Abstract:
The invention concerns a robust communication system (1) for transmissions through a noisy environment, which comprises a cascade of a discrete signal source (2), a chaotic modulator (3) for the signal, a noisy transmission channel (4), and an incoherent discriminator or receiver (5). Advantageously in a preferred embodiment, the incoherent discriminator (5) comprises a high-pass filter (6) effective to remove the lowest frequency harmonics of the received signal, a rectifier (7) providing the absolute value of the wave, a low-pass filter (8) carrying out a mean of the rectified wave, and a comparator (9) placed after the low-pass filter.
Abstract:
The invention concerns a new method of controlling the movements of a multi-actuator electromechanical system which has a matrix of locally interconnected analog cells associated therewith. Each cell represents a hardware implementation of a model of fuzzy inference rules. The model is constructed as a fuzzy circuit architecture which is implemented, preferably, in the form of an integrated circuit with VLSI CMOS technology which generates and controls the reaction diffusion mechanism typical of auto-waves by means of a fuzzy neural network. This fuzzy neural network defines the functional relationships which can duplicate simultaneous reaction diffusion equations, and the duplication of such simultaneous equations is provided by two sets of fuzzy rules processing, in a linguistic manner, the state variables of the cell plurality and imposing on each cell a dynamics of the oscillatory type wherein two dynamic processes having different kynetic characteristics coexist.
Abstract:
A step-up continuous-mode DC-to-DC converter with integrated current control, comprising means (3) for comparing a voltage signal (Vout) in output from said converter and a reference signal (Vref) for generating an error signal (Ve) and means (7) for generating a compensation ramp which are suitable to generate a ramp signal which is added to a signal which is proportional to a current ramp that flows across the converter, the signal in output from the comparator means and the signal obtained from the sum being sent to additional comparator means (6), the output whereof, together with an oscillator signal (8), is sent to locking means (9) for driving a power transistor (20) of the converter, characterized in that fuzzy logic control means (11) are interposed between the comparator means (3) and the additional comparator means (6), the fuzzy logic control means (11) receiving in input the error signal output by the comparator means (3) and emitting in output a signal (Ve) which depends on the variation of the error signal over time to be sent to the additional comparator means (6).
Abstract:
The sensor (1) comprises an array of photo sensitive elements for acquiring images of the passenger compartment in a motor vehicle and a circuit (20) for processing the signals corresponding to said images generated by said photo sensitive elements. The processing circuit is configured according to a cellular neural network (CNN) processing architecture of said image signals and can generate an output signal indicating the decision on whether to deploy an airbag (3) to which the sensor is associated or to control the explosion of the airbag. Preferably, the photo sensitive array and the processing circuit are comprised on a single integrated component, preferably implementing CMOS technology.
Abstract:
Described herein are a molecular memory obtained using DNA strand molecular switches and carbon nanotubes, and a manufacturing method thereof. In particular, the nonvolatile memory is manufactured according to an architecture that envisages the use of carbon nanotubes as electrical connectors and DNA strands as physical means on which to write the information. In other words, the nonvolatile memory is made by means of a set of molecular DNA strand switches, the addressing of which is controlled by molecular wires made up of carbon nanotubes.
Abstract:
Monolithically integrated pressure sensors of outstanding quality and versatility are produced through micromechanical surface structures definition techniques. A microphonic cavity in the semiconductor substrate is monolithically formed by
cutting by plasma etching the front side or the back side of the silicon wafer a plurality of trenches or holes deep enough to extend for at least part of its thickness into a purposely made doped buried layer of opposite type of conductivity of the substrate and of the epitaxial layer grown over it; electrochemically etching through such trenches, the silicon of the buried layer with an electrolytic solution suitable for selectively etching the doped silicon of said opposite type of conductivity, making the silicon of the buried layer porous; and oxidizing and leaching away the silicon so made porous.
Preferably, the trenches or holes for accessing the doped buried layer are cut through the epitaxial layer and not through the rear of the monocrystalline silicon substrate thus avoiding the burden of precisely aligning the mask on the rear surface with the masks that are used on the front surface of the substrate. Moreover, the thickness of the substrate is normally much greater than that of the epitaxial layer and thus the need to cut relatively deep and narrow trenches requiring the use of special plasma etching equipment is avoided.
Abstract:
The sensor (1) comprises an array of photo sensitive elements for acquiring images of the passenger compartment in a motor vehicle and a circuit (20) for processing the signals corresponding to said images generated by said photo sensitive elements. The processing circuit is configured according to a cellular neural network (CNN) processing architecture of said image signals and can generate an output signal indicating the decision on whether to deploy an airbag (3) to which the sensor is associated or to control the explosion of the airbag. Preferably, the photo sensitive array and the processing circuit are comprised on a single integrated component, preferably implementing CMOS technology.
Abstract:
The invention relates to a novel neuro-fuzzy integrated architecture which permits of on-line self-training. The architecture comprises at least one microcontroller (11) of the fuzzy type (fuzzyfier/defuzzyfier), dedicated to fuzzy rules computing and integrated monolythically on a semiconductor together with a non-volatile memory (9). Also provided within the same integrated circuit, are a microprocessor (5), a volatile memory unit (2), and an arbiter block (3) which is linked to a bus (4) interconnecting the fuzzy microcontroller (11), the microprocessor (5) and said volatile memory unit (2); the arbiter block (3) acts to control access to the memory unit (2) by the microprocessor (5) or the fuzzy microcontroller (11). An additional fuzzy co-processor (6) is connected between the fuzzy microcontroller (11) and the microprocessor (5) for handling the fuzzy logic operations.