Multispectral imagers collect a hypercube of data, where the spatial image is along two-dimensions and the spectral information is in the third. Two main technologies are used for multispectral imaging: sweeping, where the hypercube is built by scanning through different wavelengths or spatial positions and snapshot multispectral spectral imaging, where the 3D cube of images is taken in one shot. Sweeping imaging systems tend to have more lines and better spectral resolutions whilst snapshot cameras are often used for dynamic analysis of scenes. A common method to obtain the hypercube in snapshot imagers is by pixel level filtering on the sensor chip. Pixel level filtering, where the filter is placed directly on the pixels are intergrated into the wafer-level making processing making them difficult to customize. Therefore, these sensors tend to aim for equally spaced spectral lines in order to cover many applications. This results in an often in an unnecessarily large data cube when only a few spectral lines are needed, moreover the spectral lines are not adapted to the specific application. In this work we propose a multispectral camera based on plenoptic imaging, where the filtering is done in a front-end optics module. Our camera has the usual advantages of a snapshot imager, and the added advantage that the spectral lines can be both reduced and tailored to the specific application by customizing the filter. This procedure reduces the hypercube whilst keeping performance by selecting the relevant data. Moreover, the filter is interchangeable for different applications The camera presented here is built with off-the-shelf components, shows >40 spectral channels, image sizes are 260x260 pixels, with pixel limited spatial resolution. We demonstrate this technology by fruit quality control using machine learning algorithms.

We present an optical system which integrates a plasmonic sensing surface and an angular tracking system to enable a compact refractive index measurement. A refractive index change at the surface of the sensing membrane causes a change in the angle at which monochromatic light is transmitted through the membrane. This transmission angle is measured by the angular tracking system. We show good theoretical and experimental agreement of the transmission of the plasmonic sensing surface at different angular illumination of the membranes. Using this compact optical setup the embedded angular tracking system has an accuracy of <10-4 deg. This corresponds to a sensitivity <10-5 refractive index units. Finally we demonstrate this measurement technique using different concentrations of saline solution.

The miniaturization of photodetectors often comes at the expense of a smaller photosensitive area. This can reduce the signal and thus limit the image quality. One way to overcome this limitation is to reduce the photosensitive area but with no reduction of signal i.e. harvest the light. Here we investigate, theoretically and experimentally, light harvesting with nanostructured metals. Nanostructured metals can also give additional functionality such as polarization filtering which is also investigated. After defining the figure of merits used when characterizing light harvesting and polarization filtering structures, we detail the fabrication and measurement process. Structures were made on glass substrate, as a post process step on CMOS fabricated detectors and directly in the CMOS fabrication of the detectors. The optical characterization results are presented and compared with theory. Finally, we discuss the challenges and advantages of integrating metallic nanostructures within the CMOS process.

A universal software framework for hierarchical object recognition has been devised based on V. B. Mountcastle's observation (Mountcastle, 1978) that the human cortex consists of the same basic functionality which is used to subdivide the complex computations into elementary matching tasks, independently whether auditory, visual, olfactory, haptic or any other sensory information is presented at the input. Combined with a unique vision sensor system that is capable to directly extract contrast magnitude and direction, a powerful combination of hardware and software for real-time pattern-recognition tasks at very low power-consumption levels is presented

This paper describes a sensor interface for metal-oxide chemical gas sensor for pollution detection. The function of the ASIC is to control the sensor working temperature by applying a programmable voltage with 10 bit resolution, to measure the resistance of the sensitive elements ranging from 5 k/spl Omega/ to 100 M/spl Omega/, measure the ambient temperature with an external NTC thermistor and offer a fully digital user interface. It gives the possibility to make low power and low cost high performance gas sensing microsystems for consumer application.

This application-specific integrated circuit (ASIC) computes two optically encoded positions, each engraved in a pattern onto one of a two-cylinder assembly. The cylinders are coupled by a torsion bar of known stiffness so that the difference between the two extracted positions provides a torque measurement to be used in next-generation automobile electrical power assisted steering (EPAS) systems. The cylinder positions are defined by 11b absolute-position encoders plus 7 additional bits of interpolated (relative) position between adjacent codes. The circuit is controlled and accessed via a serial peripheral interface (SPI).

The optical character-recognition (OCR) system described consists of a CMOS retina, an analog classifier IC and a microcontroller. The retina converts the parallel optical input into an oriented edge representation which is processed further and recognized by the classifier in real-time. The microcontroller postprocesses the time sequence of classifier outputs and provides timing and control. The system can be used as a handheld OCR scanner which processes the field of view 1000 times per second. The system output is the character string being scanned.

This short paper explains how the Khepera robot was developed, from the initial idea to the its commercialisation by K-Team. The goal of this paper is not a scientific analysis but an historical overview of the steps made in the development of this robot since 1991. The papers introduces first the situation of the team who started the development, then decisions made in creating the actual Khepera are briefly described, as well as some important steps in the commercialisation of the robot. We conclude with the current status of Khepera, and introduce other products that have evolved from Khepera.

This short paper explains how the Khepera robot was developed, from the initial idea to the its commercialisation by K-Team. The goal of this paper is not a scientific analysis but an historical overview of the steps made in the development of this robot since 1991. The papers introduces first the situation of the team who started the development, then decisions made in creating the actual Khepera are briefly described, as well as some important steps in the commercialisation of the robot. We conclude with the current status of Khepera, and introduce other products that have evolved from Khepera.

Keywords: [MOBOTS] ; Khepera robot Reference EPFL-CONF-200915doi:10.1109/ROBOT.1995.525757 Record created on 2014-08-20, modified on 2017-05-10

To get a better understanding of the application of artiicial neural networks to the robotics eld, the rst part of this 3-year project supported by the NFP-PNR23 project was devoted to the study of some applications of neural networks to motor control , sensor fusion and image processing and to the development of sensor interfaces and neural network dedicated accelerators. The second part of the project, described in this paper, takes advantage from the experience gained in the rst part and studies the application of ar-tiicial neural networks to a particular sub-eld of robotics: mobile robotics. Most of the problems of the domain seem to be particularly adapted to neu-ral network solutions. To illustrate this, we present in the rst part of the article an example implementation of a very simple behaviour in a real robot using several design methodologies. The robot used for these experiments is described. In the second part of the document we present some more complex experiments based on a distinctive capability of neural networks, that is the ability to learn from examples. In the last two sections we present the beginning and the direction of our future work: The goal of this project is to study the neural networks approach applied to a group of mobile robots. In section 5. a rst experiment involving ve robots is presented, while the following section outlines the future strategy .