Teresa Escrig

News and oppinion about Cognitive AI & Robotics

We need Service Robots to feed disable students


Dear Teresa, My name is Paul Doyle and I am Head of Access R&D at Hereward College in Coventry. Hereward is a residential college that supports disabled students. We have for some years developed a keen interest in the use of robotics as an assistive technology.

I have been in contact with many providers of robots over the years from the PR2 at Willow Garage to the Care-o-bot by Fraunhofer with little tangible progress. What we have failed to achieve to date is to embed and evaluate an actual device in a real care/living/education environment such as Hereward to see if it actually works and if it is financially viable!

I would like to challenge any robot for example to help with the scenario I posted recently on a Linkedin forum:

Today when I was having lunch in our refectory I observed a number of students (with a variety of physical disabilities) waiting in an orderly queue for a human career to help feed them their lunchtime meal. Due to a shortage of careers some of the students waited for a very long time before a staff member could ask what the student wanted from the menu, picked up the chosen meal from the counter and then fed the student in an appropriate manner (food at the right temperature consistency and rate).
This situation led me to ponder the questions could a robot have helped carry out these tasks to some degree, and bearing in mind the care staff are paid not much over minimum wage, when (if ever) will a robot alternative be a financially viable?”

I would hope manufacturers could see this exposure to a group of users as a development resource, as we have a residential care and education setting where such technologies can be tested in a managed and safe environment.

Many of the young people at Hereward will eventually be the recipients of assistive robot technologies if and when they come online, so hearing what they need/want would I imagine provide a useful insight to product developers.

Perhaps we could explore and identify where your technology could and should be used in our college and test it with our students, care/academic staff and in house therapists and technicians.

We also have links with Coventry University (http://www.hdti.org.uk/) and have co-operated extensively on assistive technology evaluation projects over recent years.

I think my challenge to the robot community today is to get out of the laboratory and do something tangible!

I and my organization would be more than happy for any trial to be used jointly as a vehicle to publicize your device and raise the awareness of assistive robotics in the UK.

If you would like to discuss this matter in more detail please do not hesitate to contact me.

Kind regards,


Video: Hereward College Personal Robot User Survey

PDF Files with the answers of the students:

Hereward Project_Part1

Hereward Project_Part2

My answer:

Hi Paul,

Thank you for your touching email and bringing this to my attention. This is an important area that I was not previously aware of. One that clearly needs to be addressed.

We do not work with arms (as PR2 from Willow Garage) yet. I see this project as a collaboration among several companies/ research labs, experts in different areas.

We certainly can provide:

  • Intelligence perception: Qualitative interpretation of the distance sensors, the Kinect sensor and the visual sensor.
  • Reasoning with the landmarks of the environment.
  • Learning to repeat what works best.

We have solved autonomous navigation using the previously mentioned aspects of intelligence.

I know they can also be used to control arm manipulators, but we haven’t tested this at all.

I would love to be able to collaborate with other companies/ research labs to provide all the pieces of the puzzle together to fully address this problem.

I hope we can help you soon enough.

Thank you again.

Kind regards,

To the community: This is a real need, that deserve the integrated effort of technologist, scientist and investors. I am up for it. Who else?

by Craig Hines (LinkedIn Group: Robotics Guru) – What might be helpful is to encourage kids with gross motor coordination to hit a switch to tell the robot to feed them the next mouthful. Even if it is a kick switch, or a proximity switch in a tray, the feeling of control over action and reaction can give them a feeling of control of one little part of their destiny. This can be converted into a wheelchair driving switch for instance if they can show consistency of control.
Another useful switch would be in the back of the head rest, if they keep their head in the back position the machine feeds, if they move forwards the machine retracts. This is for safety, if they are tempted to lean forwards but don’t have enough head and neck control they may loll forwards into the feeder.
I may be making too many assumptions about your clients here?

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – Actually Craig has a good point. I mean pre-processing the food in such a way that it is able to be served from safe devices so the risk of injury is kept to a minimum. Liquid dispensers however can be tricky to handle, maybe the ball sized food can be a better approach, if we can find a design for a safe dispenser. I know it’s not the place to make fun or laugh about this but a scene from a movie keeps popping in my mind. It is Charlie Chaplin’s marvellous masterpiece “modern times” where he is attached to a feeding machine. Please don’t understand me wrong but if they coud think of such an easy automated device back then why can’t we adapt such a machine to aour current needs? i attach the link to an youtube video showing the machine in action . http://www.youtube.com/watch?v=pZlJ0vtUu4w

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Robotics Guru) – Agree with Craig. To have a bit of control can increase the level of safety and will make kids feel more comfortable with the feeding machine. It could be one of their buddy at school.
We could also use computer vision to recognize face, locate the mount, and also be a switch for the next feed. The machine can see when they turn their face away or lean forward to get more food. I’m not a computer vision expert, so I would like to hear from an expert whether it is practical or not.
The next questions would be what type of disability and food we are talking about here. In the email said there are a variety of physical disabilities, so it could be more types of feeding machines. If so, we can start from the easiest one or the majority type of disabilities.

by Demetrescu (LinkedIn Group: Robotics Guru) – Iulian@Panvadee there are some libraries and tools like OpenCV that can ease the burden of face recognition. More important is the communication between the disabled person and the machine. How and when one needs to be fed, how much, what happens when something goes wrong (choking,sneezing,etc).

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Robotics Guru) – Thanks for your response, Demetrescu. I heard about OpenCV. Sorry I was not clear about my question. Is the computer vision practical/possible in this application in term of processing capacity and respond time (how fast the machine can react from what it see). We would not want to put a super computer in every feeding machine.
The communication between the user and the feeding machine and how we gonna feed them bring us back to the question “what type of disabilities and what type of food we want to work on here”. I think the machine should be able to recognize when someone is choking or sneezing (capture quick movement for example), and may need to let the carer know when there is a problem. The carer should be able to observe around the room and see when there is a problem while they are feeding other students or doing other tasks. By the way, I don’t want it to be too complicated now, simplest is the best. We should start from a physical switch as Craig’s suggestion together with remote and local emergency stop switches 🙂

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – So in theory we have the basic “effector”. A robotic arm coupled to a SBC. A moderate resolution camera will capture the subject reaction, eventually interaction with the subject can be provided by the means of a standard LCD panel and audio speakers, wireless or wired action switches connected to the effector SBC.
But we are a little ahead of this project. We should establish somehow a project , a website, some hardware requirements, etc in order to proceed further.

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – I propose the name “RAFE”. Stands for Robotic Aided Feeding Equipment. We can setup a website and start.

by Teresa Escrig  – Hi all, I really appreciate your contributions and enthusiasm. I asked directly to Paul Doyle, the Head of Access R&D at Hereward College in Coventry about your questions and he will respond to them. He has been out of office.  I am not going to be able to lead that web site that you are talking about, but I can certainly lead the research and provide our Cognitive Brain to the project. I can happily be an adviser, a contributor.  I would not start something unless I know that we are able to accomplish something (but that’s just me).

by Paul Doyle (LinkedIn Group: Robotics Guru) – Thank you. I am really excited to see there is so much interest in our plight. I wholly endorse the idea of a vehicle to facilitate this discussion and a project site is certainly a good idea I am not sure we have the capacity to develop such a site but am more than happy to support the the development of the project by bringing the real needs of physically disabled individuals and those who care for them to the table.
What I have suggested to Teresa is that a site amongst other uses could also serve as a means of bringing the user and the developer together, in order to best capture the needs and aspirations of young disabled people.
I stress young as so often the development of assistive technologies focus on older users,but our students will hopefully be the using a range of robotics technologies for many decades so the y have a vested interest in their development.
I think feeding is an essential activity as well as accessing hydration and these activities of daily living serve as a realistic starting point for the development of useful cost effective personal robotics. Again I emphasise cost effective as to be of practical use such technologies need to have a sound financial model underpinning their provision. Needles to say safety is at the core of the project too.
If you would like to know more about the needs of our student population please do not hesitate to contact me directly.
Lets see where this goes. Many thanks. Paul

by Paul Doyle (personal email) – Perhaps we could link our students up to the group of experts?
A virtual discussion, or even a Skype session? I think the idea of a project site would be fantastic, if students were able to use it to communicate their needs directly it would help break down some of the barriers and misconceptions that have traditionally arisen when technical experts try to visualize the issues encountered by the disabled community. Getting to know real users will help contextualize some of the issues we have raised. Kind regards, Paul

by Teresa Escrig – Paul, that would be absolutely great! It can change the way things are going in technology in many ways: first users and technicians will be closer, so we can solve what it’s really needed; and second, we might attract some investor who likes the synergy created and get funded to make it really happen.
Please, ask your students if they would like to get involved.
I will post this communication to LinkedIn and my blog.
Thanks, Teresa

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – Teresa, Thank you for the reply. I was personally interested in your opinion regarding the entire project. I agree with you that we first have to hear from Paul, the actual problem issuer if he is willing to embark in such a project. As far as i understood from the letter Paul wrote to you, we are talking mainly about children or young people, who have the greatest chance to recover from their condition. Personally i am willing to put up the website providing space, domain registration, storage, maintenance and collaboration tools in order to move this project ahead. Also i have some expertise in embedded systems, electronics design, embedded programming, skills that can be of great help in such an endeavor.
Of course we can use any resource and knowledge you can share and we could at least outline some directions and sketches before committing to something. If we can reach a theoretical model that can work, i think we can risk a PoC system that can be minimal in terms of hardware requirements but with the core features present.
If Panvadee steps in, it will be great because he has also experience in robotics and can help with the hardware part also. As a side note, I will move to Australia in the next month and i will be closer to him, so we could meet either in Sydney or Canberra in order to test and debate over hardware/software issues 🙂 .
As for the accomplishment, if we understand each other, if we debate all the facets of the project and reach a conclusion, we can accomplish what we want. All we have to do is to really want to help.

by Teresa Escrig (LinkedIn Group: Robotics Guru) – @Demetresku, I really appreciate your passion and enthusiasm.
I have been directing many big projects and teams to solve real problems including cutting-edge technology, like this one. As I know what it takes to make this kind of things happen, i want to make sure we have or can have all we need to succeed, before we start to dedicate time to this project.

by Craig Hines (LinkedIn Group: Robotics Guru) – From what I can gather from the web site, Hereward College is for the 18 year old and up group (“College” has different meanings in different countries.) Disability at Hereward could mean any condition from autism, blindness, to MS or spinal injury. From my experience in customizing power chair drive and powered seating systems and electronics, disabled people come in all shapes and forms. You can’t count on their head being in centre position or even facing forwards or in an upright position in their natural relaxed state. Some conditions make people so weak it literally takes a minute for them to get their hand to push a button. Others can’t do fine movements, when they do their muscles spasm and break equipment – this condition is a permanent workout, their muscles can be surprisingly strong. Some conditions come with a childish irresponsibility, though not intellectually handicapped – the sort of person that shouldn’t have been left in charge of an automotive air horn in a supermarket (it happened!)
What I am getting at is, I think the feeder needs to be adaptable to a variety of positions and external equipment that might get in the way. It needs actuator force to be applied by spring loaded mechanisms so that a lurch forwards does no harm to the user or equipment. Each user may need their own settings, and if there are several machines, a networked profile lookup could help so that any machine can work with any user. Caregivers and teachers probably aren’t technicians, if the machine doesn’t do what they expect, it will end up in the broom cupboard. Industrial robotics these days is more user programmable. If the caregivers can train the machine by physically moving the arm and feeder mechanism they will be more confident about using it. Some users may be able to do this for themselves, so the machine simply assists them where they wouldn’t normally be able to hold their hand still enough to self feed.
On a practicality point, it seems that possibly parallel prototypes could be built in Sydney and in the UK (Coventry), maybe at a certain point a media release could drum up some funding.
It is so encouraging to see so many experts being inspired by this challenge. I don’t know if I can contribute much to the technical development, but I hope to learn from it.

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – @Craig, your experience in customization can bring a great deal of inside information in the behaviour and different needs of each case you encountered over the years. Please don’t minimize your work since, as I see it, is a valuable source of information. For the identification issue I was thinking at a comfortable wrist RF-ID tag similar to those used in some hospitals. That tag can carry patient information and also can be used in future home-friendly appliances ( intelligent house,etc), I’m dreaming yes i know…
Also a great deal of information will have to come from some specialized doctors that can give us a medical view of the problems we are facing, risks associated with each disease we are trying to manage. We have to be aware of the fact that we are ( or at least we think of) creating a complex device that will interact at an intimate level with human beings. Our conception will have to pass a huge set of safety rules and regulations due to the fact that it will be perceived as a medical equipment. We need a close relationship with the medical environment if we want to have a chance to make this work.
It’s not important where the device/devices hardware will be created. We will share the bits and pieces of hardware and software ( despite the great distance between UK and New Zealand, India, Australia,etc). We have to ask Teresa to lead the way and eventually at some point let’s have a conference over skype to set some things up.
Also we have to attract more developers and robotic technicians because ” more eyes have a bigger picture “.
In order to attract some of the funds we most probably need, we can try first at the humanitarian organizations. These associations can help us buy at least the more expensive items that for sure will not be donated. Of course at some point we can have a press conference and if we can achieve at least a limited working prototype until the next year, we can enroll a talk on TED (www.ted.com) which will for sure attract more developers and founders.

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Robotics Guru) – Please count me in. My experience is around developing cost-effective robots and bespoke devices (mostly Defense projects). Having a clear scope and requirements of a product will reduce the cost from unnecessary and/or expensive solutions. One of the most important factor for a successful product is the interaction between the end users and the developers (and we will have that). If we have a limited budget, we will need to narrow the scope down and try to maintain the must functions. It does not sound great, but it is a reality. In the car industry, we have affordable car and luxury car. Same here.
I think I understand Teresa’s point. At the moment, we still have many holes in the project. I think we can start small, and I will promote this project among my connections as well. My skill set is technology integration, system design, mechatronic design, electronic and PCB design, logistic and manufacturing. I hope that I could help more or less. My next step would be to summarize the requirements from what we have so far.
@Demetrescu, I can help you with the website if you want. I’m not a programmer, but I can help with words and design (and if you need a second opinion). Good news that you are moving to Australia, how’s convenient. We can arrange a meeting when you are here. BTW, I’m a she not he.

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Robotics Guru) – @Demetescu,
>>For the identification issue i was thinking at a comfortable wrist RF-ID tag similar to those used in some hospitals. That tag can carry patient information and also can be used in future home-friendly appliances ( intelligent house,etc), i’m dreaming yes i know…
No, it is real now. It has already been used in home automation. RFID solution is not expensive, we should be able to use it.

by David McMillan (LinkedIn Group: Robotics Guru) – As a professional user of heavy industrial robots, I have to confess that the idea of putting robots into direct contact with young people, especially disabled ones, makes me extremely aprehensive. Safety for a RAFE system would would be absolutely critical, but also difficult to achieve.
Some of this can be addressed at the design level, by making the manipulator low-powered and compliant (a bit like Willow Garage’s PR1), but even that has limits — a robot capable of putting a small spoonful of food into someone’s mouth is easily capable of sticking that spoon in their eye. And errors *will* happen.
In an industrial environment, general practice is for anyone working within reach of a robotic manipulator to carry a “deadman switch”. This multi-stage switch must be held partially closed (first stage) in order for the manipulator to have power. Releasing the switch or squeezing it too tightly (ie, a panic reaction) kills power to the manipulator, freezing it in place. For the proposed RAFE system, I imagine that some equivalent would be absolutely necessary, but given that we are dealing with disabled persons, and a wide variety of disabilities, creating such a safety presents a serious ergonomic challenge. This is one of those areas where the design engineers would need to work closely with the people experienced with the physical disabilities in question. For example, I’m having a very difficult time envisioning a safety-qualified “deadman” for a person whose disability leaves them only able to manipulate suck-and-puff controls. But the Hereward staff, being experienced in this area, might well be able to provide the critical insights necessary to create a functional safety interlock.
Getting the right food to the right people could also be a safety-critical item. I once was peripherally involved in a case where an elderly man whose swallow reflex was impaired after a severe stroke was accidentally fed something other than thin, clear fluids, and who developed severed respiratory complications as a result, which probably hastened his eventual demise. This is another area where the Hereward staff would bring absolutely critical knowledge to the endevour.

by Paul Doyle (LinkedIn Group: Robotics Guru) – David, I agree safety should be at the heart of the project. Managing risk is something we all do on a daily basis and this will apply to Hereward students too. The designing out of risk is an issue that the powered wheelchair industry have years of experience of but accidents still happen. You are right to point out that getting the balance right (balancing risk against benefit) is an outcome worthy of consideration.

by Paul Doyle (LinkedIn Group: Robotics Guru) – I would be interested in the safety protocols adopted during this experiment!

by Craig Hines (LinkedIn Group: Robotics Guru) – These are definitely interesting times for emerging technologies for the disabled and the cross over with robotics and bionics. But even once this technology, and exo skeletons become more mainstream, it will be a long wait till it is affordable for most. Powered wheelchairs have been around for decades but are out of reach for many who don’t quite fit funding criteria in NZ.
Back on safety and economy. Both could be addressed to some extent by just focusing on people who can move their heads towards the spoon for the final gulp. A simple collision avoidance IR distance detector could freeze the robot arm about 30mm in front of the user’s mouth, and when their head moves back again the arm retracts for another spoonful, food ball or whatever.
Let humans handle the trickier cases.

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – From my point of view, spoons and forks are out of the question in our system. Serving liquids hot or cold from spoons can lead to unpredictable results( spills, burns, etc) if the patient moves unexpectedly and the robot arm decides to abruptly stop the move. Better go on the already paved road with a recipient and a straw. It’s already verified and works. Also the robot arm must be provided with protection against accidental collision with human body parts because the shear brute force of an uncontrolled human arm or head slaming into the robotic arm can cause severe injuries. Maybe some sort of IR proximity sensors that can trigger a swift retract of the entire robotic arm when such a human move is encountered. Since i have dealt with MS i can tell you that sometimes muscles can contract or extend uncontrollably generating hieratic movements of the limbs, head, torso, etc. For reference we can try to observe the feeding process of a small child that refuses food. His head and upper limbs move unpredictably but the person feeding the child in most cases is able to avoid any fatal collision between the feeding arm holding the spoon and the child. Then we can build avoidance patterns that can be programmed inside the AI of the robot in order for this to react when sensors are triggered.
If we are talking about solid food, here the problem is even more difficult to attack. Different diets, type of food, consistency, etc will generate a whole load of types of effectors that can grab and serve that kind of food. But if we look at the problem from another angle, we might get away with just one kind. Reverting back to the feeding process of a child, i recalled the times when my son was young and also the times when i was taking care of my Ex. In both situations the solid food was server directly with the hand in form of small portions, enough to be eaten comfortably without the risk of choking or producing injuries. I think that by creating an end effector with a general shape of human hand ( not the entire hand, just three fingers counting from the big one which are the most used) elastic enough to bend if forced gently but stiff enough to hold and direct food, can be a real success, again from my point of view. This can be unappealing to many including the patient but i think that the human hand holds the answer to this issue. It is used to feed ourselves on a daily basis from ancient times and rarely fails if handled correctly.

by Paul Doyle (LinkedIn Group: Robotics Guru) – Demetrescu, We need to be careful not to neglect the emotional needs of the individual/user in a pursuit of functionality. I say this because having worked with disabled people for some years I have found some technology is not used not because of what it does but instead of what it means. We have used the following device for many years http://www.neater.co.uk/products.htm but a lot of our students still prefer a human feeder for a host of reasons even though technically this device and others like it can enable them to feed themselves. I think there is a kind of subtle emotional transaction taking place whereby the user needs to want to sacrifice the benefits of interacting with a human carer for the ability to do something unaided. In an ideal world there would be space enough for both choices but as we are in difficult financial times and the reservoir of carers is diminishing whilst the cared for population grows this is not the case.

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Robotics Guru) – Have you seen “Luke” arm?
It is a bit old, but relevant here. The test pilot and engineers look happier in the video 🙂

As we all seem to relate the solution to a feeding robotic arm, I am just curious.
@Paul, any other processes apart from feeding will be worth looking at? You wrote about asking the student what they want from the menu and bring the food from the counter. If we automated that process first, will it help a lot? I can imagine that feeding would take the longest and would be a bottle neck here, but the other processes will be easier and will take less time to develop.
We can also work with food scientists to develop a easy-feed-and-taste-good food for the machine.

by Demetrescu Iulian (LinkedIn Group: Robotics Guru) – @Paul, i would never neglect such an important issue. That is why i think it is imperative to involve the students into the development process. Any idea, even a crazy idea that can make them feel safe and confident in using this device can be adapted to our system. The student have the most crucial impact to this project. They will teach us how to help them not the other way around, us teaching them to eat or dress or use the facilities. We have to keep in mind that their main goal is not to live in a “special world” they want to live and share what we call our “normal world”. They will agree to change their food consistency to some degree but in order to be successful they have to be able to eat “normal” food as well.
Maybe the greatest idea of this project is actually the deep involvement of the students in it. We should ask them what features they want inside, we should put their minds at work in designing some or all of it. I’m sure they will com up with some idea that can be put to practice.
@Bua for sure we will ask them once we have the occasion what other processes can be robotized but Paul can give us an insight on the day by day activity in the college.

by Wolfgang Heller (LinkedIn Group: Personal Robotics) – Take a look at Bestic http://www.bestic.se/en/home/ a nice solution from Sweden. Links to other meal-assistance devices can be found at http://webbrain.com/u/135I

by Craig Hines (LinkedIn Group: Personal Robotics) – We had a speaker at a company meeting who was an ex client, power chair user with muscular dystrophy. He had to make the transition from mobility scooter to power chair, as insisted by his occupational therapist. Once he had his power chair he left it sitting in the corner for about six months. To use it, he said in afterthought, was to accept that his condition had worsened, but also, the power chair meant that he was “disabled”, unable to use his old mobility scooter. Finally, he had a crash on his scooter and was unable to raise a hand to fend off the ground and gravel or protect his head. When he finally accepted the power chair he wondered why it had taken him so long – it actually gave him more freedom and more ability.
Along the lines of what Bua said, if the arm could perform other tasks and become a personal piece of equipment rather than the shop feeder maybe it would be accepted more widely. Send txts, operate iPad for speech assistance, iArm. Probably getting too complex with this though.

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Personal Robotics) – If the kids can play games, watching kid’s programs or some interaction activities while being fed, it may make them want to be fed by the machine. It would be a plus if we can add some of these functions into the machine.

by Demetrescu Iulian  (LinkedIn Group: Personal Robotics) – @Craig & Bua i couldn’t agree more. we should’t limit to the feeding exercise but expand the capabilities of the device. This is why i think it’s imperative to talk to the students. Eventually Paul can try and start some sort of brain storming with the students in order to gather some thoughts and ideas from them.
Last evening i was reviewing the entire discussion and it appears to me that the route of some impersonal cold metallic robot arm will not be so appealing to anybody. We have to think out of the box and come up with a solution. In my wildest dreams i would go as far as building something around the human body not external to it. I mean a partial exo-skeleton to smooth/enhace/correct/revive an arm for example. I know it souds SF but looking at the movies and products presented so far, the most successful solutions tend to be the direct replacement foR the missing limbs with their robotic counterpart. If we can conceive and exo skeleton capable of controlling the student’s arm, for example, i think we can have a lower rejection figure from the ancients. This is not so far from the original idea but it’s another angle of approaching the problem. If we go this way the students will be able to feed themselves, and do many other things they couldn’t do until now.

by Demetrescu Iulian (LinkedIn Group: Personal Robotics) – @Bua, my thought was more on the line of an exo skeleton like http://www.theverge.com/2012/5/9/3008979/paralyzed-woman-bionic-suit-marathon
but designed to be concealed under the clothes. Very light builds from aluminum with carbon fiber, maybe using muscle wire, that was the idea…

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Personal Robotics) – The video showed that the students want a personal robot (RoboBuddy) rather than just a feeding machine; however, to come up with all of those requirements in one go would take us forever. In my opinion, we should separate those requirements into sub-systems; for example, AI module, robotic arm (manual), vision system (integrated with other sub-systems to create a smart system), physical movement (including exo skeleton), Speech assistance, etc. Targeting a shorter goal and creating a usable equipment as soon as possible will bring the advantage to both the students and our group. The student will have a reliable user-friendly equipment for the most needed task sooner, while we will have some success to attract more people to get involved.
Our group is still too small comparing to the tasks ahead. First of all, I think we should help Teresa promoting this campaign. She prefers to use Tweeter. I will start from getting a Tweeter account and help re-tweet hers. Secondly, I want to finalize the requirements and get the specification for the first device which has a possibility to have a prototype within a couple years. We can start with a device that involve only one sub-system or two and add it up step by step; for example, a robotic arm (just feeding or pick up things) with AI module or a speech help with AI module. At the end of the day, we will have to consider developing a device that suite the team’s skills and budget as well. Then, I would like to hear from the group what do you think we should develop for the first step, based on students’ needs and your skills. In the future, if we can get other companies or people to help out, we would have many sub-systems developing at the same time.
I am a system (mechatronic) engineer. This is what I do; develop requirements, create the product specs, benchmark, choose to develop in house or using off-the-shelf technologies, design the whole system (hardware and software) with the team, assign tasks, design E&E system, design electromechanical system, system integration, electronic design, building prototype and test, engineering commercialization and manufacturing. I can help designing a robotic arm, vehicle, or a device, and basically anything but the high level programming (I know how it work. I can edit and understand the code, but I have never written the code myself). I usually do C in micro-controller. If we have something solid, I can also talk to my employer about this project. They are a bunch of investors.

by Teresa – @Bua, your way of approaching the project is the one that makes more sense.
I know we want to help, therefore the effort, the goals and the resources (both human and financial) need to be in place to succeed. Being backed up by investors would help a lot Bua!
The solution the students were talking about in the video is very challenging for the current state of robotics nowadays.
The questions were very vague to make sure that if we provide something similar to what they are asking, they are not going to get bored very easily. The survey was a starting point, but in my opinion, not enough to justify the project yet.
Let’s continue working, talking and getting clearer.
MIT researchers have created a robotic elephant trunk that could be useful to pick up things from the floor: http://www.gizmag.com/universal-jamming-gripper-trunkbot/22610/

by Panvadee (Bua) CHUASRITRAKUL (LinkedIn Group: Personal Robotics) – Below is my draft requirements for the first version robot. From this requirements, we will do the brainstorming, to collect possible solutions, to create the specification (product’s requirements in details) and then we will have a product concept and proposal to ask for funding. As always, I am open for feedback and opinions.

RAFE V1 (Robotic Aided Feeding Equipment version 1) Requirements Summary:
[We need a new name because now the aim is not just feeding]

1. An equipment/machine that help users with disabilities (age 18+) to be able to feed themselves and help pick up a small thing from the floor or on the shelf as closest as the majority of people can do. The users can be: any wheelchair users with various physical disabilities except vision impairment; must not be a person with cognitive/emotional issues, refers to medical personnel.
2. User interface should be vary depends on the users’ abilities. There are two groups of users: (1) users with muscle-control difficulty and (2) users without muscle-control difficulty.
3. The machine should be able to interact with users while feeding them. In another word, the machine has to make the user feel that they are not alone or forget that they are alone.
4. The machine needs to look friendly and to make users feel safe and comfortable around it.
5. The machine has to be extremely safe and robust for the users. For example; it has to be soft-touch and able to handle an unexpected force. [need more details]
6. The machine has to be cost effective and reasonably comparable with the stuff rate [required stuff rate and number of students per staff], may be a break-even within .. years.
7. The machine has to be easy to operate and to perform minor-maintenance by non-technical staff.


6 Responses to 'We need Service Robots to feed disable students'

Subscribe to comments with RSS or TrackBack to 'We need Service Robots to feed disable students'.

  1. This is a great talk. We’re building Intelligent robots perform, sometimes, unnecessary jobs – while there’s a great needs that we don’t care about! Thank you so much, you made me think different now!


    13 May 12 at 11:37 am

  2. It’s great that the students can give us feed-back!
    Are they excited too?

    I am not sure where this is going to go, but at least there is a pure intention to help behind the experts in that LinkedIn threat. And that’s meaningful.

    Let’s support it.

    Kind regards,

    Teresa Escrig

    16 May 12 at 6:53 pm

  3. That will be a great step …


    16 May 12 at 7:03 pm

  4. This is a very interesting and worthwhile project. Is there a website setup for the project?

  5. As part of my MS Project, I had built a prototype of a robotic human feeder, which could be operated by a single switch. It also had sensors to detect obstacles. The robotic feeder could also be paused by the user. The robotic arm moved from the mouth to the bowl and back, with a spoon attached to scoop food from the bowl. The height of the spoon , however is fixed and may have to be customized as per the user.

    Unfortunately, i couldnt realise the end product due to lack of funds and the prototype is lying in my cupboard. 🙁

    I have had some previous exposure working with spastic kids and have designed cheaper alternatives of the Jelly Bean switch to replace regular keyboards and computer mouse.

    Jude – Bangalore, India

    Jude Pereira

    12 Jun 12 at 3:42 am

  6. Hello to everybody, my name is Luis Garcia and I’m developing a robot that might help children to have a fun and happy meal time.

    The name of the robot is HelloSpoon, and it’s intended to be an affordable solution for children.

    I’m starting the testing stage (I’ve been working in HelloSpoon a bit more than a year by my own) and I would like to share here with you my last video:


    The idea behind HelloSpoon is offer something cute, fun and affordable. And it’s based in a mobile phone, so all the audio feedback can be enabled/disabled if the user feels awkward or something similar.

    Please I invite you to subscribe to the Youtube channel to see the following testing with children.
    Or send me an e-mail to:

    Thank you very much! and I hope it doesn’t look like spam hehe

    Luis Garcia

    3 Jul 14 at 9:53 pm

Leave a Reply