A new signal processing tool for the detection of pulses travelling through media with complex or unknown dispersion properties was developed by the group of Prof. Gal-Yam, originally for detecting radio bursts in astronomical observations.
Pulses are applied in various fields such as oil & gas exploration, detection (e.g. sonar, lidar and radar) and communication. When pulses pass through dispersive media, the arrival times at the detector of different frequency components may differ, and as a result the pulse may become degraded (e.g. transformed to a longer pulse with reduced intensity), even to the level of becoming indistinguishable in terms of signal to noise. This problem becomes even more challenging when detecting short pulses that travel through complex or unknown media.
The new method presented here provides a proven and efficient solution that can be applied for different scenarios where short pulses dispersed by complex media are used.
- Detection and surveying technologies- sonar, lidar, radar etc
- Efficient, requires limited computational resources
- Generic, can be applied to various setups
- Easily implementable into existing systems
The method includes obtaining an input array of cells, each indicating an intensity of a frequency component of the signal at a representative time. A fast dispersion measure transform (FDMT) is applied to concurrently sum the cells of the input array that lie along different dispersion curves, each curve defined by a known non-linear functional form and being uniquely characterized by a time coordinate and by a value of the dispersion measure. Application of FDMT includes initially generating a plurality of sub-arrays, each representing a frequency sub-band and iteratively combining pairs of adjacent sub-arrays in accordance with an addition rule until all of the initially generated plurality of sub-arrays are combined into an output array of the sums, in which a cell of the output array that is indicative of a transmitted pulse is identified.