Date of Award

January 2014

Document Type


Degree Name

Master of Science (MS)


Electrical Engineering

First Advisor

Saleh Faruque


In this thesis work, a bandwidth efficient coded modulation technique has been proposed for LASER based communication. It offers reliable error correction and bandwidth optimization for free space connectivity. Laser communication has been emerging as a potential alternative of Radio Frequency (RF) communication for long-haul connectivity. It provides several advantages over RF systems. Some of specific advantages are its high speed data transfer, higher bandwidth, robust security, immunity to electromagnetic interference, lower implementation cost. However, as any other communication system, it requires an error control system that detects and corrects error in the data transmission. The proposed technique blends well with the laser communication systems to provide a very good error correction capability. For a decent code length it can correct around 24 percent of error. The proposed method is named as Orthogonal On-Off Keying of LASER and uses a Bi-orthogonal matrix to generate the encoded data for laser. While decoding, it uses the cross correlation between two orthogonal codes to detect and correct any errors during transmission. Orthogonal codes also known as Walsh codes are used for error correction in Code Divisional Multiple Access of RF communication. But it has not been investigated for laser systems. Besides on-off keying this code also has enormous potential in other modulation techniques as well. Here in this thesis project, the error control performance has been verified using an experimental setup. Furthermore, a method for more bandwidth efficient transmission of data using laser have been proposed and discussed. This method of bit splitting offers bandwidth efficiency of unity code rate.

Altogether, the proposed technique is a promising solution for error detection and correction and at the same time a bandwidth efficient system.Another advantage of orthogonal coding is self-synchronization capability as the modulated signals share orthogonal space as well. All the codes in orthogonal matrix are unique as per their properties and can be identified separately. As a result, it does not require any synchronization bits while transmission. This results in reduction of complexity in implementation and thus yield savings economically.

Like orthogonal matrix, traditional block coding also uses a block of information bits. Say, they are segmented into a block of k bits. This block is transformed into a larger block of n bits by adding horizontal and vertical parity bits. This is denoted as (n,k) block code. The problem with this type of coding is, the added (n-k) bits carry no information and only helps in error detection and correction. However this bigger block of data can detect and correct only one error in the transmission and it fails, if more than one error occurs.

On the other hand, the data is mapped using an orthogonal table in this type of coding where codes have unique properties. No redundant bits are required to be added as they possess parity generation property with themselves. Also, the distance property of orthogonal codes makes it stronger for error detection. For codes of greater length in size, such approach is capable to correct more than one error.

The test bed is implemented using a LASER transmitter and hardware interface that includes a computer to receive the data and transmit. The operations such as data capture, modulation, coding and injection of error are carried by a software written in MATLAB®. On the receiver side, a high speed photo-detector is placed with a hardware interface with another computer. This one has the other part of the program to receive the bits and decode to extract the transmitted data. To test the error control capability, errors are intentionally transmitted by altering number of bits in the modulated signal. Upon reception, the data is compared with the transmitted bits and evaluated. This test goes through 8, 16, 32 and 64 bits of orthogonally mapped data, several different speed of transmission and a range of error percentage. All the results were compared and investigated for prediction and error tolerance.