Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method and Means for High-Fidelity Digital Signal Injection into Hard Clipped Sensor Data

IP.com Disclosure Number: IPCOM000117128D
Original Publication Date: 1995-Dec-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 2 page(s) / 101K

Publishing Venue

IBM

Related People

Buratti, RJ: AUTHOR [+3]

Abstract

Many sensor information processing applications including radar, sonar, scismic profiling, radio astronomy, bio-engineering and infrared imaging require front-end injection of generated sensor signals for training of operators and for system testing. Often, these systems involve a large number of sensors and a broad information bandwidth to obtain a large spatial and temporal processing gains required to identify the presence of small, application dependent signals. Processing technology implementation constraints coupled with the need to preserve the large gains inherent in these systems often result in the use of digitized, compressed sensor data, including single bit clipped sensor data.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Method and Means for High-Fidelity Digital Signal Injection into
Hard Clipped Sensor Data

      Many sensor information processing applications including
radar, sonar, scismic profiling, radio astronomy, bio-engineering and
infrared imaging require front-end injection of generated sensor
signals for training of operators and for system testing.  Often,
these systems involve a large number of sensors and a broad
information bandwidth to obtain a large spatial and temporal
processing gains required to identify the presence of small,
application dependent signals.  Processing technology implementation
constraints coupled with the need to preserve the large gains
inherent in these systems often result in the use of digitized,
compressed sensor data, including single bit clipped sensor data.
Formerly, signal injection for these non-linear hard clipped systems
was performed using an analog approach injecting signals at the
sensor level prior to digitization.  In this disclosure, presented is
an improved method for signal injection done digitally at the clipper
output with substantial cost savings over the former method and with
essentially no observable impact on the original systems performance.
Key advantages of the design are:
  1.  a fail-safe interception of the normal system datastream,
       insuring no impact on system operation when the injection
       capability is off;
  2.  a simple gate array implementation of a Filter/Adder/Clipper
       (FAC) for combining the original signal with the injected
signal,
       providing a minimum hardware implementation, effectively
       eliminating the need for reclipping the composite signal and
       eliminating any further clipping loss; and
  3.  use of negatively correlated injection noise, which is
eliminated
       by subsequent system processing minimizing any corruption of
the
       original system signals.

      This invention relates to a method and implementation for
digital insertion of high fidelity signals into hard-clipped sensor
data information processing systems.  When compared with former
analog insertion techniques, the digital insertion implementation is
significantly simplified.  This provides a minimal implementation
that exceeds the performance of the former method while having no
noticeable impact on normal system operation.  Key elements on the
invention which distinguish it relative to former techniques include:
  1.  use of a passive, fail-safe signal insertion into the normal
       sensor signal data flow;
  2.  use of an innovative Filter/Add...