Spacecraft Mission Assessment and Replanning Tool SMART

1
Spacecraft Mission Assessment and Replanning
Tool (SMART) A Real-Time, Intelligent Autonomous
Flight Management (AFM) System for Increased
Safety and Performance of Human Spaceflight
Vehicles Jeremy Hart/JSC-EG jeremy.j.hart_at_nasa.gov
Howard Hu/JSC-EGRyan Proud/JSC-DM
Lyndon B. Johnson Space Center Houston, Texas
2
Contents

• NASAs Human Spaceflight Goals
• Decision-Making Process
• Levels of Autonomy
• SMART Concept
• SMART Products
• Current Status of SMART Development

3
NASAs Human Spaceflight Goals

• To enhance crew and vehicle safety
• To reduce development and life cycle costs
• To increase mission success

• NASA is evaluating the ability of autonomous
  systems to help achieve these goals
• One autonomous system concept, SMART, is being
  developed to evaluate the benefits of
  computer-based decision-making for human
  spaceflight vehicles

4
Generic Decision-Making Process
Observe the world around you
Analyze current and future possibilities
Decide on thebest course of action
Command the action to be taken
Iterate on the process
5
Specific Decision-Making Process
ExternalInformation
Observe
Orient
Decide
Act
Vehicle Constraints Flight Rules Analyses and
Synthesis Previous Experience New Information
UnfoldingInteractionWithEnvironment
UnfoldingInteractionWithEnvironment
Feedback
Feedback
Decision does not necessarily lead to actionit
can also spawn new observations or analysis tasks
(Orient) The Observe, Orient, Decide, and Act
(OODA) loop is based on work performed by Col. J.
R. Boyd
6
Human Spaceflight Decision-Making Functions
7
Specific Example
Abort Decision-Making Function
8
Space Shuttle Abort Decision-Making

• One of the most demanding tasks performed by the
  Space Shuttle flight crew is the process of
  determining whether, when, and where to abort the
  vehicle should engine or system failures occur
• The Shuttle Abort Flight Management (SAFM)
  application was developed to help maximize crew
  situational awareness and reduce flight workload
  by autonomously providing the crew dynamic
  information about the abort capability of the
  Shuttle during ascent and entry
• Previously this task was performed only by
  ground-based flight controllers
• SAFM automated the abort decision-making process
  through the Decide step

9
SMART - Natural Extension of Current Work

• SMART is intended to evolve a step further by
  collecting a group of decision-making functions
  and incorporating them into an Autonomous Flight
  Management (AFM) system
• In the case of the abort decision-making
  function, SMART will not only provide information
  but also assess the information and act upon the
  resulting decisions

SMART AFM
Abort Decision-Making Function
10
Levels of Autonomy

• Increasing computer involvement in the
  decision-making process does not necessarily
  eliminate the need for human interaction. Some
  level of involvement will be required across all
  four OODA types
• During the initial SMART development process a
  standard method for determining the appropriate
  level of autonomy (LOA) for all of the vehicles
  AFM functions was required
• Through our research, we were unable to find
  method that met our criteria for use during the
  design phase of a human spaceflight vehicle.
• Therefore, we created our own method, the
  Function-specific LOA Tool (FLOAT)
• The FLOAT uses an approach that balances how much
  human operators will trust the system with the
  best cost/benefit ratio for building the system

11
Level of Autonomy Scale
12
SMART Vision

• SMART is an AFM concept that incorporates
  advanced performance assessment technologies,
  autonomous decision-making algorithms, and past
  flight experience to improve crew/vehicle
  survival and reduce life-cycle costs

13
SMART Concept
SMART
Priority Listing of Mission Objectives
Database of Rules
SMART Decision-Making Algorithm
Opportunities
Orient
Vehicle Management System (VMS)/ Integrated
Vehicle Health Management (IVHM)
Observe
Problems
New Planning Problem
Nominal Flight

• Plans
• Objectives
• Constraints
• Capability

Decide
Act
New Plan
Crew Inhibit
Data
Prelaunch
Ascent
Entry
Orbit
Decisions
Flight-Phase-Specific Algorithms
14
SMART Phase 1 Prototype

• MATLAB Simulink simulation
• Earth Model
• Booster Model
• Service Module Model
• Capsule Model
• STK graphics capability
• Guidance Models coded in C
• Booster, Capsule, and On-Orbit Guidance
• Code is interfaced with Simulink using
  S-Functions
• Capability exists for future CEV work
• Select_site.c and GNC_Apollo.c
• ADEPT capability with increased stability
• Existing simulation framework to build from
• Experience base with ADEPT and sim

15
SMART Products

• LOA Assessment Tool (FLOAT) created for designing
  autonomous systems

AFM functional requirements document that
includes the appropriate LOA

• AFM prototype based on the SMART concept
• Serves as the key test bed for determining AFM
  requirements, quantifying cost/benefit metrics,
  and evaluating algorithms

16
Current Status

• Hubble Return Vehicle (HRV) Feasibility
  Assessment Effort
• Team is responsible for providing a feasibility
  assessment of the HRV to perform an autonomous
  rendezvous and capture of the Hubble Space
  Telescope (HST)
• Assessment of Level of Autonomy (LOA), using
  FLOAT, required to achieve mission success
• Design and develop the autonomous flight
  management system to account for time critical
  contingencies during prox ops and docking.
• Mini-AERCam Mission Management System Development
• Support Mini-AERCam Shuttle DTO (CDR-late 2005,
  Delivery to KSC-late 2006) by designing and
  developing a flight management system to perform
  hazard analysis, GNC FDIR and autonomous
  contingency execution

17
Current Status (continued)

• Shuttle Autonomous Rendezvous and Docking (ARD)
  DTO
• Identify requirements and functions needed to
  automate the current Orbiter rendezvous and
  docking capabilities
• Demonstrate and mature ARD technologies to
  benefit SSP and Exploration Program customers.
  This includes advancement of navigation sensors
  and AFM algorithms.
• The SMART Phase 1 prototype is in place for
  future CEV AFM development
• Developing SMART as a pathfinder for
  demonstrating the value of autonomous systems
  will enabled NASA to start reaping the benefits
  inherent to an AFM system for the next human
  spaceflight vehicle.

18
BACKUP CHARTS
19
Autonomous Decision Making and AFM

• Autonomous The crew and vehicle systems perform
  functions independent of the ground
• Decision-making The process of evaluating the
  current situation, extrapolating results into the
  future, and determining the correct path
• Autonomous decision-making The process of
  evaluating the current situation, extrapolating
  results into the future, and determining the
  correct path onboard using both humans and
  automated computer systems
• Autonomous Flight Management System A system
  comprised of flight mechanics functions that
  execute the autonomous decision-making process
  and commands the vehicle GNC system to perform
  the action that has been decided