Optimized Contact Scheduling for NOAA Search and Rescue Overview - - PowerPoint PPT Presentation

Optimized Contact Scheduling for NOAA Search and Rescue

Orbit Logic Proprietary

Overview

2

Orbit Logic Proprietary

Not Just Clouds

• NOAA satellites

don’t just collect weather data

• NOAA satellites

also assist in Search and Rescue

3

Orbit Logic Proprietary

What is SARSAT?

• NOAA operates the Search And Rescue Satellite Aided

Tracking (SARSAT) system to detect and locate distress signals around the world

4

Orbit Logic Proprietary

Importance of Maintaining Contact

• Satellite data is only

useful if it can be sent to a ground station via a space to ground contact

• Maintaining consistent

contact means more data is available

5

Orbit Logic Proprietary

The Problem

• For the best chance at

finding distress signals, the space to ground contact schedule for the GPS satellites must be

• ptimized to maintain a

low DOP

6

Orbit Logic Proprietary

What is DOP?

• DOP
• Dilution of precision
• Measurement of error

propagation when making position measurements via satellite

• DOP values range from 1 to ∞
• 1 is ideal
• 2-5 is reliable for navigation
• 5-10 is moderately reliable for

some calculations

• 10-20+ is poor

7

Orbit Logic Proprietary

System Design

8

Orbit Logic Proprietary

The Team

• Orbit Logic
• Isabel Martinez, Zachary Roberts, Ella Herz
• NOAA
• Jesse Reich, Larry LeBeau, Kris Tai
• AGI
• Chris Mulcahy, Reggie Pforter, Jens Ramrath

9

Orbit Logic Proprietary

The Solution

• Orbit Logic, in collaboration with NOAA and AGI,

developed a software suite to solve this problem:

• STK
• Coverage tool for DOP calculations
• Scheduler
• Manages scheduling resources
• Deconfliction for scheduling tasks
• Plugin Code (python / C#)
• Manages creation of STK scenario
• Tracks DOP values
• Adjusts scoring of tasking opportunities based on DOP

10

Orbit Logic Proprietary

Inputs and Outputs

• Inputs
• Satellite orbit data
• Ground station location and obscuration masks
• Ground targets / Areas of Regard (AOR)
• Outputs (UI, .csv, keyword/value text files)
• Contact schedules

11

Orbit Logic Proprietary

Workflow

User configures inputs Access and DOP Reports SARSAT service performs

• ptimization

Schedule generated

12

Orbit Logic Proprietary

The First Approach

13

Orbit Logic Proprietary

Proof of Concept

• Did initial prototyping demo using a python script
• Controlled STK and Scheduler via API commands
• Created tasks in Scheduler based on satellite to ground station access times
• Queried STK for DOP reports, adjusted task timeslot scoring in Scheduler
• Used default algorithms (One-Pass)
• DOP optimization done in script, not in Scheduler itself
• Contact tasks were broken up into 9 minute chunks
• Each satellite to ground contact was a minimum of 9 minutes
• Original optimization of schedule took 11 minutes from start to finish
• Original script was not parallelized
• Used an average DOP value for several targets across continental US

14

Orbit Logic Proprietary

Original Demo Screenshots

15

Orbit Logic Proprietary

How’d It Do?

• Pros
• Okay performance (took 11 minutes

for whole optimization run)

• Some optimization based on good

satellites to use for good DOP

• Cons
• Timeslot scores were all set to 100

(highest) as long as they used any of the best satellites

• Did not include actual DOP values for

timeslot scoring

• Not easily configurable
• Points in US had to be chosen prior to

scheduling

• Doesn’t check if best satellites

actually have access to a ground station

16

Orbit Logic Proprietary

The Second Attempt

17

Orbit Logic Proprietary

Goals For This Attempt

• After Orbit Logic received the official statement of work, it was

time to make some updates to the proof of concept

• Goals for the first round of updates included:
• Checking that there was actually access between the ground stations

and the chosen best satellites

• Tracking the actual computed DOP value for better use in optimization
• Creating a mesh of points based on user-defined AOR instead of

selecting all points manually

• Converting python to C# for performance purposes
• Also, NOAA said that the scheduling system no longer had to

schedule LEO contacts

• Instead, Scheduler will ingest the LEO contact schedule
• The LEO contact schedule would become windows of unavailability on

the resources

18

Orbit Logic Proprietary

Script Updates

• Of the previously listed goals, all

were accomplished

• In the transition from python to C#,

Orbit Logic also updated the following:

• DOP value now tracked for each

point, and the average DOP value across all points was used for scoring timeslots

• Made input parameters configurable
• utside of code
• Refactored method used for setting

access restrictions in STK which boosted performance

19

Orbit Logic Proprietary

But Was It Better?

• Pros
• Very fast performance (~2

minutes)

• Optimization now based on

actual DOP values and best set of available satellites

• Easily configurable
• Cons
• Poor coverage in some

locations

• Antennas not fully

scheduled for contacts throughout the entire scheduling period

20

Orbit Logic Proprietary

Closing the Gaps

21

Orbit Logic Proprietary

What Was Going Wrong?

• NOAA’s analysis showed

there was poor DOP coverage near Alaska and the dateline

• But in the schedule itself, it

appeared all satellites that had access to that region were already contacting the ground station for the full time of access

• Needed to investigate the

problem setup in Scheduler

22

Orbit Logic Proprietary

Fixing Access

• The first problem was related to how the access

time constraint was used in STK

• The access check added in the previous major update

would send all access windows per satellite into STK separately

• STK would in turn intersect all of the provided access

times

• This overly constrained the tasks to only occur if all

ground stations had access to a given satellite

• Changed the access constraint
• Access windows were unioned prior to going to STK
• Now it allows access as long as any one of the ground

stations can see the satellite

23

Orbit Logic Proprietary

Allowing multiple Satellite Contacts

• NOAA pointed out that satellites were

allowed to contact multiple ground stations at a time

• Original implementation assumed each

satellite could only contact 1 ground station at a time (ie. accommodation = 1)

• Updated the scheduling method
• First algorithm run assumes all satellites

have accommodation = 1

• Locks all assignments in place from first

algorithm run

• Second algorithm run assumes all

satellites have accommodation = unlimited

• Second algorithm run fills in the gaps
• This method prevents contact over-

assignment for satellites

• Blue = Locked tasks from first run
• Green = Assigned tasks from second run

24

Orbit Logic Proprietary

How Are the Results Now?

• Pros
• Much longer contacts
• DOP improvements
• Cons
• Gaps that are less

than the minimum contact duration length are still unscheduled

25

Orbit Logic Proprietary

Discussion

26

Orbit Logic Proprietary

Further Work

• Improvements needed to resolve the small

gaps

• Could be problem setup updates in Scheduler
• Could be algorithm updates for Scheduler

27

Orbit Logic Proprietary

Conclusions

• Final software suite does create optimized contact

schedules based on DOP

• Needed to iterate on the problem setup many times based on

results from testing

• Each iteration improved the DOP optimization
• Final software suite is easy to use as well
• Minimal user involvement needed to create optimized

schedules

• Scheduling method could be applied for any system to
• ptimize contact schedules for DOP
• Configurability allows users to run analysis for any satellite

constellation, ground station and AOR combination

28

Orbit Logic Proprietary

Rescues This Year

29

Orbit Logic Proprietary

Contact Us

Orbit Logic 7852 Walker Dr., Suite 400 Greenbelt, MD 20770 www.OrbitLogic.com Email: Isabel.martinez@orbitlogic.com sarsat@orbitlogic.com Phone: (301) 982-6232 x7103

30

Orbit Logic Proprietary

Backup

31

Orbit Logic Proprietary

Acronyms

• AGI
• Analytical Graphics

Incorporated

• AOR
• Area of Regard
• DOP
• Dilution of Precision
• GEO
• Geostationary Orbit
• GPS
• Global Positioning System
• LEO
• Low Earth Orbit
• NOAA
• National Oceanographic and

Atmospheric Administration

• PDOP
• Position Dilution of Precision
• SARSAT
• Search And Rescue Satellite

Aided Tracking

• USMCC
• United States Mission Control

Center

32

Orbit Logic Proprietary

Glossary

• Accommodation
• Number of tasks a resource can support simultaneously
• Resource
• Any object or entity required for schedule creation (ex. a

satellite)

• Task
• An action in the schedule that requires resources to complete

(ex. satellite contact)

• Timeslot
• Windows of time in which task assignments may occur. They

are constrained by access and resource availability times

33