Science
ACTUAL PROBLEMS OF AVIATION AND AEROSPACE SYSTEMS
Kazan Daytona Beach


Some control problems

for low-orbit satellites of Earth remote sensing

R.N.Akhmetov

State Research and Production Rocket-Space Center "TsSKB-Progress"

The article herein tackles crucial issues of remote sensing satellites motion control in normal conditions and contingencies. Approaches to ensuring survivability on the basis of autonomous control of operation state of onboard systems have been studied.

Maintaining autonomy of an unmanned satellite is of paramount importance at all phases of its life-cycle, and, first of all, it is linked to intellectualization of control procedures. Therefore it is interesting to consider the following problems concerning the Earth remote sensing satellites:

-         satellite autonomous attitude control;

-         autonomous solution to ballistic and navigation tasks;

-         in-flight coordinates calibration;

-         improvement of satellite survivability;

-         development of signature remote sensing;

-         development of criteria for estimation of a remote sensing satellite autonomy.

Pattern of a remote sensing satellite control procedure is in many respects determined by the selected control method of aiming of the observation equipment (sensing equipment - SE) optical axis at observed objects.

Historically, the first control method for aiming of the SE line-of-sight at a target was software-based temporal method when the satellite work program was formed at the ground mission control center and time of imaging equipment actuation as well as duration of its operation were transmitted aboard via the radio channel.

Attitude control of modern remote sensing (RS) satellites is performed by coordinate-temporal method which, on the one hand, assumes assignment of coordinates of survey routes from the Earth via radio link, and on the other hand, autonomous creation of an attitude control program (ACP) for imaging aboard the satellite. All data necessary for ACP autonomous aboard synthesis are also determined and calculated independently, i.e. without support from the Earth. These data include:

-         motion parameters (coordinates, velocities) of the satellite mass center (MC) in the Greenwich coordinate system (autonomous navigation task), time;

-         satellite angular position and angular velocities in the inertial coordinate system (attitude control system (ACS));

-         distance from the Spacecraft MC to a point on the Earth surface to be imaged;

-         prediction of the Spacecraft MC motion and attitude for the moments of imaging.

Note that ACP should allow for the peculiarities of control both on the routes and on inter-route gaps. To provide effective control on inter-route gaps it is advisable to minimize satellite slew time and also time needed to meet the second of the above mentioned limitations.

Satellite attitude and stabilization relative to MC is done in an orbital coordinate system with respective rotation of satellite body axes relative to the orbital ones as per each of the three axes. Modern remote sensing satellites make use of booster guidance, meaning that all calculations are executed in an inertial system of absolute coordinates.

The logic of target acquisition and tracking depends on the adopted method of imaging. There are: target and route imaging, ground area imaging, stereoscopic and random-azimuth imaging. Note that aiming of SE axis-of-sight at the route central line and guaranteeing the required velocity of sighting in the longitudinal direction is performed with programmed pitch and roll rotation; and guaranteeing minimum velocity of optical image crosswise motion in focal plane is performed with a programmed yaw turn.

For autonomous ACP synthesis one should know and predict a Spacecraft MC motion parameters and results of image motion parametric analysis, and that requires solving the ballistic and navigational tasks.

Analysis of Spacecraft flight control shows that ballistic and navigational information is required in general for rapid coordinate-and-temporal referencing of a satellite in the course of its orbital operation. It is performed by integration of GLONASS and GPS sensors in the Spacecraft control circuit thus creating an onboard satellite navigation system (SNS) for efficient execution of the following tasks:

-         navigation measurements according to global navigating satellite system signals and direct navigational definitions of the satellite MC motion parameters (MCMP);

-         refinement of the satellite MCMP by results of statistical processing of direct navigational definitions of the satellite MCMP;

-         updating the satellite MCMP for users of the onboard control system;

-         generation and accumulation of navigating and checking information for downlink to the ground mission control center.

Resurs-DK1 onboard navigational system consists of software installed in the onboard control computer system, and measuring hardware, i.e. onboard time-coordinate synchronizer . Navigating data are generated by means of Satellite motion parameters prediction using its motion model and environment model.

Motion parameters are updated in the onboard control system according to results of SNS periodically with the corresponding interval in order to keep their required accuracy.

Every day of flight the onboard SNS compiles several files with motion parameters for the given moments of time. These files are incorporated in data to be downloaded to the ground mission control center. Minimum and maximum time for statistical processing of direct navigational definition outcomes and, correspondingly, periodicity of motion parameters update in the onboard control system is determined pursuant to:

-         necessity of urgent schedule and coordinate satellite control within a time interval, comparable with one revolution along the orbit;

-         reaching the required accuracy of motion parameters by the beginning of time period allotted for fulfillment of a certain functional task;

-         possibility to conduct sessions for navigational data definition taking into account discontinuity of GLONASS navigating field, operating at the current moment of time of the Spacecraft orbital flight.

Over three years the navigational satellite system built according to the indicated principles has successfully supported fulfillment of target tasks and continuous autonomous operation of Resurs-DK1.

To increase autonomy and survivability of the satellite it is necessary to guarantee durable fulfillment of navigational tasks in conditions of partly deployed GLONASS system and possible local malfunctions of GPS system.

Besides, usage of radio navigational fields of these systems does not contribute to high-precision positioning of a spacecraft. Therefore, more challenging is creation of integrated inertial-satellite systems able to cope with navigational and positioning tasks.

This may be feasible in case of multipurpose utilization of measuring information received from inertial sensors, star trackers and global navigational satellite system. Inertial sensors and star trackers are the basic sensing elements of modern Spacecraft motion control systems.

At the present stage of remote sensing satellites development a contradiction appeared between attained pinpoint accuracies provided by the onboard hardware (seconds of arc) and mutual misalignment of their sensitivity axes when operating as a part of a satellite, attaining several minutes of arc that results in reduction of the satellite orientation accuracy and aggravates quality of the target information.

It is not obviously possible to eliminate such a contradiction barely by certification of the onboard hardware in ground environment. The effective approach to this problem consists in geometrical coordinate calibration for the onboard hardware (including imaging equipment) directly during satellite orbital flight.

We consider in-flight coordinate calibration for the onboard hardware as a verification process performed aboard satellite aimed at detection and estimation of onboard hardware intrinsic errors including installation errors of imaging equipment and onboard segment of the attitude control system, their mutual alignment and certification, with the subsequent consideration of these ambiguities while controlling satellite attitude in real time.

Our company has been studying scientific and applied aspects of in-flight calibration of satellite onboard hardware since the 80s of the last century, in order to reduce deviations of the satellite strap down attitude control system based on standard measuring instruments (sensors) of its angular velocity. There were developed calibration procedures for other gyroscopic measuring devices using sensors acquiring external information, in particular, dynamic star trackers. In modern remote sensing satellites, hardware-algorithmic structure of the attitude control system has undergone considerable changes that have demanded new approaches to the in-flight calibration.

Due to constant toughening of quality requirements on imagery data received by modern remote sensing satellites, increased attention is paid to calibration of the onboard equipment. Research guidelines in this sphere are the following:

-         calibration of attitude control system instruments (namely, EBG, SLSST, FOARM gyroscopic and star trackers) directly by developers at the ground special-purpose test-benches;

-         definition of the satellite configuration and implementation of corresponding design and technological procedures allowing for correct calibration (e.g. placing of sensitive elements of an attitude control system on a common platform - thermo stabilized plate, imaging equipment base frame, mutual alignment of the instruments' base axes, etc.);

-         calibration of onboard hardware during the orbital flight performing special calibration procedures and specification of models and algorithms hardwired in the instruments' processors;

-         development of special angular maneuvers for satellite calibration and estimation of errors of attitude control system and imaging equipment that differ in their nature;

-         development of special onboard systems and certain calibration modes that allow combined operation of the attitude control system and imaging hardware;

-         development of algorithmic support for in-flight calibration procedures via virtual attitude control system devices (i.e. their algorithmic models);

-         development of in-flight calibration techniques implementing regular angular maneuvers performed by satellite in the process of its functioning.

There were considered techniques and results of EBG in-flight geometrical coordinate calibration performed during Resurs-DK1 flight qualification test. Such calibration is carried out by stellar monitoring and taking into account a corresponding deviation mathematical model (DMM) adapted for conditions of orbital flight and developed at EBG design stage. DMM coefficients were defined at EBG development tests. DMM and corresponding coefficients are stored in memory of the EBG central computer and underlie the calculation algorithm of gyro deviation during an orbital flight. Peculiarity of this procedure is the following: during calibration and reference orbital revolutions the satellite angular position is simultaneously measured by EBG and STC, the obtained data are down linked via telemetry to a Ground Mission Control Center and after processing are up linked to be used during satellite real-time attitude control.

At stellar monitoring and adjustment of EBG gyros measuring axes position, the data are used which were obtained from external orientation sensors (that is from onboard STC star coordinators) as reference values of satellite angular position. However, accuracy of EBG measuring axes positions relative to inertial coordinate system is not the same for different axes: accuracy of optical axis positioning (direction perpendicular to the sensor placement location) is by an order of magnitude higher than the positioning accuracy of two other axes. Required accuracy of satellite attitude position may be reached if the information is used from two synchronously operating STC, mounted aboard satellite at different angles. But the period of time when synchronous operation of two sensors is possible, is limited as well (due to the Sun, the Moon, the Earth overexposure).

Some works deal with calibration of the outer positioning sensors, STC in particular, in a novel way using generation and implementation of a "virtual" device (STC mathematical model). It guarantees accuracy of the Satellite attitude position definition corresponding to the accuracy of two synchronously operating STC. In this case, having defined discrepancies in readings of the real and "virtual" STC, it is possible to control satellite by measurements of a single SLSST. Note that the number of periods for which it is possible to have two synchronously operating STC is considerably below the number of periods when one STC may operate. Such an approach allows to consider errors of STC sensitivity axes set-up relative to the Satellite axes and start a quasi-continuous control of the satellite based on STC measurements.

Mathematical aspects of the in-flight geometrical coordinates calibration based on data from both the space telescope and star tracker system are considered early. Analyzing heritage of Resurs-DK1 it is possible to conclude the following:

-         it is necessary to provide an automatic onboard calibration of EBG without on-ground processing of telemetry data; that will essentially raise efficiency of calibration and will increase productivity of a satellite;

-         at designing of new satellites and in order to reduce offset of EBG and STC measuring axes it is expedient to provide possible installation of these sensors on a common platform.

The essence of geometrical coordinate calibration for the onboard hardware of attitude control system and imaging equipment mounted on a uniform platform (main frame or thermo stabilized plate) consists in the following. SLSST directly determines satellite position (or SLSST instrument axes) relative to the inertial coordinate system. By means of FOARM (or EBG) the satellite position is determined at the intervals between stellar monitoring by satellite angular velocity measured relative to FOARM sensitivity axes (by integration of the known satellite motion equations at the initial angular position measured by SLSST).

In the case under concern two modes are analyzed. These are: the so-called mode of stellar monitoring and matching of star tracker imaging equipment axes (SMMA) when position of their measuring axes is determined by stars. And there is a mode when the current position of the main axes of platform/imaging equipment is controlled by means of an autocollimation measuring system and matching of axes (ASMA). The SMMA technique incorporates two phases. The first phase is implemented directly during the satellite flight and includes imaging of celestial map simultaneously by star trackers and imaging equipment, i.e. an optron telescope system. The second phase implies combined on-Earth processing of celestial imagery data acquired by star trackers and optron telescope. Here the data received by ASMA technique are also considered. The ASMA technique guarantees matching of the attitude control system onboard hardware platform and optron telescope optical axes with an error of 3 seconds of arc.

However, the time interval from the moment of reception (measurement) of data up to their application turns to be considerable; that leads to unsuspected mismatch of the onboard equipment axes. The problem to be solved is elimination of this interval transferring this task aboard the satellite.

On completion of SMMA mode the orientation of optron telescope base line and star trackers base line for a certain instant of time relative to inertial base line I turn to be the known values. Orientation of the platform (P) relative to optron telescope base line is determined by results of ASMA mode. On the basis of SMMA mode a current relative position of the optron telescope and star trackers base lines was calculated.

Position of attitude control system onboard equipment relative to optron telescope optical axes may be computed basing on data received in SMMA modes; whereas variation of the optron telescope and star tracker current position is estimated by means of ASMA. The acquired data are used in the attitude control system to control satellite position and promote the higher quality imagery.

The strategy of survivability improvement includes the following:

-         in-flight on-board monitoring of satellite functioning;

-         effective testing of satellite systems fitness;

-         compilation of an onboard database of contingencies;

-         autonomous analysis of satellite systems status in contingencies;

-         compilation of onboard correct reference points database;

-         autonomous rollback to reference points in contingencies;

-         autonomous rearrangement of controls to recover satellite functions;

-         balance of centralized /decentralized control principles;

-         implementation of standard functional patterns (modes) of the satellite onboard tools in contingencies;

-         autonomous calibration of the satellite onboard hardware.

Signature imaging is remote sensing with imagery data preprocessing aboard a satellite before its transmission to the Earth; consequently only a part of imagery data is transmitted which demonstrates changes in the state of targets if compared to previous instant, or it may not be transmitted at all if no changes were revealed. Such an approach allows to transmit only useful information and not to stuff the downlink channel with useless data. On the other hand, it demands availability of a dedicated system aboard the satellite which would allow real time determination of changes in the status of targets and other elements of the scene under observation.

An object signature is considered to be an object characteristic (descriptor) with the help of which the object can be found out, discriminated, classified and identified. As a rule, it is not enough to have only one characteristic, therefore an object should have the whole set (tuple) of signatures. A tuple of signatures characterizing an object and used for its detection, discrimination, classification and identification is called a pattern (template).

Thus, signature imaging is obtaining of information on signatures of real targets, their comparison with the earlier accumulated and stored in the knowledge base signatures (templates) of objects, discrimination of objects, determination of changes if any, and decision making about transfer of the appropriate information to the user. The user can receive either complete information on the observed scene (objects, background conditions, obstacles), or partial information about changes (signatures, parameters), or a message on absence of changes, or no messages at all.

There are geometrical, time-space, spectral signatures, and also power, dynamic and fractal signatures.

Development of onboard autonomous systems of signature imaging requires a great many technological aspects to be properly resolved, namely:

1.         Compilation of an onboard knowledge base including:

-          a priori spectral characteristics of objects, backgrounds, obstacles (background - target - obstacle environment) for various spectral bands, observation conditions (flight altitude, sun aspect angle, observation angle, atmospheric status, and contrast), modification of the optron system parameters and recognition system parameters;

-          methods, algorithms and criteria for automatic spectral selection of objects in various background-target-obstacle environment and observation conditions;

-          minimum necessary tuple of standard reference signatures sufficient for detection, recognition of the object class, type, size, structure, status, which on the one hand, supports the required quality performance of the onboard autonomous system of signature imaging, and on the other hand, does not complicate its design (ensuring specified probability of recognition and fitting restrictions on weight, power consumption, fabrication cost and system maintenance).

2.         Software for the onboard autonomous system of signature imaging (operational environment, reduction of message redundancy and compression of imagery data, decision making on imagery data modification and transfer to the ground-based imagery data reception and processing center, interface with the satellite onboard / ground-based control system);

3.         Imagery data receivers (multispectral optron systems, hyperspectrometers), onboard special purpose computers.

4.         Ground-based center of the onboard autonomous system of signature imaging optimization, test and support.

5.         Onboard and ground-based training systems with self-correction, self-adjustment and self-training elements.

So far criteria estimation for autonomy of a remote sensing satellite and its onboard systems has not been duly covered in scientific research. Below some approaches are discussed to determination of autonomy criteria.

Criterion of power autonomy.

The criterion is characterized, on the one hand, by an autonomous set of devices and equipment necessary to maintain the required energy profile of the satellite: chemical power sources (nickel-cadmium, nickel-hydrogenous, lithium-ionic accumulator batteries), photovoltaic cells (silicon with 14.5% efficiency, arsenide gallic with 26-28% efficiency), solar arrays, power supply systems, automatic voltage stabilizers, and on the other hand, onboard software which activates autonomous modes of power balance maintenance. These are: solar arrays deployment technique (two degrees of freedom), special modes of satellite turning with the purpose of power accumulation and reaching max cosα, maintenance of a power balance and fault recovery in power supply systems.

Criterion of functional autonomy.

This is determined as relation of a scope of tasks competed by an onboard control system to the total amount of tasks fulfilled by the satellite autonomous control system (onboard and ground-based control system). Such distribution of tasks became possible due to transfer of attitude control, satellite MC motion control, satellite control in various modes, and great number of testing and diagnostic tasks aboard the satellite.

Criterion of informational autonomy.

From the informational approach point of view, two different circuits should be discriminated: imagery data circuit and satellite/orbital constellation control loop. Let us discuss the first one.

Imagery data circuit may be considered in two aspects: in respect of creation, accumulation, compression and down linking of imagery data; and in respect of improvement of imagery data descriptiveness (resolving capability, ground resolution). As for the first aspect, at first sight it may seem that now this circuit operates in autonomous mode. However, not everything has been done here. For example, acquired imagery data require ground-based processing, namely - matching of imagery fragments. It is not absolutely acceptable for users who need real-time information. Besides, currently signature imaging technique is under research.

As for improving of imagery data descriptiveness, special imagery restitution methods shall be implemented for denoising (for example, elimination of smearing effect).

Satellite and orbital constellation control tasks are distributed between the ground control system and the onboard control system. They interact via information flows between mission control center, receiving station and satellite, satellite and special purpose center. Here, besides transmission of control operations during communication sessions and operating programs, for arrangement of onboard and research hardware functioning and single commands, the telemetry information and housekeeping data are transmitted. Autonomy in this case is possible only when informational interchange between ground-based and onboard control systems fails by some reason. In case of telemetry data lack, necessary data may be obtained together with the housekeeping data. If it is impossible to uplink data about the scheduled areas to be imaged, it is feasible to have satellite operating completely autonomously, i.e. imaging the previously scheduled regions.

Criteria of in-built self-integration level.

Control system response to modifications normally leads to consequences suppression. Therefore it is necessary to develop control methods responding not only to modifications, but also to the rate of these modifications. Here comes the necessity to develop self-organizing systems. In the case under consideration it is possible to have structural, engineering, parametric and probably informational self-organization, autonomous reconfiguration of onboard systems, functional and control processes. It is especially urgent in case of anomaly, when it is necessary to maintain integrity and continuity of functional tasks solution.

Numerically this criterion can be determined by various indices, namely:

-         coefficient of satellite operational readiness to execute basic functional tasks;

-         minimal time required for a satellite to perform restorative function in case of anomaly;

-         quantity of possible structures and ways of reconfiguration of satellite systems;

-         relative volume of onboard software realizing satellite control pattern in case of anomaly;

-         number of descriptors (faults in the satellite hardware) necessary for switching from nominal flight to anomaly mode;

-         amount of checking information about satellite functioning sufficient for application of "reverse engineering" during analysis and recovery of the satellite functionability;

-         minimum intensity of satellite switching from nominal flight to anomaly mode;

-         decrease of the ratio of severe faults to the total number of failures.

To formulate an integrated criterion for estimation of system autonomy it is necessary to standardize all partial criteria, e.g. to make them dimensionless. Then the integrated criterion may be determined by any known way (e.g. additive or multiplicative convolution of partial criteria, distance estimation between alternative under consideration and its ideal representation).

Criteria estimation of autonomy by a numerical method (criteria convolution) is problematic, as observed criteria are of different nature and possess multifactor indeterminate form. In this case application of artificial intellect and, in particular, fuzzy logic methods where not exact numbers but loose linguistic variables are used, is challenging.

The paper covers a wide range of problematic aspects regarding autonomous functional control of low-orbit satellites, in-flight coordinate calibration for the attitude control system, onboard signature imaging systems, adaptive autonomous remote sensing space systems with intelligent control. The approaches are suggested for determination of autonomous navigation criteria.

Not all the aspects are comprehensively studied. Many of them require more thorough theoretical and experimental research, involving some dedicated companies.

 

 

Ravil Nurgalievich Akhmetov, First Deputy Director General (FSUE SSP RSC "TsSKB-Progress"); Chief Designer, Head of "TsSKB-Progress"); Candidate of technical sciences; Full Member of K.E.Tsiolkovskiy Russian Academy of Cosmonautics, Full Member of the International Academy of navigation and motion control, Corresponding Member of the Academy of Technological Sciences. His professional interests: design, control and survivability of complex rocket-and-space systems.



[Contents]

homeKazanUniversitywhat's newsearchlevel upfeedback

© 1995-2008 Kazan State University