Добавил:
ivanov666
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз:
Предмет:
Файл:Readings in Robotics Engineering. Учебное пособие
.pdf
2.10. The challenges of controlling robots using our thoughts
71
terms of just seeing a performance that I’m generating, obviously, with the as-
sistance of the team. The difficulty being that I understand the more we can do
this, the better our results will be. And just living with the normal health issues
that come with somebody who’s got my level of disability. It can slow you
down a little bit and become quite frustrating. There’s obviously some very
experienced teams out there. We met a couple in the test run. None of the pilots - just some of the tech leaders.
3:45 And I understand that there’s universities out there that do nothing
but BCI research. And they’re obviously in a much stronger position than the
majority of the other teams. But we’re going over there to try our best, we’re
not just making up the numbers. We’re going to go see if we can get a podium
finish. Aim for the top and see what happens. The potential for the BCI system
and the EEG system we’re using, I mean, there’s obvious potential in medical
fields and diagnosing or helping people with brain injuries and so on. But the
fact is the applications are limitless. We’re in such an early stage of its re-
search, and that research could take you anywhere.
4:22 It could take you to playing computer games, like we’re doing. It
could control wheelchairs in the future. It’s literally limitless, the applications,
we’re only dipping our toe really and playing about compared to what it actual-
ly could be applied to in real life.
The challenges of using mind control
Generally, in BCI, brain activities are recorded using EGG signals. In
this activity, we'll uncover why dealing with EEG signals is such a challenging task.
2.10. The challenges of controlling robots using our thoughts
In recent years, BCI research has lead to a large variety of proof-ofconcept BCI systems. However, despite this impressive expansion, none of the
BCI systems described in the literature are sufficiently mature for daily use out
of the laboratory.
To make BCI an intuitive interface for use in either control, communication or rehabilitation, improvements are required in several areas, such as usability, signal acquisition techniques, hardware development, machine learning
and signal processing, and system integration.
Part 2. Learning from nature
72
Generally, in BCI, brain activities are recorded by EEG due to its high
temporal resolution, non-invasiveness, relatively low cost, and portability. Despite these advantages, dealing with EEG signals is a challenging task.
Noise in the nervous system
The EEG signals are very noisy. The information of interest is normally
mixed up with several irrelevant but concurrent neural activities called central
nervous system (CNS) noise.
CNS noise may interfere with the information of interest due to overlapped characteristics. For example, BCI systems that work based on imagination
of body movement involve the brain patterns in the μ-rhythm (i.e. 8‒13Hz).
However, the α-rhythm, which represents the visual and mental efforts, has the
same frequency range as the μ-rhythm. Although the α-rhythm originates from
the occipital lobe of the brain, due to the volume conduction, it interferes with the
μ-rhythm that originates from the motor cortex.
Besides, the α-rhythm is quite prominent, whereas the μ-rhythm is
weaker and can only be observed after appropriate signal processing.
Non-stationarity
Another problem related to classification of EEG signals is nonstationarity. Variation in the EEG properties, called non-stationarity, can result
in deteriorated BCI performances as most of the machine learning algorithms
implicitly assume stationary data.
This means if the model is fitted to the subject’s training data, it may not
be optimal for a new session (recorded on another day) or even new trials in the
same session. The plausible reasons for the changes in the dynamic of the EEG
signals are as follows:
• The physical properties of the EEG electrodes change over time. For
example, the impedance of the electrodes can change due to drying up of the
conductive gel, or the position of the electrodes can change when the EEG cap
is reused in a new session.
• Neurophysiological conditions (e.g. fatigue) can be variable over time.
• Psychological parameters, such as motivation, attention and task in-
volvement also display a large variability.
• Artifacts caused by body movements or muscular activities such as
swallowing or blinking change the signal properties.
2.10. The challenges of controlling robots using our thoughts
73
These drawbacks can clearly impede the continuous use of the BCI
system. BCIs in real-life situations must be accurate and robust in their application. A patient needs a system that functions accurately all the time and is
able to adapt to new circumstances. Therefore, it is necessary to investigate
these issues and find solutions for improving the accuracy and robustness of
the BCI systems.
The solutions include developing advanced signal processing and machine learning algorithms as well as training users in generating stable brain
patterns over time using feedback training approaches.
Reading: ROBOTS FOR HUMAN ENVIRONMENTS
Do the preparation task first to help you with the difficult vocabulary. Then read the article and do the exercises to check your understanding.
Preparation
Match the vocabulary with the correct definition and write a‒h next
to the number 1–8.
1. huge a. task which is difficult but stimulating
2. to cope b. very big
3. to tie (smth) down c. to confront a problem successfully
4. a pattern d. to attach something with ropes so that it cannot move
5. a challenge e. a regular or repeated way in which something is done
6. a tool f. artificial intelligence
7. robotics g. an instrument that helps us do a task
8. AI h. the design and construction of robots
According to Dr. Nick Hawes, the future of robots is going to be
huge. Here he explains the challenges
‘There’s this huge excitement around robots,’ says Dr. Nick Hawes,
Senior Lecturer in Intelligent Robotics, School of Computer Science, Universi-
ty of Birmingham. ‘Everyone really believes, as we do ourselves, that robots
are going to have a huge impact on our future - in workplaces, in roles in vari-
ous industries.’ But there is one problem that motivated Dr. Hawes and the
group at Birmingham in their research. ‘The fact that these robots would only
function for two hours, and only do one useful thing once, made us think that
we are not getting close to doing the science that will allow robots to have this
huge impact.’
Part 2. Learning from nature
74
The result of this was ‘Bob’ the robot, which received much media at-
tention as ‘Bob’ was shown working as a security guard, patrolling the offices
of a security company. ‘Bob’ was designed as part of a research group called
STRANDS, and is made up of seven universities across Europe. ‘The
STRANDS project has focused on what we can do to make a robot function for
more than an hour or two - for days, weeks and months,’ says Dr. Hawes. There
are two interesting things about this, he says.
Controlled environment
Firstly, there’s the science and engineering challenge of making an ‘au-
tonomous robot, a robot that can do things for itself, function for that length of
time in an environment it has no control over. Normally when you put robots
into places you have to control everything, to tie things down, make sure no-
body gets in the robot’s way. You want to be able to make a robot cope in a real
human environment.’
Secondly, there are real advantages when a robot can function for an ex-
tended period of time. ‘The robot can start to learn things about this environ-
ment that it wouldn’t ever see normally,’ says Dr. Hawes. ‘It gets to see daily
routines and patterns: what time people come and go; where you put your mug
of tea on your desk every day; things that humans have a common-sense understanding of, but robots don’t. Our aim is for robots to learn that over time.’
From driverless cars to washing machines
But do we have a fixed image of robots and have certain expectations of
what they look like and how they operate? Is our image of robots too human-
like, and is that a negative factor? ‘I think that is the way science fiction has
shown them to date,’ says Dr Hawes. ‘I’m not really interested in robots that
look like humans. There are some advantages to having human-like features,
humans naturally understand other humans from their physical movement.
Having a robot with some human-like physical movement – with eyes looking
at places, positioning your body to look. Humans understand that as having
some meaning, and just generally it makes them feel more comfortable about
other robots.’
He points out that our environment, from door handles to cupboards, is
created for humans so having a human-type robot is easier to fit in. ‘But at the
same time,’ he argues, ‘robots are really tools, a technology. Their form should
be dictated by their function. A driverless car is a driverless car and that’s a
2.10. The challenges of controlling robots using our thoughts
75
robot. To some a washing machine is a robot. There are many autonomous intelligent machines that can do things on their own. Some of them may end up
looking like humans but most of them won’t.’
The next step with Bob is extending the amount of time he can function.
Other STRANDS partners are working in the area of ‘care’, in a hospital in
Vienna where a robot is doing some work. ‘There is going to be a big industry
building and programming robots so we think our students need to be looking
at that,’ says Dr Hawes. So at the university they ‘teach robotics; we have a
Robot Club where students work on Bob and similar robots. We are really trying to get everyone, from 18 upwards, working on this technology because it is
going to be huge.’
If you're interested in science check out the British Council's science
magazine called Cubed.
1. Check your understanding: true or false
1. People think that robots are going to have a big effect on some areas
of our lives in the future.
2. According to Dr. Hawes, robots were limited because they could not
work for a long period of time.
3. There was little interest from newspapers and TV when Bob the ro-
bot started working.
4. A robot normally has no problems functioning in a place where hu-
man beings live and work.
5. If a robot can function for a longer period of time, it can learn new
things it didn’t understand before.
6. Dr. Hawes says that the majority of robots in the future will look
like human beings.
7. The next thing that needs to be done with Bob the robot is to make
him able to work for a longer period of time.
8. Dr. Hawes thinks that only a small team of experts should work on
this technology.
2. Check your vocabulary: gap fill
Read the text and write the correct form of the word in brackets to
complete the gaps.
Look at the example at the beginning of the text.
Part 2. Learning from nature
76
Many scientists feel very ______ (EXCITE) about the future of robots.
According to Dr Nick Hawes, robots are going to be of great (IMPORTANT)
in many industries. However, one thing that was (PROBLEM) for scientists was the fact that robots could only work for two hours and do one specific
task. It was important to invent a robot that had the (ABLE) to function for
more than two hours and so Bob the robot was invented. A robot which can
function for a _____(LONG) of time superior to two hours can begin to learn
new things. The _____(OBJECT) of the Bob the robot project is to see if Bob
can learn about his environment by seeing what is going on around him. It’s
probable that most robots will be _____(LIKE) how many people imagine
them and most will not look like humans at all. Although there are some advantages to having robots which look like humans, robots that look more like
machines than people shouldn’t be seen _______ (NEGATIVE). Dr Hawes
thinks that the future of robot technology is very important, and this should
give (ENCOURAGE) to students who are interested in this field.
Discussion
Would you like a robot to help you in your daily life?
What would you want it to do for you?
Would you prefer your robot to look like a human or a machine?
www.britishcouncil.org/learnenglishteens
The future of learning from nature
We look what the future holds for biomimetic research and discover
some of the ways this technology could shape our daily lives
2.11. What does the future hold for nature inspired research?
Biomimetics is a research field that is achieving particular prominence
through an explosion of new discoveries in biology and engineering. The field
concerns novel technologies developed through the transfer of function from
biological systems. Over the last decade, there has been an explosion of important discoveries within the field of biomimetics. The societal and economic
impacts expected to emerge from these advances will have future benefits for
our health and quality-of-life, due to advances in information and computation
2.11. What does the future hold for nature inspired research?
77
technologies, robotics, brain–machine interfacing and nanotechnology applied
to life sciences.
In 2013, Dr Nathan Lepora, Professor Paul Verschure and Professor Tony Prescott investigated the impact of biomimetics within engineering and related sciences, analysing a comprehensive database of publications on biomimetics.
In particular, their study focused on a few key questions.
• How rapidly is the subject of biomimetics expanding?
• What subjects does biomimetics encompass?
• Are there research communities within biomimetics?
How rapidly is the subject of biomimetics expanding?
From a relatively small field of tens of papers in the mid-1990s, biomimetics has exponentially expanded thereafter to now reach nearly 3000 papers
per year.
The subject area has doubled in size every 2–3 years, far outstripping the
modest expansion of about 6% per year for science in general (Larsen and von
Ins 2010).
Based on this finding, there is a boom in bioinspired research, with leading discoveries in biomimetics laying the foundations for large areas of present
and future research.
What subjects does biomimetics encompass?
The results of this analysis are displayed in a word cloud of frequent
terms in biomimetic research.
Popular topics in biomimetics
The word cloud shows the popularity of terms occurring in the titles of
papers on biomimetic research. The word size is proportional to the frequency
of word occurrence.
Part 2. Learning from nature
78
Word clouds, and data clouds more generally, are a visual depiction of
the frequency of words within a larger set obtained by scaling the font size of
each word within the cloud by its frequency of occurrence
As expected, the word biomimetic is the most popular word. Then, per-
haps more revealingly, other leading terms are ‘robot’ and ‘control’, which sug-
gests that a main thrust of biomimetic research is to take inspiration from how
animals control their bodies and sensory systems for application to robotics.
Concepts from control engineering and artificial intelligence are also
represented, including model, network, algorithm, simulation, learning, adaptive and optimization.
Are there distinct research communities within biomimetics?
This question was addressed with techniques from network theory applied to a graph of frequent biomimetic topics linked given by common pairings within the titles of papers. Terms that are strongly connected can then be
pulled together on the graph, while disparate topics are pushed apart.
Connectedness of popular terms in biomimetics
Applying a modularity analysis to this network showed that the field of
biomimetics was well connected and may thus be considered a single discipline. Underlying this inter-connectivity was a community structure into five
identifiable research themes (shown in the network graph above):
2.12. How might bioinspired robots help us in our future? Video
79
• Robotics and control ‒ general robotics and control, not specifically
bioinspired or bio-related (shown in blue)
• Ethology-based robotics ‒ robotics based on the study of animal be-
haviour (shown in black)
• Biomimetic actuators ‒ synthetic actuators that mimic biological ac-
tuators, such as muscle (shown in yellow)
• Biomaterials science ‒ materials and processes associated particularly
with biology, such as tissue or adhesion etc. (shown in red)
• Structural bioengineering ‒ structures and movements associated
particularly with biology, e.g. wing or flapping (shown in green)
Conclusions
Biomimetics is a research field that is achieving particular prominence
through a wide variety of new discoveries in biology and engineering.
There has been a rapid expansion of publications on biomimetics from
the mid-1990s to present day, doubling every 2–3 years to now reach a mature
field of nearly 3000 papers per year. Furthermore, the field is still expanding,
and so more growth can be expected. There are a number of distinct themes
into which biomimetics can be partitioned; robotics and control, ethology-based
robotics, biomimetic actuators, and biomaterials science and structural bioengineering. Taken together, these findings indicate that biomimetics is becoming a
dominant paradigm for robotics, materials science and other technological disciplines, with the potential for significant scientific, societal and economic impact over this decade and into the future.
Adapted from: Lepora, N. F., Verschure, P., and Prescott, T. J. (2013)
The state of the art in biomimetics. Bioinspiration & Biomimetics, 8(1). © IOP
Publishing. Reproduced with permission. All rights reserved.
2.12. How might bioinspired robots help us in our future? Video
We take a look into a realistic future for robots inspired by nature.
Emily Collins is a PhD student in Psychology and in this video, she explains her research which explores how human psychology and emotion can
improve the biomimetic robot, Miro.
Part 2. Learning from nature
80
We’ll then hear insights from educators on some of the future applica-
tions of bioinspired robots.
Transcript
0:05 I work with a robot called Miro that was developed here at Shef-
field Robotics in conjunction with some external partners. And inside Miro
there is a model of emotion that’s based on the continuum model of emotion,
and this is the idea that instead of having discrete emotions, all emotions sort of
work on a continuum. And we use the Posner’s affect emotion model within the
robot. It’s very, very simple to programme such a thing. We use coloured lights
in the shell of Miro in order to express the affect. You can slow down or speed
up the particular colour lights in order to display, alongside its basic behavioural movements, an emotion.
0:50 And when I say emotion, I of course mean “emotion,” because it’s
whatever the human user is perceiving. But if you imagine that the Miro robot
is flashing, quite quickly, a green colour, which is a calm colour, he’s sort of in
happy playful mode, whereas if you switch that to a red colour, he would be a
bit angry. We can use really nice, simple models from psychology in order to
build models of affect in robots, and that’s one of the great uses that psycholo-
gy has within robotics.
1:23 What we’re introducing into our environments is a new type of so-
cial agent with which a human being can have a relationship, which you can’t
influence to the extent you would your mobile phone, but which does have the
potential to influence you in being somewhere along the lines of being alive but
not being alive. As researchers start to explore that dynamic more, I’m hoping
there’ll be enough understanding that we’ll be prepared for whatever kind of
results will occur in the future on the basis of the introduction of such advanced
technology. Robots are going to be useful in educational settings, helping
teachers, particularly giving one-on-one attention to children, scaffolding their
learning of things like reading, writing, arithmetic, perhaps even more advanced subjects.
2:16 Robots right now are not very intelligent. We can’t expect a robot
to act like a teacher, to understand the world like a teacher, and to be able to
communicate. So what a robot can do is it can help a child that is trying to learn
by being a kind of co-learner and by giving encouragement. It can also help
with tasks that involve repetition and going over things like times tables. In
Соседние файлы в предмете [НЕСОРТИРОВАННОЕ]