrth ymlawn i bardzo ei mewn gwir. ac ydych yn oed yn y bwysig ychydig yma. A dyna'n meddwl ystod o'r cyfnodd angen a sut ymgyrch iawn y bwysig mae'r bwysig yn y rhai. Ac rwy'n ganddo nhw'n golygu'n gweithio'r bwysig yn golygu'n cyllidau. Mae'n ffordd i'r clywed. Mae'r bwysig yn y bwysig? Rwy'n gallu'n ddod. Yr hyn yn gweithio'r bwysig. Hello, I'm Ed Rogers, I'm the director of Bristol Braille Technology and we are working on the Canute. The Canute is a radical new Braille e-reading device. It's like a kindle for blind people. It makes digital Braille far more affordable than anything that can exist on the market by having multi-line Braille, which means that you can see a whole page of Braille and it will cost hundreds of pounds rather than thousands of pounds which the existing models currently cost. Now we want to take the prototypes that we've already made over the last couple of years. We're on the Mark 9 now. We want to take this prototype and we want to make a batch of them and trial them in schools around the UK, in Ireland, maybe in America. And we want to get the feedback from these schools about how to make the Canute into the perfect device for learning Braille and for literacy in general. We want the Canute to be a tool which can help reverse the decline in Braille usage and therefore increase literacy around the world. So we've been working with Braille. The Braille is a community organisation established in 2014 designed to give a voice to blind people so that they can be involved in the creation of their own technology, especially around digital Braille. Having an affordable multi-line Braille display would be transformative, I think, for Braille. A multi-line Braille would certainly make studying science and technology subjects more feasible, I think, for blind students. Take computer programming for example, been able to review code across multiple lines of Braille. That would be a very powerful use case for multi-line Braille. Braille for children in school, they could do maths on it. Well, they could get a hold of material, be loaded onto it electronically so that it would be readily available in Braille. It's very disempowering to be in a room and to be told well, all the information you want is on the screen. But that doesn't help a blind person to feel that they're part of the meeting or part of the experience. I use Braille every day of my life. Well, hopefully it would mean that I'd have access to more books. And I can think of practical uses for it, like for instance I read in church once a month and my Braille missile at the moment is 15 volumes and searching for things in it is a nightmare. So it would be much quicker if I could load the missile on to something like Canute. Same with recipes, I could take it to the kitchen and look up things instead of having to trade something downstairs to my computer to get the Braille stuff I want. Well, for the last two years the developers of Canute Bristol Braille technology have worked closely with the Braillists who are a community of, I think, around about 200 Braille enthusiasts, some of us are Braille readers, some of us teach Braille, some of us are parents, children who read Braille. So Bristol Braille technology have paid very close attention to the needs and the views of that community in the development of Canute. I think very early on in the development of the Canute blind people were involved. I certainly remember Ed contacting me and saying he needed to involve vision impaired people in testing and giving feedback and generally being involved at every stage of the development. We told them what we wanted in terms of the number of lines, the length of the characters and they've gone away and tried to build this. It's very gratifying to see what's taken place, what's actually come about. For me it's a great encouragement and I think I feel it's really valuable. I feel like I have a sense of ownership of this project and a commitment to see it through to success. I certainly feel glad that I've been a part of it. I've been able to show off to some of my friends who don't live in Bristol who have said, oh, I wish I was a part of that. I'd like to be a part of that. It would be a good thing for me as well. It's a bit longer than I remembered but that's why we're doing the project. To summarise the goals, it's to increase literacy amongst blind people because there's an overall decline in the use of Braille which sort of cuts off access to books and other things. We're also hoping that through involving them with the development of it we could help bring more of the blind community into the hackspaces. As a point of interest, one of the other goals is when the project is at a more finalised stage is to release the knowledge in a CERN-based open source hardware licence. Moving on to some more technical things. The Braille is on a 2.5mm dot pitch and then there are six pins or six bumps per cell. There's another grade of Braille that has eight pins but we're working on six because it's mechanically easier to deal with. The page we're working on is an American standard of 40 characters per line often referred to in Braille as cells and then 13 lines per page. The interesting thing there is you have 3,840 individual dots to address and control for each page refresh. The equivalent of turning over the page in your reading book or diary would involve the control of 3,840 dots. The thing that crosses everybody's mind especially if they have some electrical or technological background is it can be done with magnets or solenoids and I'm sure that any one of you in the audience would be thinking much the same. However, magnets just don't work because by the time you've built your electromagnets down to a 2.5mm pitch so as they're sitting in rows 2.5mm apart you'll get magnetic crosstalk. The core of the solenoid being the bit that is actuated by the magnetic force becomes so thin that the resultant force on it is nowhere near enough to take the force of somebody's fingertips going over the top. So really the idea of using solenoids even though it's the most obvious is difficult. The other thing is cost and things cost and lots of things cost. So if you use an individual actuator per dot not only is there a lot of them but however inexpensive they are having almost 4,000 of them is going to be prohibitively expensive to get something into the front room of your average guy who either needs it or who is interested in learning the medium of Braille. So our project has gone down the path saying well can't we be more efficient with the use of actuators and can't we address more than one dot or one bump with one actuator and we've come down to what we think is a minimum which is using a matrix of actuators one up the side of the machine and one across the top to make a matrix of 16 row actuators and 40 column actuators making a grand total of 56 actuators that would control individually the 384 bumps that would make a page of Braille and we've done that by implementing and developing a cunning arrangement of gears and levers and to coin a phrase that often ban round is using the sort of tried and tested Victorian horology clockwork mechanisms with the modern control of microprocessors and so far we've had promising success with it. The other task is to make the machine repeatable one of the ideologies of the whole project was to not only open up literacy Braille to electronic devices in the developed world but also make it available for people in the developing world where I believe there's a higher instance of visual impairment and for this we'd have to be able to make each machine with accessible tooling such as laser cutters that you now find in every hackspace, CNC mills which are almost an everyday thing and 3D printing but through our path of producing the machine and the prototypes we discovered various materials limitations such as nothing is flat which might sound obvious and it might not if you've never needed anything to be flat we've come across problems of laser cutters distorting thin material if you have a very thin sheet and you put lots of holes in it it then warps in the middle because it's distorted being melted on the edges you can make very complicated acrylic shapes you can see some in some photographs in a moment however acrylic is fragile and it snaps and if you get the wrong kind of solvent on it it self destructs in front of you the other obvious material is acetyl it's kind of a wonder milling machine millable material however acetyl bends it's soft material so if you make a precise machine out of acetyl it will no longer be precise in a day or two because you put it down irregularly and it's slowly creep crept and distorted out of shape the other engineering problems we've had are problems of miniaturisation in that we'll test an idea at five times scale so you can pick it up and handle it but then when you miniaturise it it gets smaller yet the accuracy of your machine stays the same so the actual tolerance and scale are related so as it gets smaller the inaccuracy, for instance your laser cutter becomes more and more significant and we've come across various problems with that as well the other approach we've made is whenever we've tested a new idea we've tested it on one braille cell with six pins so we've got a lot of support around a very small area but then when you increase that area to 40 cells which is about a foot, you know, 30 millimetres then all of a sudden the force of resting a finger on the middle of a bar that's only three millimetres thick for instance, is significant and not only does it distort under the pressure of reading but then any machine moving on this support rail underneath is applying force along the whole length and then it bends and we've ended up having to use in the prototypes ground carbide rods which are prohibitively expensive for the final machine and that presents another engineering problem we're going to have to get over before we can go into mass production and the phrase at the bottom we've not been able to run test of miniaturisation as we don't have the resources is a combined problem of working on limited budget with the turnaround cost of having a lot of things rapidly prototyped especially now as the machine has become so sensitive to tolerance that we've had to have parts for it made on some of the top 3D printers in the country and we just don't have the resources to say we'll have one of those, then we'll have one slightly differently and then slightly differently again, we have to make an entire machine and almost guess that our intuitive calculations on why it went wrong the previous time were right so to highlight some of the prototypes we've been through is this is one of the earlier ones some of you might recognise the simple stepper motors on the end there's commonly available hacketype ebay purchase, Chinese motors and then the whole machine at that time was made out of acrylic which was fairly good, it did the job but if it suffered any kind of shock or blow it would break and also as we strive to get the machine more repeatable and more accurate we noticed that the laser cut edges were nowhere near as accurate as you would think and that was a bit of an opener to us all because we'd all made many things on laser cutters personal things, projects, laser cut stuff in our everyday life but then when you start actually measuring it the machines almost give you a different cut every day you go to them despite nobody using it in between so it's things like that so we'd make one prototype and we'd make a duplicate and they'd behave differently because they'd be cut out slightly differently this particular machine uses brass pins and solenoids so the rows up the side were controlled by stepper motors and then the columns for one of a better word were controlled by a solenoid actuation this wasn't anywhere near the figures I mentioned earlier for a number of actuators but it was still quite low this is another prototype we made, much bigger stepper motors more like the ones you'd find on rep-wrap 3D printers only these are slightly bigger the other problem that this brought with the laser cutter grillic was stepper motors moving steps which is an audible sound then when you have a bunch of them in a machine that sides are made from loosely clipped together or tightly clipped together acrylic sheets you have a sounding board like a musical instrument and the thing would sit there and buzz and click and were like some sort of late 80s sort of rave booming box it was quite comical but nothing you could actually ever give out to anyone and say here this is a braille machine because it would disrupt anyone else in the room that might be with you one of the materials we moved on to was a seatle because it didn't break, we were having breaking problems with some of the smaller parts they were just not durable so we used laser cutter seatle and this was fantastic however it was apart from being slightly more expensive it was bendy and again it was laser cut and the problem as I mentioned before with laser cuts is it's not as accurate as you would think and this is compounded by the fact that the delrin or the seatle melts as it's cut so when you're making gears out of a seatle and interacting parts is there's a greater level of unknown than you would like I put this on because I found it quite a comical photo we had to demonstrate this to some partners or colleagues in America and basically the evening before of the flight we were working with this prototype and someone said how are you going to carry it through customs because you can't possibly put this on the plane a briefcase appeared out of nowhere and it was put in there like some child's comedy spy film briefcase and we thought what with the current state of things because this only flew earlier in the year this one we thought it would produce all manner of security problems taking it to a demo which was very important because that's the sort of forum and the place where we can demonstrate our research and prototypes in order to gain more funding and credibility amongst the communities we're trying to support and really it produced no trouble at all in airport security and getting it on an airplane which is a prize for us so we're not quite sure what they're looking for there this is the first working prototype we were able to give this to people and they had it in their homes for a few days one guy a few weeks and this has Raspberry Pi in it that has a select number of books that we've converted the format on and there's various you see the buttons up and down each side the black buttons for selecting menu options however, unforeseen problems is we developed this one for several months put it in the box, tested it, gave it to people, tested it it was fine because everyone treated it like it was worth a fortune like it was some precious artefact and then one day someone was walking with it and did not much more than a comedy trip taking it to someone for a demo and the pins bounced higher than they bounced before and rested on another internal part of the mechanism and stayed there on that occasion we had completely dismantled the machine reset the pins, put them in with tweezers because I think there's about 500 on this particular machine and once they're back in, assemble it, take it out to the next demo and on the second time it did it we decided that it wasn't worth the three days of disassembly and reassembly so we had to move on to the next prototype and completely eradicate that as a possible failure for each idea we have we've had to test it unfortunately this isn't as clear as it was but we're testing a rack and motor mechanism so instead of solenoids moving to more controllable motors and often when we produce things like this is we set them going we put some kind of feedback on them whether it be optical, with a Raspberry Pi camera or merely contact switches and leave them running for several days to prove to ourselves that the idea is going to be successful and again this is one of many prototypes and mock-ups that we've had to make the other way of testing ideas again this is looking at a rack and pinion idea is you make something larger this looks like it's about five scale even with the best CAD images there's nothing like making something and having it in your hands and poking it, prodding it, flexing it and just seeing how it works when it moves and interacts with the rest of the mechanism these also proved invaluable because we demonstrate every couple of months in various braille groups that we've set up around the country and it's almost empowering to give something like this out to a braille audience and say this is what we're talking about this is what's on the screen and you can feel the gear and the pinion interact and with not so much this one but with other ones this is how we make the dots rise up and down because without that it's very difficult to explain to a visually impaired person what exactly it is you're talking about and this is another idea of gears and interaction and assembly is a separate problem it's all very well designing these mechanisms but the assembly of them is crucial and I've put this one up because it's made of many laminates the whole machine was made I think there were four or was it five acetyl laminate sheets for each braille cell and then we wanted 40 of these in a row and these were held in alignment we actually used the bicycle spoke it was the right diameter and their hardened material with bicycle spokes threaded through long ways so if you needed to maintain or inspect a portion in the middle of the machine you had to slide these spokes out and then the machine probably fell apart it had some sort of crazy origami failure and again you don't think about these things until you've built a machine and it's ready to assemble and then you want to take it apart again and it's unforeseen problems I've mentioned several times the problems we've had with laser cutting some of the things we do are not suited for CNC milling so more recently about springtime this year we started moving to 3D printed parts and it's beyond the scope of any DIY 3D printer because of the tolerance and the accuracy we need we've ended up having the centre of the machine now all 3D printed which is a step away from having it made in a back of a hackspace in a developing country but our development and our learning prove that that was an unachievable target and these particular 3D parts are produced on a machine that at its highest resolution give 16 micron layers and then an overall accuracy across the full area it works on of something like 0.3mm which is about within what we can accept so the project has gone from laser cut parts through its range of materials and then on to 3D parts unfortunately I was going to bring a demo in I was going to bring what we call a POC a proof of concept machine which is a piece of hard board with circuit boards screwed onto it and maybe three or four cells in the middle working but we posted that to a researcher in America last week and the actual demo is with some braille lists I believe it's in one of the schools in Worcester at the moment so I couldn't bring that in and I didn't think it valued to bring in just a few bits so unfortunately I have no physical demo for that so the next stage after using 3D printed parts is then to have them injection moulded which is the only way you can produce things of this accuracy with any kind of repeatability and cost advantage which is something we hope to move into next year so that's about as far as I've got with what I was going to talk about so if anyone wants to ask any specific questions I'll write you a microphone, yeah, cheers What sort of production volumes are needed and how many do you expect to make? At the moment we're making six working prototypes which will be the first ones that will go out to people on a sort of beta test and we've got some of the braille groups will share one we've got schools and colleges and I believe there's at least one going to American colleagues but the overall goal which in the short video I showed you at the beginning is after is thousands and he wants it to be mass produced and we've got the support of the R&IB in this country we've got personal friends as the wrong way of putting it but we're now on talking terms with their research guys they're interested in the project and he has a similar interest from the equivalent of the R&IB in America so when it hits production hopefully it will be mass produced so I hope that answers your question Two questions and then something to mention so questions, what is your target price for this and is the machine going to also have a braille keyboard that is the typical 10? The target price is very difficult to predict but we're aiming for the target price to be about 440 pounds around the 500 pound mark and that was for a full page which is why did you miss the beginning? No, I saw the 16 line by 49 which is one of the major cost benefits is a reduced number of actuators yet more moving parts so the idea is to open up literacy or braille to people who are just interested as well as people who depend on it to talk about the keyboard is we plan on having remote links or radio links, whether it be wifi, Bluetooth, network or whatever to other devices so for instance you can have your text go straight to it which means it looks quite bulky at the moment but hopefully it will be laptop size so as you can read your text for instance on a bus without the guy next to you having to listen in on every occasion if it's through a text-to-speech type thing Have you compared to existing refresher brails where they have like a one or two line display and a nine key keyboard? It's not a full keyboard, it's the braille keyboard It's based on the Perkins machine It is something we talk about but we're so concentrating on making the display that almost a row of buttons on the front is a technical triviality It might be a huge, huge feature and very useful but from a development point of view it's something we can't focus on because we have a fair amount of money but it's still limited for the development and once it's on the machine it's difficult to remove the feature theory it's difficult to take away it's very easy to add it afterwards version two has a keyboard and we'd have to do that Fair enough Also just separately one thing to mention is I'm running crash courses in blind navigation hopefully not too crash ad hoc so I actually do use a cane so if you would like to know how to do so just come find me, it's on the schedule Two questions How much harder would it be to do eight dot braille and do you anticipate getting the power consumption down to the level where you can reasonably run it off batteries At the moment we're focusing on the one standard of braille I can't recall the different grades of braille but at the moment there's not much interest in art ways It's not really not interesting it's at the moment we're focusing on six dot braille Sorry what was your second question? The power consumption yes I mean at the moment the power consumption is quite high it's in the order of 120 watts that's because we've got many stepper motors in it One of the things I didn't mention earlier is it's fantastically expensive to have motors made to a specification so we've had to use motors that are readily available because we need them to be cheap to fit our design budget as well as the overall budget of it being a cheap machine to produce eventually that only leaves us mass produced motors so that the main actuator motor in there is from I believe a CD draw and we've also used the motor that scans a floppy drive head but the target is battery powered and a day's use is what we're aiming for because of course it consumes no power once it's got a display on it it switches off internally I think we're probably going to have to wrap that up I'm afraid but if anyone does have any more questions then I'm sure Russell will be available to answer them afterwards I'm going to hang around the tent afterwards if you want to chat Cool okay thank you very much again Russell