Anyone who has worked in the aerospace industry for even a few years knows that the industry has a lot of ups and downs. One year your company is “THE COMPANY” for capability X. The next year capability X is out of fashion or saturated. Aerospace companies are not traditional manufacturing companies but service companies who provide hardware as well as a final report. The service they provide is the necessary expertise to push the envelope – push the envelope on capability, on reliability, or on price.

The ups and downs of the aerospace industry make staffing decisions difficult and risky. When you wish to pursue new areas of technology or new markets you need new people. But with salaried employees come long-term commitments and what to do with these new experts while you wait on a contract award?

No commitments…

As a consultant I provide your company with expertise when you need it and only when you need it. I come in for the proposal and I leave while you wait for a contract award. Assuming you win, I am happy to come back for program execution but there are no hard feelings if you decide to hire a full time salaried employee instead. I’m available when you need me, for the projects you need me and no more.

Whether you are discussing a spacecraft or aircraft the platform will have rigid and flex body motion. The rigid body motion is the motion you learn to describe in undergraduate kinematics and is typically straight forward to understand. The flex body motion will add additional uncertainty to your sensor readouts.

Experience on several programs has provided ControlTheoryPro.com, LLC Consulting with the following LOS skill set:

• Alignment Acquisition and Stabilization
• Controller Design (for steering mirrors, gimbals, etc.)
• Jitter Suppression (through feedforward commands, Outer/Inner loop control, and Inertial Stabilization)
• Modeling and Simulation (MATLAB and Simulink)
  • Sensors
    • Position Sensors (DITs, Resolvers)
    • FPA
    • Custom Sensors
  • Mechanisms
    • Steering Mirrors
    • Gimbals
    • Constant Rate Spin Mechanism
  • Design of basic (PI controller) and optimal controller design
• Inertial Stabilization
• Sensor Fusion
• Stability Analysis
  • Conservative stability margins are a phase margin of at least 40 degrees and a gain margin of at least 6 dB

Sensor fusion is used when no single sensor can do the job. In the aerospace industry it is not uncommon that no single sensor exists that can accomplish the task at hand over the entire required frequency range. When no single sensor can accomplish the necessary feedback task then multiple sensors are blended to create a single feedback signal.

The simplest form of sensor fusion is a matter of two or more sensors which are filtered so that their strengths (good responsivity and low noise) are used while their weaknesses are filtered out. Often times sensor fusion is nothing more than simple second order low pass or high pass filters with their outputs added together. While sensors are often blended with simple filters more complex Linear and Non-Linear Kalman filters are sometimes required.

ControlTheoryPro.com, LLC Consulting brings several key skills to solving your Sensor Fusion problem:

• Experience with sensor blending for Laser Communications and Inertial Stabilization
• Experience with both traditional and MEMS based sensing technology
• Experience design and validating sensor fusion filters
  • Model and analyze before you build – even small phase and magnitude errors can create large problems

Simple Sensor Fusion Example

A simple sensor fusion example is available on ControlTheoryPro.com’s educational website: Sensor Fusion Example.

Guidance, Navigation & Control (GNC) or Attitude Determination and Control (ADC) is a difficult subsystem to design and manage for any spacecraft. During my time working GNC for the International Space Station (ISS) we were busy trying to determine the cause of a Control Moment Gyroscope (CMG) failure and prevent any more from failing. Picking an anomaly out of a data set with something like 20 different significant contributors was nearly impossible to do with the naked eye. So a Kalman Filter based approach was created for anomaly characterization.

On the Attitude Determination side of the GNC equation Kalman filters are used to blend a number of different sensors measuring different things at different rates. It is not uncommon for a spacecraft to have a star tracker with approximately 1 Hz for an update rate combined with gyros or accelerometers updating at up to 1 kHz. The Kalman filter blends all of these signals into a coherent attitude for your GNC subsystem.

The GNC System Definition Process is as follows (Ref. 1):

1. Define navigation and orbit-related top-level functions and requirements
2. Do pointing and mapping trades to determine preliminary navigation (position) accuracy requirements
3. Determine whether orbit control or maintenance is needed
4. If yes, do trade on autonomous vs. ground-based orbit control
5. Determine where navigation data is needed
6. Do autonomous vs. ground-based navigation trade
7. Select navigation trade
8. Define GNC system requirements

Guidance, Navigation & Control Mission Types

There are several different mission types and each type creates different requirements for the Guidance, Navigation & Control subsystem.

• Orbit Maintenance
  • Maintain orbital elements
• Altitude Maintenance
  • Maintain altitude to prevent orbit degradation
• Stationkeeping
  • Keep a spacecraft within a specified box
• Constellation Maintenance
  • Keep a spacecraft within a specified box which is moving, i.e. the box needs to be defined as a position relative to another spacecraft in the constellation

References:

1. Larson, Wiley, and James Wertz. Space Mission Analysis and Design. Torrance: Microcosm Press, 1999.

A thorough understanding of a given system can only be achieved through concerted effort; often times only through modeling do we truly understand the interactions of each piece. Tracing signal paths, with appropriate units and coordinate systems, provides for a more in-depth appreciation of all aspects of the system and can inform other teams such as embedded software (i.e., number of filters, coordinate transformations, or unit conversions that need to be performed by software).

ControlTheoryPro.com, LLC Consulting brings the following skills:

• Non-Linear, Discrete and Continuous Domain Modeling
• Automation of Time Consuming MATLAB and Simulink tasks
• Automated report generation (mostly, flexibility requires that it can’t all be automated – yet)
• Frequency and Time Domain Modeling (MATLAB & Simulink)
  • Frequency Domain models are linear models built for a quick turn-around
    • Useful for exploring architecture trades
    • Useful for identifying and quantifying the individual contributors to the final residual LOS error (jitter)
  • Time Domain models are non-linear models built for simulating the effects of non-linear system aspects
    • Examples of Non-Linear System Aspects: Analog-to-Digital Conversion, Mode Transition (sensors and controllers may change)

Control Systems Modeling: Block Diagram Basics

The image above is the basic model of a feedback control system. Each piece plays a part. I’ve described the various transfer functions below for the uninitiated.

Feedforward Transfer Function

This is the system response should the system lack or lose feedback.

Open-Loop Transfer Function

The open-loop transfer function designating the response of all components after the error signal is created. The Nichols and Nyquist plots are generated from open-loop transfer function data. And the gain and phase margin are determined using the open-loop data – preferably the margins are derived from a Nichols or Nyquist plot rather than from the open-loop bode information.

Closed-Loop Transfer Function

The closed-loop transfer function designates the response of the system when it is closed around a feedback sensor. The shape of the closed-loop transfer function will tell you how well the system will follow your commanded inputs (r).

Disturbance Rejection Transfer Function

The disturbance rejection transfer function shows how well the system, under closed feedback control, will reject system disturbances (at d_o). This is especially important to LOS, Directed Energy, Imaging, and Inertial Stabilization applications.

Proposals Require a Fast Response

The time frame between RFP release and final submittal is almost always too short. Even when you know an RFP is coming out the details can change at the last minute. Your proposal teams need flexibility but proposals can also be a distraction to your often over burdened staff.

When you need extra people to tackle that Line-of-Sight (LOS), Directed Energy, Inertial Stabilization, or vibration control proposal effort contact ControlTheoryPro.com, LLC Consulting. The temporary controls and modeling expertise could be the difference between winning and losing that proposal. Also, you don’t have to overtax your staff (risking burnout) or worry about what to do with that engineer while you wait to find out if you’ve won.

Key Proposal Skills

Experience with several recent proposal efforts and automation efforts over the past several years have led to quick turn-around, high quality models, analysis and graphics. Also, experience with the follow-on programs has led to a modeling style with ample, clear, and thorough commenting throughout the model code as well as logical model code organization. Don’t underestimate the value of proper coding practices; poor coding practices can make the model nearly useless to anyone but the author. The models are developed to be handed off to another engineer or picked up easily months later after a contract win.

Speed is of the essences with proposal efforts. Often multiple design iterations are required in a span of just a few weeks. That is why all ControlTheoryPro.com, LLC Consulting models are created with as much automation included as possible. As the design changes the amount of model rework is minimized. For example, graphics (system, results plots, etc.) for PowerPoint and Word are typically generated automatically in a well formatted manner. Code is developed to identify and create the majority of desired plots so that most plots are automatically generated no matter the system design; the result is that the final report is largely a matter of culling unnecessary graphics and adding commentary to the results.

Finally, having worked several proposal and programs which incorporated LOS controls, Directed Energy, and Inertial Stabilization means a minimal amount of time spent “coming up to speed”. This is crucial because then the right design questions can be asked and answered early in the proposal process preventing the need for design changes late in the proposal effort.

In Aerospace applications such as Laser Communications we are provided with an initial Line-of-Sight (LOS) pointing vector to a target. The initial vector has a certain amount of uncertainty to it and thus the system must start with a wide view of the target area and refine it’s pointing down to the true target position. However, during the initial stages when the position uncertainty of the target is large the system disturbances (such as spacecraft platform motion) can introduce a substantial error between the desired LOS vector and the system’s true LOS. Inertial Stabilization allows the system to point at the initial LOS pointing vector despite disturbances; this pointing is maintained until a return signal from the target can be acquired (a Tx beam from the target or a reflectance) for feedback purposes. ControlTheoryPro.com, LLC Consulting brings the following crucial skills to your Inertial Stabilization program:

• Artificial Star
  • Using a Fast-Steering Mirror as a platform inertial sensors are used to measure the mirror’s motion and controller/compensator is designed to counter this motion
• LOS Stabilization (Laser Communications, Directed Energy)
  • Prior to target acquisition (and a feedback signal) the system must maintain a pointing vector such that the LOS is centered on the best available target position; maintaining this position requires a feedforward implementation
• Inertial Stabilization Simulation (Simulink)
  • Since sample rate and computational delay time are crucial the utility of a, purely linear, Frequency Domain model is limited. Therefore modeling and analysis is primarily done through Time Domain Simulation
  • Time Domain models are non-linear models built for simulating the effects of non-linear system aspects
    • Examples of Non-Linear System Aspects
      • Analog-to-Digital Conversion
      • Computational Offsets
      • Mode Transition (sensors and controllers may change)

• LOS Pointing Control (Jitter Suppression/Disturbance Rejection)
  • High-Accuracy applications such as Laser Communications (residual jitter of approx. 1 microradians)
  • Medium-Accuracy applications such as Imaging and Directed Energy (residual jitter of 10s of microradians)
• Basic Optical System Architecture
  • Experience with 2 Directed Energy programs (MLD, HEL-TD)
  • Experience dealing with issues such as Beam Walk and optical bench disturbances
• Frequency and Time Domain Modeling (MATLAB & Simulink)
  • Frequency Domain models are linear models built for a quick turn-around
    • Useful for exploring architecture trades
    • Useful for identifying and quantifying the individual contributors to the final residual LOS error (jitter)
  • Time Domain models are non-linear models built for simulating the effects of non-linear system aspects
    • Examples of Non-Linear System Aspects: Analog-to-Digital Conversion, Mode Transition (sensors and controllers may change)
• Inertial Stabilization
  • Using inertial sensors such as gyroscopes and accelerometers a feed-forward architecture can be designed to hold the LOS inertially stable