Controls & Computing

Learning Outcomes

Get to know issues and options of On-Board Processing

Understand system design characteristics of software

Understand architecture types and also modern implementations

Discuss software development challenges

Failures

💥

Top 3 Failures: Software, Voltage, Launch

Whats the commonality between Beresheet, Mars Climate Orbiter, and Astro-H?

They all malfunctioned because of software issues:

Beresheet (Israeli Moon Lander)

• A sensor malfunction triggered a software command
• This command unexpectedly shut down the main engine
• The lander crashed due to lack of engine power during descent

Mars Climate Orbiter

• Ground software used imperial units instead of metric
• This mismatch caused incorrect trajectory calculations
• The spacecraft entered Mars’ atmosphere at the wrong angle and disintegrated

Astro-H

• Faulty software misread the satellite’s rotation
• Attitude control system overcompensated, causing a spin
• Incorrectly configured thrusters worsened the spin, leading to breakup

🛠

That’s why we need good design and thorough testing

Computer Systems

Think of mission computer systems of all components, from the spacecraft itself, to development tools and testing systems.

Mission Computer Systems have many interfaces. Managing these interfaces and keeping them interoperable uses a lot of resources.

⚠️

Flexibility

Having a flexible system has its advantages, but it’s also is usually much more complex than something that does only one thing right

Below you can find a hierarchy of the three pillars, Hardware, Software and Documentation:

Hardware

• Hardware Configuration Item (HWCI)
• Computer Board
• Computer Chip Set/Analog Devices/Logic Components/Discrete Components

Software

• Computer Software Configuration Items (CSCls)
• Computer Software Components (CSCs)
• Computer Software Unit

Documentation

• Requirements Specification
• Design Documents
• Detailed Design Documents
• Interface Control Documents (ICDs)

Design Drivers in Computer Systems

⚠️

When flowing down mission requirements, we must be careful to design hardware and software with the “ilities” in the fourth column in mind

Mission Requirements	System Level Processing Requirements	Computer Level Requirements	Additional Requirements (“ilities”)
- Customer Needs

• Expected Availability
  • Weeks
  • Months
  • Year or more
• Number of Satellites
• Number and location of ground stations
• Level of Autonomy
• Security Requirements
• Programmatic Issues
• Cost
• Schedule
• Risk | - Functional Capabilities
• Processing Partitioning
• Payload vs. Spacecraft
• Onboard vs. Ground
• Physical Characteristics
• Size
• Weight
• Power
• Radiation
• Communication Protocol
• Commercial Digital Standards
• Commercial Analog Standards
• Protection / Encryption | - Throughput
• Memory
• Radiation Hardness
• Development Tools
• COTS Software availability
• Emulator / Engineering Model availability | - Testability
• Feasibility
• Usability
• Reusability
• Reliability
• Flexibility
• Maintainability
• Interchangeability
• Replaceability |

System Requirements for Computer Systems

We can approach the system requirements definition by asking the following questions:

Questions to Ask	Parameters to Review
What Must The System Do?	Evaluate and establish basic functional requirements.
Why Must It Be Done?	Establish traceability from functions to mission objectives. Be sure to challenge the requirements and assess their validity.
How Can We Achieve It and What Are The Alternatives?	Evaluate candidate architectures including network topology, data flow, control flow, and UML diagrams. Investigate emerging technologies and perform trade studies as needed to ensure a robust solution. A combination of traditional architectures is an acceptable “hybrid” solution.
What Functions Can We Allocate to Which Parts of the System?	Perform functional partitioning to development block diagrams of ground station and vehicle, as well as spacecraft and payload.
Are All Functions Technically Feasible?	Determine if the value of state-of-the-art technology outweighs the risk. Evaluate the potential use of off-the-shelf hardware and software. Look for data flow bottlenecks and reallocate functions to evenly distribute the data flow. Review baseline block diagrams for potential holes.
Is The System Testable?	Develop non-intrusive test plans and procedures that ensure the system will meet mission objectives. Are test points available outside the system for easy “black-box” testing?

Computational Architectures

Centralized Topology

In a centralized topology, units connect directly to a central hub or management computer, offering high reliability for a limited number of well-defined subsystems.

However, it requires extensive wiring and makes adding new components complex, needing additional wiring and software updates.

Ring Topology

A Ring Topology allows data to travel around a ring, delivering the same information to multiple nodes. It uses smaller, distributed wiring harnesses, ideal for spacecraft layouts.

However, each transmission relies on the previous node, which can be a limitation if a node fails.

Testing Computersystems

🛰️

You can use flatsats or testbeds that fully integrate sensors, hardware interfaces and allow for testing “as system.” This makes it possible to inject all nominal and off-nominal data at interfaces.

Test bed persists after launch and becomes ground-test opportunities before uploads into space

What is testing: Testing starts with low-level software (unit, module) and hardware (component, acceptance) tests, progresses to stand-alone functional tests for ground- and space-based systems, and culminates in system integration, testing, and on-orbit calibration, verifying mission requirements.

Key Characteristics and Challenges

• Traditionally, there is limited power and limited mass. Sometimes, also size limitations
• Speed of processor limits/enabled options of computational system
• Size and type of storage
  • Size of instruction memory
  • Size of storage
• Most Earth-based storage capabilities are not hard

Key Computation Options

1. Spacecraft Ephemeris
   • Interpolation of Hermitian Polynomials
   • Onboard propagator using Doppler shifts
   • GPS signals
2. Star trackers
   • Digital trackers outputting full quaternions
   • Picture-based with onboard catalog
3. Filesystems
   • None (for simpler systems)
   • Fully fledged (for complex missions

Software Issues and Challenges

• High labor costs
• Delayed testing in integration flow
• Shortage of skilled software engineers
  • Continuous workload post-launch
  • Poor cost estimation

🚨

Spacecraft design often focuses on hardware, leading to reactive software development

Software Architecture Options

Layered Services

• Standard API for each layer and service
• Encapsulated implementation details
• Allows for internal changes without affecting other layers

Bus Architecture

• Standard-sized, message-oriented communication system
• Minimizes interdependencies
• Enables independent testing and “plug and play” capabilities

Core Flight System (cFS)

• NASA-developed reusable software framework
• Modular architecture for scalability and portability

Controls + Computing Summary handwritten

https://drive.google.com/file/d/17y8FClK89jioalq_K4p5-qYetWLbv9Em/view?usp=sharing