Dynamic Scan Scheduling Specification

1
Dynamic Scan Scheduling Specification

• Bruno Dutertre
• System Design Laboratory
• SRI International
• E-mail bruno_at_sdl.sri.com

2
Dynamic Scan Scheduler
External Parameters
Performance Requirements
Schedule Construction
Assessment
Scan Schedule
EW Receiver
Signal Data Processing
Detected Pulses
Detected Emitters

• Assessment function
• determines when rescheduling is required
• specifies requirements for the new schedule
• Schedule-construction function
• generates a schedule that meets the requirements
• under real-time constraints

3
DSS Specification

• Objective
• Determine requirements for the schedule-constructi
  on function (on a single platform)
• Key Issues
• Schedule representation
• Metrics for expressing performance requirements
  and measuring schedule performance

4
Schedule Representation

• Strongly periodic schedules
• Defined by pairs dwell time/revisit time for each
  frequency band
• Feasibility results show that this is too
  restrictive
• Pattern-based schedules
• Defined by a basic pattern repeated periodically
• The pattern describes a finite list of successive
  dwell intervals

Dwell interval
Pattern
5
Performance Metrics

• Two emitter types
• Search and track radars produce successive
  illuminations (rotating beams)
• Missile radars continuously illuminate their
  target
• In both cases, detection and tracking require
  intercepting a minimal number M of pulses in a
  single dwell
• Good performance requires a high probability of
  intercepting M or more pulses in dwell intervals

6
Periodic-Illumination Emitters

• First Metric Coverage
• The probability of intercepting at least M pulses
  from a single illumination

?, m emitter parameters L, n, A parameters
derived from the schedule pattern
This gives an estimate of how well the schedule
does at detecting an illumination from an emitter
not already detected
7
Extensions of Coverage

• The previous metric can be generalized to
• Coverage with respect to successive illuminations
  (probability of detecting an emitter after a few
  illuminations)
• Relative coverage estimate of how good the
  schedule is for tracking already detected
  emitters (uses information about the likely time
  of occurrence of future illuminations)
• Probabilistic coverage to deal with emitters
  whose characteristics are not known with
  exactitude, but with some probability
  distribution
• All these metrics can be computed from the
  schedule pattern

8
Metric for Continuous Emitters

• Requirements for continuous emitters
• A good schedule must minimize detection delays
• Associated metric
• Expected delay between the activation of the
  emitter and the interception of at least M pulses
  in a single dwell

? pulse repetition interval of the emitter L,
n, parameters derived from the schedule
pattern
9
Global Performance Constraints

• We can partition emitters in two classes
• Emitters already detected (that need to be
  tracked)
• Emitters likely to be present (that need to be
  searched for)
• This gives three sets of constraints on the scan
  schedule
• Tracking constraints maximize the relative
  coverage for each tracked emitter
• Searching constraints for continuous emitters
  minimize the expected detection delay for each
  probable emitter
• Searching constraints for periodic emitters
  maximize coverage for each probable emitter

10
Conclusion

• New results
• Analysis of scan-schedule performance
• Metrics for evaluation of pattern-based schedules
• Requirements for a schedule-construction
  algorithm
• Future work
• Algorithm development and experimentation
• Extension to the multi-platform case