Abstract:
In adaptive beam forming, the beam produced by single sensor is cumulative result of all
elements in SONAR. The side lobes produced by any sensor is cancelled by the
neighboring sensors. When any sensor is defective then the interference is not cancelled
and the beam pattern produces major degradation in beam pattern by producing the side
lobes. Due to this factor, the detection of target becomes difficult. In this research, we
have applied two types of filters blind and non-blind adaptive filters that cancel the
interference caused due to defective sensors in SONAR. We have used adaptive filter
when the desired beam pattern is known at the receiver or it is not known at the receiver
side. It has been observed that even if up to 50% of the sensors are degraded, proposed
system works at its optimum level. Results show that adaptive filter is computationally
efficient, works in presence of noise and operates in the environment when the actual
beam pattern is known or not known at the receiver side. It has been noted that filter
produces output that is very near to its optimum value. The results show that it is feasible
to construct a real time acoustic beam former with realistic parameters using parallel
processing architecture.
