Abstract
Recent work has highlighted the benefits of exploiting robust
Capon beamformer (RCB) techniques in passive sonar.
Unfortunately, the computational requirements for computing
the standard RCB weights are cubic in the number of
adaptive degrees of freedom, which may be computationally
prohibitive in practical situations. Here, we examine recent
computationally efficient techniques for computing the RCB
weights and evaluate their performances for passive sonar.
We also discuss the implementation of these efficient algorithms
on parallel architectures, such as graphics processing
units (GPUs), illustrating that further significant speed-ups
are possible over a central processing unit (CPU) based implementation.
Capon beamformer (RCB) techniques in passive sonar.
Unfortunately, the computational requirements for computing
the standard RCB weights are cubic in the number of
adaptive degrees of freedom, which may be computationally
prohibitive in practical situations. Here, we examine recent
computationally efficient techniques for computing the RCB
weights and evaluate their performances for passive sonar.
We also discuss the implementation of these efficient algorithms
on parallel architectures, such as graphics processing
units (GPUs), illustrating that further significant speed-ups
are possible over a central processing unit (CPU) based implementation.
| Original language | English |
|---|---|
| Publication status | Published - 2015 |
| Event | Underwater Defence Technology (UDT 2015) - Rotterdam, Netherlands Duration: 2015 Jun 3 → 2015 Jun 5 |
Conference
| Conference | Underwater Defence Technology (UDT 2015) |
|---|---|
| Country/Territory | Netherlands |
| City | Rotterdam |
| Period | 2015/06/03 → 2015/06/05 |
Subject classification (UKÄ)
- Signal Processing
Free keywords
- Passive sonar
- computationally efficient robust adaptive beamforming.