Nadav Kahanovich

Research

Feb 24th, 2021
Can robots understand what we want?

Dr. Avishai Sintov has created a wearable device that can detect the movements of the forearm muscles and communicate the intentions of human activities which a robot will sense and reciprocate.  

  • Research
  • Mechanical Engineering

Human- robot collaboration requires intuitive understanding of human movements. Since our early childhood development, we learn that the different kinds of movements we perform in a certain way, will generate a certain response, more often than not, without any verbal communication. For example, if you offer a cup of coffee to your friend, they will intuitively reach their hand out to take the cup. In fact, by just looking at the cup in your hands, your friend will understand the intention and predict the trajectory of your movements, acting in response.

 

Publication

The research findings of Dr. Avishai Sintov and graduate student, Nadav Kahanovich, will be published in the The IEEE Robotics & Automation Letters.  

 

Dr. Avishai Sintov is the Head of ROB-TAU lab lab at the Mechanical Engineering at Tel Aviv University. Dr. Sintov and his team of researchers study the prediction of human intentions in non-verbal communication, and aim to “teach” this to robots..

 

In their article, the researchers showed that a simple and inexpensive wearable device can detect different motions of human forearm muscles. The device contains 15 power sensors that stick to the skin and detect muscles contractions when a person takes various objects and moves a hand. Using their algorithms, the robot receives information about the objects in real time which imply about the intention of the person. The algorithms include a neural network that is trained to read and interpret the information from the device and accurately predict the intention of the interaction in real time.

 

The ability to predict the intentions of human activities helps the robot plan the movement for efficient and quick assistance. This approach will allow robots to work intuitively without verbal communication and cameras.  

Photo: The robot identifies the objects and assists the person

 

Future Vision

In the future, this technology can assist people with disabilities, for example a wheelchair-mounted robot arm aid to assist with daily activities. It can be also applied in the factories and hospitals. Can you imagine a robot replacing a nurse and assisting a surgeon during medical procedures? We applaud our engineers and students for their fantastic discovery and research which will only enhance our world.

 

The research is funded by the Israeli Science Foundation. 

 

Loving the problem is the greatest way to invent

Research

Jul 16th, 2015
New Cyber Security study aims to identify malicious websites through their

Web Mining utilizes data mining techniques to discover and automatically learn information on the Internet (www). One of the most challenging applications of Web Mining is identifying the type of a specific website based on the category to which it belongs, and in particular identifying malicious websites. Malicious websites are websites created with the intent of harming users, stealing information or conducting other undesirable activities when the user enters the website, and afterwards.

 

In recent years, many methods have been developed for identifying website categories. Some are based on analyzing the website's textual content, its users' navigating profiles??, suspect traits found in the website itself, IP forgeries and many other properties. At the same time, the already enormous amount of information existing on the Internet continues to rise rapidly and exponentially, making the identification task ever more complex – often requiring a great deal of expensive resources.

 

A study carried out by Doron Cohen, under the supervision of Prof. Irad Ben-Gal and Prof. Shulamith Kreitler, tested a new method for identifying websites, based on an analysis of their graphic design, via methods of data mining.  To this end, an algorithm was constructed, which receives a URL as input, and produces an output by retrieving and processing all graphic design features and storing them as tables in the server. The researchers examined hundreds of home pages of websites taken from Google's top 1000 sites. For each of these, over 1000 graphic design features were examined, such as: size of area covered by each color, font size, number of characters, standard deviations, quantity and type of elements etc.

 

After processing and analyzing the data, the researchers built a predictive model based on a decision tree, and then cross-validated the model. In the first experiment they discovered that classification based solely on graphic design enables relatively high prediction of all five examined website categories (including that of malicious websites). Another experiment found that adding graphic design features to another objective prediction method can improve the precision of identification – specifically of malicious websites - by 95%, and in a statistically significant manner, using low-cost resources and low runtimes. A possible explanation for these findings is that malicious websites apparently try to conceal keywords to avoid detection by search engines, while a search based on so many graphic features can identify repeated patterns which are more difficult to conceal.

 

The study revealed that graphic design, especially colors, plays an important part in the prediction of website categories. Adding graphic properties to other predictive systems can improve their accuracy, and is therefore highly recommended. The study will be presented at the national cyber conference scheduled to take place in September at Tel Aviv University.

Research

May 18th, 2015
New Optical Materials Break Digital Connectivity Barriers

From computers, tablets, and smartphones to cars, homes, and public transportation, our world is more digitally connected every day. The technology required to support the exchange of massive quantities of data is critical. That's why scientists and engineers are intent on developing faster computing units capable of supporting much larger amounts of data transfer and data processing.

A new study published in Nature Photonics by Tel Aviv University researchers finds that new optical materials could serve as the nuts and bolts of future ultra-high-speed optical computing units. According to the research, led by Dr. Tal Ellenbogen and conducted by group members Nadav Segal, Shay Keren-Zur, and Netta Hendler, all of the Department of Physical Electronics at TAU'sSchool of Electrical Engineering and TAU's Center for Nanoscience and Nanotechnology, these "nonlinear metamaterials," which possess physical capabilities not found in nature, may be the building blocks that allow major companies like IBM and Intel to move from electronic to optical computing.

At his TAU lab, Dr. Ellenbogen studies the interaction between light and matter at the nanoscale level in order to explore underlying physical mechanisms, which can be used to develop novel optical and electro-optical components.  

Read more:

(1) From American Friends of Tel-Aviv University: https://www.aftau.org/news-page-computers--technology?=&storyid4702=2176&ncs4702=3

(2) N. Segal, S. Keren-Zur, N. Hendler, and T. Ellenbogen, "Controlling light with metamaterial-based nonlinear photonic crystals," Nat. Photonics 9, 180–184 (2015).

Research

May 18th, 2015
A journey from conceptual research to product

Funded by Momentum Fund, a revolution of the compact camera device is developed in the Faculty of Engineering. Few years ago, Professor David Mendlovic with Ariel Raz developed a new concept for optical imaging. Although Color imaging is a standard, significant penalties are connected with achieving the color information such as reduced resolution and poor low light performance. In Smart Imaging System (SIS) project, Mendlovic and Raz take the new concept few steps further to the market. A mixed team that includes students and engineers investigates the various practical aspects of the new technology. The vision is to own a collection of intellectual properties that enables commercialization of the new technology. With business development support from RAMOT, discussions with partners and potential customers helped to refine the product specifications. As anon conventional action, the project will be presented in the coming Mobile World Congress including demo system and impressive list of meetings with customers.

 

The trigger of this project is the present capability of image sensors to capture many frames per second.

Common commercial grade cameras operate by acquiring all color components at once, on the expense of resolution – each color is sampled at a pre-determined fraction of the overall resolution. This method, invented by Kodak in 1976, degrades the overall image resolution by a typical factor of 4. Recently available frame rates, however, allow dividing the color information across several consecutive frames.  In essence, each frame is taken at a different color and at full resolution. Then, each set of frames is converted into a full resolution color image. As each frame is taken at full resolution, the SIS camera exhibits significant resolution improvement compared to present cameras. In addition, we have experimentally proved that acquiring additional frames at non-standard colors, such as white, clearly improve the overall image quality. This is done by removing image noise from the red, green and blue frames using additional information obtained from the white frame. Moreover, our system supports acquisition of near IR frames (color shades that are invisible to the naked eye), thereby improving the resulted camera sensitivity. In parallel, good image quality requires fast and accurate color filter. Current demo system utilizes off the shelf components but we have already completed a custom design, optimized for this task.

 
Nonlinear converter - narrower than you would expect

Research

May 18th, 2015
Nonlinear converter - narrower than you would expect

Nonlinear crystals convert laser light of one wavelength to another. Usually, the nonlinear crystal also converts efficiently other wavelengths that are close enough to the designed wavelength. When the input laser has multiple wavelengths and we are interested only in the specific designed wavelength, this wide response must be narrowed so that the unwanted wavelengths would be filtered out.

However, in order to achieve a narrow conversion response, a long crystal is needed, since the width of the efficiency response (with respect to wavelength) is usually inversely proportional to the crystal's length. Unfortunately, longer crystals are more expensive, consume a larger physical size and moreover, the maximum length of most crystals is limited to a few centimeters at most.

In our research, we show that by appropriate modulation of the nonlinear crystal, the width of the converter can be made as narrow as we like, irrespective of the crystal's length. This is achieved by modulating the nonlinear coefficient with a "super-oscillating" function. Super oscillating functions are band-limited functions that oscillate locally much faster than their highest Fourier component. This work therefore brings the concept of super oscillations to the nonlinear optics regime for the first time. This high resolution nonlinear filter can have applications in optical communication, spectroscopy, quantum information, and more.

Roei Remez and Ady Arie, “Super-narrow frequency converter”, Optica 2, 472-475 (2015), https://www.osapublishing.org/optica/abstract.cfm?uri=optica-2-5-472.

Research

May 18th, 2015
Experimental and Theoretical Studies of Soliton-Soliton Collisions

Solitons (solitary waves) are self-trapped wave packets that keep their localized shape in the course of long-distance propagation, due to stable balance between linear spreading out, caused by diffraction and/or group-velocity dispersion, and nonlinear self-focusing. Solitons are fundamental collective-excitation modes in diverse areas, among which two most rapidly developing ones are nonlinear photonics and matter-wave optics (the latter pertains to coherent modes in  atomic Bose-Einstein condensates (BEC)). Once the existence of individual solitons has been established, the next issue is collisions between them. Very often, solitons interact elastically, i.e., they pass through one another and emerge from the collision with original shapes, amplitudes, and velocities, the actual result of the collision being a shift of both solitons forward or backward, which seems as a discontinuous jump. This remarkable property is closely related to the fact that, in the simplest but relevant approximation, the soliton dynamics obeys integrable one-dimensional (1D) equations, such as the celebrated nonlinear Schrödinger equation (NLS), which emerges in the same form in a slew of different physical contexts. In particular, the NLS equation is the most fundamental model of both nonlinear photonics and atomic-wave optics. The integrability of the NLS equation implies that its solutions indeed produce elastic collisions between solitons.

The simple 1D approximation holds if the wave field is strongly confined in two transverse directions. Otherwise, in the three-dimensional (3D) setting, the integrability breaks down, the most dramatic consequence of that being a possibility of the collapse, i.e., catastrophic self-compression of the field. Accordingly, soliton-soliton collisions may exhibit complex outcomes, including the onset of collapse, merger of the colliding solitons into a single one, or the quasi-elastic rebound. The BEC in the gas of 7Li atoms makes it possible to realize both quasi-1D and fully 3D regimes, a crucially important experimental technique being the Feshbach resonance (FR), which makes it possible to control the strength of the self-attractive nonlinearity by means of an external magnetic field. While solitons were created in this setup more than 10 years ago [1,2], experimental and theoretical investigation of collisions between them was reported only very recently by Jason Nguyen, Paul Dyke, De Luo, Boris Malomed, and Randall Hulet, "Collisions of matter-wave solitons",

Nature Physics, DOI 10.1038/NPHYS3135, http://www.nature.com/nphys/journal/vaop/ncurrent/full/nphys3135.html (the work was supported by the Binational Science Foundation through grant No. 20110239). Using real-time in-situ imaging, it has been shown that the collision is a complex event, the outcome of which markedly depends on the relative phase between the colliding solitons. Adjusting the strength of the nonlinearity by means of the FR, the collisions have been examined in the proximity of the 1D - 3D crossover. All the above-mentioned outcomes were directly observed in the experiment, and explained theoretically by means of direct simulations and analytical consideration, which treats solitons as quasi-particles: the collapse, merger, and repeated rebounds (in the latter case, the solitons collide repeatedly as they are confined by a weak harmonic-oscillator trapping potential. These outcomes are shown, respectively, in images displayed in panels (a), (b), and (c) of the figure.

This work has quickly drawn attention of the major professional network, PhysicsWorld: http://physicsworld.com/cws/article/news/2014/nov/06/colliding-bose-einstein-condensates-vanish-from-sight.

[1] L. Khaykovich, F. Schreck, G. Ferrari, T. Bourdel, J. Cubizolles, L. D. Carr, Y. Castin, and C. Salomon, Formation of a matter-wave bright soliton, Science 296, 1290-1293 (2002).

[2] K. E. Strecker, G. B. Partridge, A. G. Truscott, and R. G. Hulet, Formation and propagation of matter-wave soliton trains, Nature 417, 150-153 (2002).

Research

May 18th, 2015
Cryptography with Low Complexity

The efficiency of cryptographic constructions is a fundamental question. Theoretically, it is important to understand how much computational resources are needed to guarantee strong notions of security. Practically, highly efficient schemes are always desirable for real-world applications. More generally, the possibility of cryptography with low complexity has wide applications for problems in computational complexity, combinatorial optimization, and computational learning theory.

 

In this project we aim to understand what are the minimal computational resources needed to perform different cryptographic tasks. In a nutshell, we suggest to focus on three main objectives. First, we would like to get better understanding of the cryptographic hardness of random local functions. Second, to harness our insights into the hardness of local functions to derive applications in cryptography and computational complexity, with an eye towards improving the efficiency of basic cryptographic building blocks such as pseudorandom functions. Third, to expand our theoretical understanding of garbled circuits, study their limitations, and improve their efficiency.

Research

Mar 10th, 2015
Modeling of novel metal-oxide materials for photovoltaic applications

Photovoltaic (PV) cells are an efficient mean to convert solar energy into electricity. PV is one of the fastest growing renewable energy sources and shows great promise in countries with enough sun light and in isolated villages. While leading technology is currently silicon based, there is an ongoing effort to find new materials for PV cells. In this research we combine efforts of several experimental and theoretical groups in Israel to do a systematic characterization of a very large set of novel solid state alloys for PV applications. We would initially do theoretical calculations for selected choice of materials and compare our theoretical predictions for the electronic, optical and structural properties to experimental data. At later stages we would initiate high throughput calculations strategies to theoretically predict novel materials (new alloys compositions and new geometrical structures), those predictions are going to be tested by experiment. The experimental setup includes also the formation of nano-crystals and thin films – such structures have properties that are substantially different than the bulk and so an effort would be dedicated to describe also such structures and their possible interfacial effects.

 

The theoretical and experimental data is going be accessible through a database of properties that would be available to all researchers in the project and later to a wider audience. The main aim of the project is to find new alloys that would allow for cheaper and efficient future PV cells. The project is going to create also a large knowledge base that would have a huge value by itself.       
Tel Aviv University makes every effort to respect copyright. If you own copyright to the content contained
here and / or the use of such content is in your opinion infringing, Contact us as soon as possible >>
Tel Aviv University, P.O. Box 39040, Tel Aviv 6997801, Israel