Abdul Ahad Sami Awwal

Multiprocessor design using polarization-encoded optical shadow-casting

An efficient algorithm for designing a multiprocessor logic unit using a polarization-encoded optical shadow casting system is presented. Both efficient minterm pairing and associative memory mapping with dual addresses are exploited for the design. For illustration, the algorithm is used to design a parallel adder-subtracter unit.

Recoded signed-digit binary addition–subtraction using opto-electronic symbolic substitution

Two opto-electronic implementations of carry-free addition and borrow-free subtraction of recoded signed-digit binary numbers that use optical symbolic substitution are proposed.

Polarization-encoded optical shadow casting: design of trinary multipliers.

The polarization-encoded optical shadow casting (POSC) scheme is used to design a 2-digit by 2-digit trinary multiplier. Three design algorithms for identifying the appropriate source patterns, input patterns, and output mask for both serial and parallel operations are presented. The inclusion of analyzer pattern reduces the number of pixel subcells for the POSC inputs.

Gaussian-to-Bessel beam transformation using a split refracting system

A two-element refracting system is designed to transform a Gaussian laser beam into a diffraction-free Bessel beam. The resulting input and output surfaces are almost spherical, which makes for easy implementation of the system.

Two-element refracting system for annular Gaussian-to-Bessel beam transformation

A refracting system consisting of two lenses is designed to transform an annular Gaussian laser beam into a circular Bessel beam. The slopes of the input and output surfaces fit well with a sixth-order polynomial. A smooth variation of the radii of curvature of the resulting aspheric surfaces is very attractive for easy machining of the surfaces. The diffraction-free length for the designed system is 59.4735 m at 633 nm.

Digital optical processing based on higher-order modified signed-digit symbolic substitution.

An efficient higher-order modified signed-digit optical symbolic substitution technique for addition and subtraction has been derived. This two-step symbolic substitution technique, which is independent of the number of bits to be processed, saves a tremendous amount of computation time. Two separate approaches for optical implementation that use content-addressable memory and the joint transform correlator are also suggested.

Multiplier design using RBSD number system

A high-speed multiplier is designed based on recoded modified-signed digit binary number system. In the redundant binary representation each digit is represented by either 0, 1, or -1, which can be recoded into a new form, whose addition is free from carry propagation. In the recoded number, there are no two consecutive 1s or -1s, as a result, it eliminates the propagation of carry. Carry free adders based on recoded number are the basic building blocks of high-speed multiplier, which leads to a regular cellular array structure suitable for VLSI implementation. The multiplier consists of partial product generator followed by the encoder which converts binary numbers to recoded binary signed-digit (RBSD). These RBSD partial products are then added by recoded carry free adder.<<ETX>>

Fast carry free adder design using QSD number system

A high speed parallel full adder is designed which can perform carry-free addition of two modified signed digit quaternary numbers. For digital implementation, the sign digit quaternary numbers are represented using 3-bit 2s complement notation. The adder truth table with possible schemes of the electronic and optical implementation are provided.<<ETX>>

Polarization-encoded optical shadow-casting arithmetic-logic-unit design: separate and simultaneous output generation

A memory-efficient dual cell and a multioutput parallel arithmetic logic unit are designed by using a polarization-encoded optical shadow-casting scheme. The design algorithms for identifying the source patterns, input encoding, and output mask are also presented.

Enhanced artificial neural networks using complex numbers

The model of a simple perceptron using phase-encoded input and complex-valued weights is proposed. The aggregation function, activation function, and learning rule for the proposed neuron are derived and applied to two and three input Boolean logic functions. An improvement of 135% over the theoretical maximum of 104 linearly separable problems (of three variables) solvable by conventional perceptrons is achieved without additional logic, neuron stages, or higher order terms such as those required in polynomial logic gates. Such a network is very attractive for optical implementation since optical computations are naturally complex.