The creators of the robot Twendy-One and the robotic arm Stair 1.0 have a competition in your area of research: Justin robot, with strong arms, precise and sensitive that they can perform complex tasks by interacting with humans.

Justin was presented during the ICRA-2007 by Professor Bruno Siciliano, University of Naples, Italy. The robot was designed by the University and the German Space Agency (German Space Agency, DLR) within a European project called “Physical Human-Robot Interaction: Unit and Security” (Physical Human-Robot Interaction: Dependability and Safety). Read the rest of this entry »

Although computer-brain metaphor has served well to cognitive psychology, research in cognitive neuro-science has revealed many important differences between brains and computers. Appreciate these differences could be crucial for understanding the neural mechanisms of information processing, and ultimately to create an artificial intelligence. Below,he discuss the most important of these differences (and implications for cognitive psychology, if you fail to recognize).

Difference # 1: The brains are similar, computers are digital

It is easy to think that neurons are essentially binary, as a potential trigger for action if they reach a certain threshold, which otherwise would not shoot. This superficial similarity to the digital 1 and 0 “hides a wide variety of continuous and non-linear processes that directly affect neural processing.

For example, one of the main mechanisms of transmission of the information seems to be the rate at which neurons transmit nerve impulse, an essentially continuous variable. Similarly, networks of neurons can fire in relative synchrony or in relative disarray, this coherence affects the strength of signals received by downstream neurons. In the end, within each of the neurons are in a strainer integrated circuit, comprising a variety of ion channels and membrane potential in constant fluctuation.

The inability to recognize these important subtleties may have contributed to the notorious error characterization of the perceptron Minksy & Papert, a nerve network without an intermediate layer between the reception and departure. In linear networks, any function computed by a network of three layers can also be computed by a network of two layers properly sorted. In other words, it can model accurately multiple combinations of linear functions by a single linear function. Because its simple two-layer networks could not solve many important problems, Minksy & Papert reasoned that these larger networks could not. Furthermore, calculations performed by networks closer to reality (eg, non-linear) are highly dependent on the number of layers, hence the “perceptron” hugely underestimate the power of computational neural networks.

Difference # 2: The brain uses a content addressable memory

In computers, accessing information in the memory required to seek its place in memory. This is known as byte-addressable memory. By contrast, the brain uses a content addressable memory, so that information can be accessed in the memory through an “activation fuzzy” from related concepts. For example, think of the word “fox” may automatically activate and diffuse the memories associated with other intelligent animals, with riders on horseback hunting foxes or attractive members of the opposite sex.

The end result is that your brain has a kind of “joined Google, where only a few tracks (keywords) are sufficient to cause a recovery of memory completely. Of course, you can do similar things on computers, mainly with the development of huge indices of stored data, which then also must be stored and searched to locate relevant information (incidentally, what Google is almost done, with some tricks).

Although this might seem a minor difference between computers and brains, has profound effects on the neural calculation. For example, a lengthy debate in cognitive psychology focused on whether the information was lost by a simple memory decay or by the interference of other information. In retrospect, this debate is partly based on the false assumption that these options are separated, as may happen with computers. Now many are realizing that this debate represents a false dichotomy.

Difference # 3: The brain is a massive parallel computer, computers are modular and serial

An unfortunate legacy of the brain-computer metaphor is the tendency for cognitive psychologists to find a modular feature in the brain. For example, the idea that computers require memory has led some to look for the “area of memory”, when in fact these differences are much more confusing. One consequence of this simplification is that only now are learning that the regions of “memory” (as the hippocampus) are also important for the imagination, the representation of new targets, space travel, and various and other functions.

Similarly, one could imagine that there is a “language module” in the brain, as it could have programs on computers with natural language processing. Cognitive psychologists even claimed to have found this module, based on patients with damage in a brain region known as Broca’s area. More recent evidence has shown that the language is computed by widely distributed neural circuits and general domain, and that the Broca’s area could also be involved in other calculations.

Difference # 4: The brain processing speed is not fixed, there is no system clock

The speed of neural processing of information is subject to a variety of limits, including time using electro-chemical signal across axons and dendrites, the axonal myelination, the transmission time of the neuro-transmitters through the cracks synaptic differences in synaptic efficiency, consistency shooting nervous, the current availability of neuro-transmitters, and the previous history of neural firing. Although there are individual differences in something called the psychometric “processing speed” does not reflect a unitary or monolithic concept, and certainly nothing as concrete as the speed of a microprocessor. In contrast, the “processing speed” probably put a psychometric index to a heterogeneous mix of all speed limits above.

In a similar vein, there seems to be no central clock in the brain, and there is a discussion about how a watch-like devices are actually holding time in the brain. To use just one example, is often believed that the cerebellum computes information involving a precise timing, as required for the delicate movements of an engine, however recent evidence suggests that the time in the brain has more resemblance to the waves in a pond with a digital clock running.

Difference # 5: The short-term memory is not like RAM

Although the apparent similarities between RAM memory and short-term or “working” emboldened many early cognitive psychologists, a closer examination reveals surprising and important differences. Although RAM and short-term memory appear to require energy (one shot nervous for the short-term memory, and power for the RAM), short-term memory appears to contain only “pointers” to the memory long term, while the RAM contains data that are isomorphic to those in the hard disk. (Click here for more about “point of care” in the short-term memory).

Unlike RAM, the capacity limit of short-term memory is not fixed, the capacity of short-term memory also appears to fluctuate with differences in “processing speed” (see # 4 difference) and with the experience and knowledge.

Difference # 6: Can not make any distinction between hardware and software with respect to the brain or mind

For years, it was tempting to imagine that the brain was a computer where the “mental program” or “mental software” is executed. This gave rise to a variety of abstract models-like programs, of cognition, where the details of how to run the brain actually executed those programs was considered irrelevant in the same way that a Java program can accomplish the same function as a C + + program.

Unfortunately, this attractive distinction between hardware and software obscures an important fact: the mind emerges directly from the brain, and changes of opinion are always accompanied by changes in the brain. Any abstract description of information processing will always need to specify how the architecture can implement those processes nervous, otherwise the cognitive models are excessively forced. Some blame this misunderstanding for the failure notorious “symbolic AI.”

Difference # 7: The synapses are much more complicated than the electrical logic gates

Another feature of the harmful brain-computer metaphor is that it seems to suggest that the brain can also operate based on electrical signals (action potentials) traveling along individual logical gates. Unfortunately, this is only half true. The signals that are propagated along axons are actually electro-chemical in nature, and it means traveling more slowly than the electrical signals on a computer, and can be modulated in countless ways. For example, the transmission of a signal depends not only on the so-called “logic gates” of synaptic architecture but also the presence of a variety of chemicals in the synaptic cleft, the relative distance between synapses and dendrites, and many other factors . This adds to the complexity of the processing that takes place at each synapse, and is therefore profoundly wrong to think that neurons function merely as transistors.

Difference # 8: Unlike computers, processing and memory are carried out by the same components in the brain

Computers process information from memory using a CPU, and then write the results of that process in memory. No such difference in the brain. As neurons process information, they are also modifying their synapses, which are themselves the place is a seat of memory. Therefore, the recovery of memory always slightly alters those memories. (Usually makes them stronger, but sometimes less accurate.

Difference # 9: The brain is a self-organized

This point is of course the point prior experience is deeply and directly to the nature of neural information processing in a way that simply does not happen in traditional microprocessors. For example, the brain is a circuit of auto-repair, something known as “a trauma-induced plasticity” is put into operation after an injury. This can lead to a variety of interesting changes, including some that appear to reveal an unused potential in the brain (known as acquired savant), and others that can result in profound cognitive dysfunction (as is unfortunately far more typical lesions traumatic brain and in developmental disorders).

In the field of neuro-psychology is a consequence of error in recognizing this difference, which examines the cognitive performance of patients with brain injury to determine the function of the computational region damaged. Unfortunately, a poor understanding of the nature of the plasticity-induced trauma, the logic can not be so simple. Similar problems underlie the work on developmental disorders and the new field of “cognitive genetics”, where the consequences of self-organization nervosa are often ignored.

Difference # 10: The brains have bodies

This is not as trivial as it might seem, is that the brain takes surprising advantage of the fact that a body has at its disposal. For example, despite his instinctive feeling that it could close my eyes and know the location of objects around them, a series of experiments in the field of blindness has shown that our visual memory is actually very low. In this case, the brain “dump” its memory requirements to the environment in which it exists: why bother remembering the location of objects when a glance is enough? A surprising set of experiments by Jeremy Wolfe has shown that even after hundreds of times to ask what are simple geometric shapes on a computer screen, subjects continue to respond to these questions by the hearing and no memory. A wide variety of evidence suggests that other domains are just beginning to understand the importance of the body in processing information.

Bonus difference: The brain is much, much bigger than any computer

The precise biological models of the brain would have to include about 225,000,000,000,000,000 (225 trillion) of interactions between cell types, neuro-transmitters, neuro-modulators, axonal branches and dendritic spines, and that does not include the influence of geometry dendritic, or about 1 trillion glial cells which may or may not be important for neural information processing. Because the brain is non-linear, and because it is much bigger than all current computers, it seems likely to work is quite different. The brain-computer metaphor obscures this important, though perhaps obvious, difference in computational power.

“Robot Suit HAL” is a cyborg-type robot that can expand and improve physical capability.

The company Cyberdyne has developed a robotic suit, which allows to extend the power of a human being. Initially designed for people with physical disabilities, support for heavy work in the factories, and rescue support to the disaster sites, as well as in the entertainment industry. Its appearance resembles the costumes shown in the drawings Japanese manga.

When a person tries to move, nerve signals are sent from the brain to the muscles through the motor, moving the musculoskeletal system as a result. At this moment, very weak biosenales can be detected on the surface of the skin. HAL catches these signals through a sensor attached to the skin of the signals obtained in usuario. Basado, the power unit is controlled to move all the points together with the generator of muscle movement, enabling the carrier support in everyday activities. This is what we call a ‘voluntary control system’ that provides movement to interpret the intent of the bearer from biosenales, before making the move. Not just a ‘voluntary control system’ HAL has, but also an “autonomous robotic control system that provides human movements on a robotic system that works entirely with the” system of self-control. ” HAL is the first cyborg-type robot controlled by this unique Hybrid System.

HAL is expected to be applied in various fields such as rehabilitation support and physical training support in the medical field, supporting people with disabilities, support for heavy work in the factories, and rescue support disaster sites, as well as in the entertainment industry.

Automation Systems can be complex, with many solutions available to keep them running reliably.

Talk about reliability in manufacturing and the word that probably springs to mind is “motor.” Certainly, motors and other rotating equipment must be kept moving. No work is done unless a motor turns. Not surprisingly, suppliers have invested in technology to improve the reliability of these workhorses of manufacturing. Other parts of the automation system have gone under the reliability microscope, as well. As automation becomes more software intensive, it is imperative that engineers pay attention to reducing computer down-time. From sensors to software, engineers have used their ingenuity and technology to create a dependable automation system.

Read the rest of this entry »

The next annual conference of Computer-Human Interaction, held in Italy in the master-sponsored Microsoft, despite its relatively modest scale, has attracted unusual attention to the IT-Press. Perhaps the point is that in recent years just two long-developing conceptual «interfeysostroitelnye» trends dorosli to the stage of commercialization. And not just dorosli, but also implemented to ensure the success of their companies in the markets involved srazhayuschimisya fiercely competitive. Therefore, the motto of CHI’08 - «Balance of Technology and the Arts» - it would be logical complement mention more about business. It is, of course, on multisensornyh panels (multi-touch-pad), capable of simultaneously registering more than one touch, as well as input devices based on accelerometers (examples of which can serve as Nintendo Wii consoles and mouse Logitech MAX Air).

Indeed, multisensornye Ends finally ceased to be futuristic design concept and moved on the offensive in all segments of the price range. For example, in March showed the public subnoutbuk ASUS EeePC 900 at a cost of about $ 600 (spot with an impressive demonstration of the functions of his multisensornoy panel with the popular Windows-applications can be viewed here). And in April, started long ago promised a campaign to equip shopping centers AT & T in major U.S. cities terminals Microsoft Surface, providing visitors some intriguing possibilities in the field of information services. For example, putting on the screen of the terminal a couple of mobile phones, the visitor can get a comparative table of their characteristics.

Recall that the current price of these devices - ranging from 5 to 10 thousand U.S. But those who think it is excessive, can wait commercialization recently demonstrated the product, called Microsoft Touch Wall: its value, according to assurances from the developers of «several hundred dollars». Or, wait for the success of the project CUBIT multi-touch, a result which is expected to become simpler kind of Surface valued at $ 500-1000. Another significant public project to multisensornymi interfaces, called MPX (this is - an abbreviation of Multi-Pointer X Window Server), is designed to make it possible to use existing X Window-compatible applications on UNIX-like OS in several simultaneous manipulation of the cursor.

Feature CHI’08 was the installation of Natural Interaction Sensitive Table from the Italian company iO (ioagency.com), which is a sort of «Surface on steroids»: it is equipped with an array of microphones with software variable chart sensitivity, RFID-antennas, the function of speech recognition and other unique qualities.

It should also mention the recent presentation of the next OS from Microsoft (informally known as Windows 7), held under the shadow of all-out support multisensornogo interface, and plans to equip these screens 75-dollar laptop XO2, announced the project OLPC (One Laptop Per Child) .

In quantitative terms, the most popular devices multisensornymi panels in the near future is destined to become a mobile phone. According to analysts from iSuppli, during the 2006-2012 biennium. Global volume of their sales to increase from less than 200 thousand to 21 million units.

As for the «accelerometric» controls game consoles, the volume of the market is difficult to assess, because for them time and again offered the unexpected new market niches. Thus, in the U.S. in the next few months, developed an informal campaign for the introduction of Wii as a means of ridding the «white-collar» of hypodynamy (formal, it apparently will be after the economic feasibility of this method of recovery office employees will be demonstrated statistically). A recently introduced several electronic media coverage of the growing popularity of Wii … nursing homes!

The proliferation of devices based on accelerometers and multisensornyh panels will increase the demand for graphical interfaces, focusing on gesture and the development of «muscle memory», which the potential for improving efficiency of man-machine interaction are poorly implemented. In particular, multisensornye interfaces quite naturally fall menu in the form of pie charts ( «pie menus»): index finger can be used to call them in random spot of the screen, a large or medium - to select the desired item. It is expected that the passion for fashion «multi-touch-style» soon transforms usual, we kind of the taskbar and contextual menus.

All possible contribution to the popularization of Surface-like interfaces made Bork, used in concert speakers and musical instruments Reakteybl Lemur. The first - vyglyadyaschee prisheltsem from remote parallel worlds offspring of the Barcelona Institute of audio-visual technology. Objects on its surface (the so-called «tangibles») symbolize components zvukosinteziruyuschego tract. The strength of the second - the possibility of modeling the physical parameters of real devices (such as inertia polzunkovyh regulators in the Mixer)

As arranged multisensornye panel?

Like many other technology in the early stages of implementation, admire Ends modern multisensornye rampage original design. Mention a few.

* The Microsoft Surface and Perspecta of fingers is determined by the user through the analysis of the images coming from multiple cameras. In the latter case it is invited to wear special rings, markers, that is not very convenient. But thanks to his software never sputaet, say, the middle finger with the index.
* In the much less demanding (and less accurate) systems, Microsoft Touchwall area space adjacent to the plane of the image is scanned using three infrared LEDs.
* The spherical surface Touchglobe introduced an array of microscopic piezoelectric sensors. The advantage of this solution - the opportunity to register because each of the touching, and weakness - opacity of the touch panel.
* In multisensornyh displays of mobile phones are now used projection-resistive and capacitive projection technology. The latter is the higher cost and sensitivity, and even the ability to register outrage electrostatic field caused by a finger, podnesennym to the screen at a distance of about 2 mm.
* Jefferson Han (Jefferson Han) from New York University patented and used in its development of so-called full effect breached internal reflection (Frustrated Total Internal Reflection, or FTIR), with 1960’s used in biometric applications (for fingerprinting) . He proposed to place a screen in front of the glass panel (not necessarily flat), an internal volume of which is saturated with many pereotrazhaemym radiation infrared LEDs located on its perimeter. Touching the panel violated the normal course of radiation, which is possible to register using hidden video cameras for the screen. In 2006, to commercialize the design, Mr. Khan had organized a company Perceptive Pixel (perceptivepixel.com), sold in multisensornye now displays a unique quality for the price of some 100 thousand U.S.

What awaits us in 25 years?
Microsoft Surface - a computer with multisensornym interface, known to the public at large Hollywood fantasy militants «Opinion»

In conclusion pofantazirovat on a hypothetical 25-year cycle of development of interface concepts: the mouse, invented in 1963, has been widely disseminated not until mid-1980’s, and multisensornye panel, first established in 1982 at the University of Toronto, are massive demand just now. The idea of input device, built on the basis of accelerometers, also has approximately 25 years. If that pattern, rather than random coincidence, it should not assume that the perception of any ordinary radically new way of interacting with technology requires a change of generations of its users? Indeed, while the revolutionary changes associated with some of the emerging before our eyes interface technology, will occur only after a quarter century and will affect not only us but our descendants. As it will for change?

Today the notion of «Ends future» is largely associated with the means exposure to remain outside bodies smell and touch, as well as taste receptors. Anyway, Japanese experts already foresee the emergence of risovarok, allowing users to receive understanding of the taste of rice in its preparation.

As the theory II such traditional means of input devices, such as a keyboard or voice recognition system will be used to obtain information on the current emotional state of the user. This, in particular, will remove «tactlessness» voice communication systems, embedded board computers in modern cars: easy to understand that ignoring the tense state of the driver, busy with the difficult road situations is fraught with catastrophic consequences.

Earlier this year, Book Toyota announced that the new firm automatic collision avoidance will make decisions in light of the fact whether closed at a critical juncture in the eyes of the driver. Note funny (but possibly more profound than it seems at first glance) similarity of the draft with one of the exhibits CHI’08 - system Spoken Words, developed at the universities of Berlin and Potsdam. Vsmotrevshis in the user’s facial expression (through the lens regular webcam) and found both his eyes closed, it initiates reading aloud edited in MS Word current proposal.

But, more important today presented a study of interface-based systems encephalography and other techniques to study physiology of the brain and nervous system of rights.
Multisensorny 3D-monitor Perspecta company Actuality Medical (actuality-medical.com) also tends to Sphericity

In 2004, serious science and technology magazine IEEE Spectrum published a story Futurological Vernon Vindzha «Integrated intuition» (Vernon Vinge, «Synthetic serendipity»; see translated a collection of «Highlights for the year»: shop.bambook.com / scripts / pos.showitem ? v = 1 & ite = 81013). Computer keyboard its characters seen as hopeless anachronism, preferring to communicate through clothing, equipped with sensors to record the nerve impulses that reflect the intentions of the user (for example, you press a button image). Just in the same year, these science-fiction prophecy began to come true: researchers from NASA have shown a system of speech synthesis, managed by the so-called subvokalnymi nerve impulses, read from certain areas of the skin on the neck, and allowing express what you would like to say. Dumb this system is able to return the gift of speech and spetsnazovtsam provide an opportunity to talk in the crowd without the risk of being podslushannymi.

In contrast, non-invasive methods for NASA deep stimulation technology developed at Bristol University, suggests the introduction of ultra electrodes (the thickness of a human hair) directly into the brain of the patient. We show that Modulated computer electrical signals can normalize the status of parts of the brain affected in some forms of depression, which in turn leads to improved mood patient or, if it so desires, even to a state of euphoria. The number suffering from severe depression only in the UK is estimated at 50 thousand, and some experts have already foresee the emergence of social problems associated with the new «e» forms of drug addiction.
It seems that XO2 announcement marks the emergence of a new class of notebooks, which are folding multisensorny display, one of the halves, if necessary, simulates the keyboard. Note that the appearance of the latter may be determined by the specificity of the active application (the benefits of such a decision systematically presents the most popular site artlebedev.com / portfolio / optimus). Adjustment of mass release of the nice hybrid mobile computer and electronic books promised in 2010

A neurophysiology from Duke University (North Carolina) this year, managed to force the move on track anthropomorphic robot. The source of signals to control his lower limbs was the brains of live monkeys, more precisely electric potentials of approximately two hundred neurons responsible for motor activity of the animal. This achievement opens up unprecedented prospects for the rehabilitation of paralyzed patients.

Note that the people affected by paralysis and depression would require implantation of electrodes, working as a «Enter», and at the «withdrawal». Perhaps, 25 years after the current decade will be perceived IT-history as a time when overcrowding symbiotic coexistence with the IT prepared for delivery human interface technology last frontier related to the integrity of your own brain for mechanical intervention.