Using Model-Based Reasoning for Autonomous Instrument Operation-Lessons Learned from IMAGE/LENA

Model-based reasoning has been applied as an autonomous control strategy on the Low Energy Neutral Atom (LENA) instrument currently flying on board the lmager for Magnetosphere-to-Aurora Global Exploration (IMAGE) spacecraft. Explicit models of instrument subsystem responses have been constructed and are used to dynamically adapt the instrument to the spacecraft's environment. These functions are cast as part of a virtual Principal Investigator (VPI) that autonomously monitors and controls the instrument. In the VPI's current implementation, LENA's command uplink volume has been decreased significantly from its previous volume; typically, no uplinks are required for operations. This work demonstrates that a model-based approach can be used to enhance science instrument effectiveness. The components of LENA are common in space science instrumentation, and lessons learned by modeling this system may be applied to other instruments. Future work involves the extension of these methods to cover more aspects of LENA operation and the generalization to other space science instrumentation. Introduction Multiprobe missions are an important part of NASA's future. Consider the missions of the Sun-Earth Connections (SEC) theme, which include missions such as Magnetospheric Multi Scale (MMS, five spacecraft, launch 2006) and the Magnetospheric Constellation Draco (50-100 spacecraft, launch 2010). Members of NASA's Solar Terrestrial Probe line, these missions are part of a series of technologically ambitious projects that build towards the placement of a distributed sensor web that can accurately measure the mesoscale structure and dynamics of Geospace. Geospace is the region of space wherein the Sun and Earth interact to produce Space Weather. To make such missions robust, reliable, and affordable, ideally the many spacecraft of a constellation must be at least as easy to operate as one spacecraft is today. This level of performance is to be achieved in spite of full suites of scientific instruments, limited communication opportunities perhaps separated by weeks, and limited ground operations resources. Downlink bandwidth limitations reduce the coverage and resolution of the science products that missions may produce. Furthermore, U.S. Government work not protected by U.S. copyright. https://ntrs.nasa.gov/search.jsp?R=2001002026

Then we discuss the application of these ideas to LENA followed by a discussion of challenges to generalizing our model-based approach to other instruments. We close with a discussion of future work along these lines. After presenting our approach, we report on the deployed system's performance and the lessons we have learned during the work.

Model-based
Autonomous Instrument

Operations
The focus of this paper is on future autonomy for spacecraft instrument operations. A current study, reported on in this paper, focuses on a model-based reasoning approach to instrument autonomy and its application to the autonomy of the LENA instrument on the IMAGE spacecraft. Figure 1 depicts the major concepts associated with this study. The basic idea is quite straightforward.
The ground-  • Hybrid representation (a combination of the two).
In this phase of the LENA modeling work, use has been very successfully made of the internal representation approach resulting in on-board software automating various LENA functions.
The intent is to graduate to the external representation approach.
The following section will discuss what we have achieved so far in applying this autonomy concept to LENA.   Figure 3).

Model-Based
The VPI provides the capability to respond in real-time to predicted (e.g. radiation belt) and random (e.g. solar storms) conditions.
Actions that can be initiated onboard are consistent with the command authority granted by the ground-based PI ( Table 2).
The VPI is primarily tasked with monitoring and controlling three critical LENA behaviors: instrument overstimulation, high-voltage health and safety and radiationbelt induced collimator effects. Potentially damaging event rates could result from high-flux environments.
They could also be indicative of high-voltage discharges that could degrade electrostatic surfaces and damage electronic components.
In either case, the start or stop channel gains must be reduced to limit the resultant count rates.
Operation of the high-voltage systems is also monitored.
The status of each high-voltage supply is thereby derived. The state of the electrostatic surfaces can also be indirectly inferred since excessive currents or unregulated voltages may be indicative of anomalous conditions on these surfaces.
Control of these behaviors is granted authority level I. represented as a constant range (Figure 4).  This approach compromises science return.
A more robust approach is to react directly to count rates. Gains are reduced only when required.
This approach has the advantage of not only reacting to events that result from the periodic radiation encounters, but to unpredictable energetic particle events as well.
After the VPI has configured the instrument to protect itself in response to a high-rate scenario, it must determine when normal operations can resume.
Since the operational voltages have been reduced, measured count rates cannot be used directly in this determination. Instead, a model of each channel is used to predict when the voltages can be increased  to  nominal  levels  without  violating  an  overstimulation  criterion   An  occurrence  when  the  flight  system  was  overstimulated is shown in Figure 5.     The VPI compares the TOF response with a radiation response each orbit. If this signature is identified, the VPI concludes LENA is within the inner belt, and disables the collimator.
It is re-enabled when the signature is no longer detected.

Lessons
Learned-Development Through Although this approach may be more expensive in the short term, the long-term benefits are substantial.

Full Autonomy
Issues in a Space

Instrumentation Context
Four dimensions along which the path towards near-full autonomy, that can be pursued are: