الفهرس Only 14 pages are availabe for public view
 |  | from | 208 |  |  |
| | | from | 208 | | |
Abstract
Several algorithms have been proposed and studied for adaptive arrays. Most of these algorithms are intended for the case where both a desired signal and interference are being received by the array. The function of the algorithm is to control the array pattern to minimize interference and maximize desired signal at the array output. Some algorithms deal with maximizing signal-to-noise ratio at the array output such as Apple Baum algorithm and gain optimizing algorithm, but others deal with minimizing the mean-square error between the actual array output and a desired array output (called the reference signal) such as LMS algoritlu11 which use the gradient approach. Usually the LMS represents the basic algorithm and the other algorithms are related to it. The gradient approach used to solve the control problem posed by adaptive array weight adjustment is very popular, relatively simple, and generally well understood method that permits the solution of a large class of problems. Other algorithms that have been proposed for adaptive arrays include recursive, least squares estimation techniques, additional modified versions of the LMS algorithms, a controlled time delay algorithm, covariance matrix inversion methods, and guided accelerated random search techniques. For the LMS algorithms, we study the convergence behavior, transient response characteristics, and weight maladjustment and also for the constrained LMS algorithm. In chapter one we present three algorithms, the first algorithm is the constrained cascade configuration algorithm which presents a new approach to adaptive beam forming based on the Davies cascade array configuration. Each stage of this cascade network is capable of nulling out one interference source while maintaining a desired response in some other specified look direction. The structure of the resulting processor lends itself to easy computation of the weights due to the fact that the processor can consider only one stage at a time. Weights within anyone stage can be computed efficiently because we need to compute only a two-dimensional weight vector. Further, by relying directly on the correlation matrices of two-dimensional signal vectors for computing the weights, the convergence speed can be increased. The second proposed algorithm is the constrained fast null steering algorithm which uses the LMS algorithm to reject interferences in adaptive arrays with a different technique. The main advantage of LMS algorithm is its simplicity. However, the major disadvantage of LMS algorithm is that its convergence behavior is dependent on the external noise environment and thus can be very slow in severe jamming situations.speed of the LMS algorithm without sacrificing too much of its implementation simplicity, null steering beamfonners can be employed as the underlying array processing structure. The third proposed algorithm is the new constrained fast null steering algorithm which is based on the constrained fast null steering algorithm. The structure of this algorithm is simpler than that of the tree structure of the constrained fast null steering. A multiplier that is inserted in the structure multiplies the output of any summing points with the main output. The function of the multiplier is to transfer a zero when occurred at any summing points to the main output. A manual gain control is inserted in the structure for maintaining the level of directional pattern to unity.
In chapter two, the generalized constrained fast null steering algorithm is presented. It is based on the new constrained fast null steering algorithm. The structure of this algorithm is tree structure like that of the constrained fast null steering. A multiplier, which is inserted in the structure, multiplies the output of any summing points with the main output. The function of the multiplier is to transfer a zero when occurred at any summing points to the main output. Numerical evaluation proved the validity of this algorithm using four and five elements arrays. Additionally, the new algorithm has the ability to null 4N-6 interferences for N=3, 4 where N+ I is the number of elements of the array.
In chapter three, due to the increasing demand for mobile communication capacity in limited radio frequency (RF) spectrum motivates the need for new techniques to improve spectrum utilization. One approach for increasing spectrum efficiency in digital cellular is the use of spread spectrum code division multiple access (COMA) technology. Another approach is by using the adaptive antennas array in a COMA system, the amount of co-channel interference from users within the same cell as well as neighboring cells can be reduced, and therefore the system capacity can be increased. There exist many adaptive algorithms that must have the ability to separate and extract each user’s signal simultaneously.
The perfonnances of two blind adaptive algorithms using dynamic and static data blocks are investigated. The first one is least squares despread respread multi target array (LSORMTA). Its processing structure utilizes the infonnation of the spreading signals of different users in a CDMA system to adapt the weight vectors of a multitarget beamfonner. The second one is least squares despread respread multitarget constant modulus algorithm (LSDRMTCMA). Its processing structure combines the infonnation of the spreading signal and the constant modulus property of the transmitted signal in the adaptation of the weight vectors. From the simulation and the processing structure, it is clear that the infonnation of
the spreading signals is strong enough to extract a certain user from a multiuser system. Also, there is no difference in simulation using either dynamic or static data block.
In chapter four, a detailed overview is used for the various methods for estimating the direction-of-arrival (DOA) (also called angle-of-arrival (AOA)) of a radio signal using an antenna array. The array-based DOA estimation techniques considered are broadly divided into two different types: conventional techniques and subspace based techniques. Conventional techniques are based on classical beamforming and require a large number of elements to achieve high resolution. Subspace bascd tcchniques are high resolution sub¬optimal which exploit the eigen structure of the input data matrix. Also, in case of coherent signal conditions we apply the spatial smoothing technique on subspace based method to get more accurate estimatcd rcsults.
i edon I
,
I than
rin
pion
~n ~ of :
I I
tIt is
PI is lei in i ition
~ain r five
for
Chapter five uses the timc-diffcrcnce-of-arrival (TDOA) tcchnique for estimating the time difference between the received signals to base stations of a certain subscriber using the generalized cross correlation (GCe) method. The Fang’s and Chan’s methods are used for solving the resulting hyperbolic equations for finding the position location (PL) of a certain subscriber. The Fang’s and Chan’s methods give accurate and same result. This means that, an accurate position location estimation of a certain subscriber requires an efficient hyperbolic position location estimation method. It is also clarified that a more data samples give a more accurate result.
Chapter six describes a capacity improvement method applied on CDMA system. This method uses a nulling network structure which is inserted between the antenna elements array and the algorithm structure. The main role of the nulling network for a certain port is to put zeros in all subscribers’ directions except the subscriber assigned to that port. The benefit from the nulling network is to increase the signal-to-interference-and-noise-ratio (SlNR) by decreasing the interference power level for each subscriber. The structure of the nulling network is a tree structure. A multiplier, which is inserted in that structure, multiplies the output of any summing points with the main output. The function of the multiplier is to transfer a zero when occurred at any summing point to the main output. Numerical evaluation proved the validity of this nulling network structure using four and five elements array. Additionally, this structure has the ability to null 4M-6 interferences for M=3, 4 where M+ I is the number of elements of the array. Also this structure accomplished a less computational complexity.