Stochastic Resonance

1
Stochastic Resonance

• Adding Noise to a signal can help in its
  detection.
• Just the right amount of noise must be added
  (resonance)
• SR in Nature
• Periodic ice age prediction
• Crayfish warnings of approaching bass - a
  periodic fin motion
• Dogfish spit noise to better detection.
• Neurons

2
Stochastic Resonance

• What Causes SR?
• Non-Additive Noise
• Nonlinearities
• Mathematical Definition
• There is none

3

• Example Miyamoto Resonance
• Add noise to image and threshold. Vary noise.

0 15 30 60
90 120 180 240
4
Most Studied System
5
Numerics (Mataim et al. 1998)
6
Numerics (Mataim et al. 1998)
7
Numerics (Mataim et al. 1998)
8
Comparison of NP Optimal Detector

• Coherent NP Correlator vs. Stochastic Resonance
• Assumptions
• Compared NP on input and after SR nonlinearity
• Detector COHERENT
• Noise White Gaussian
• Galdi, Pierro, Pinto (Phys Review E, June 1998).

9
Comparison of NP Optimal Detectors

• Coherent NP Correlator vs. Stochastic Resonance
  (4 dB)
• Galdi, Pierro, Pinto (Phys Review E, June 1998).

10
Noncoherent

• Noncoherent SR sign detector
• Galdi, Pierro, Pinto (Phys Review E, June 1998).

11
Noncoherent Results

• Noncoherent correlator is about 3dB worse than
  coherent.
• Noncoherent SR is better than noncoherent
  correlator for low SNR
• Galdi, Pierro, Pinto (Phys Review E, June 1998).