Enhancing Multi-payload Launch Support with Net-centric Operations 1. Introduction Andrews, S.E., Bougas, W. C., Cott, T.A., Hunt, S. M., Kadish, J.M., Solodyna, C.V. MIT Lincoln Laboratory There has been a long history of academic, small scientific and small commercial efforts in space. Until the mid-1990s, however, use of space was dominated by military/government, large science, and large commercial efforts. Access to space was limited and costly, and the high cost of designing for robustness and reliability was often balanced by extensive re-use of designs. Even multi-payload launches typically involved launches of satellites that were multiple copies of the same basic design The Cosmos/Strela launches are one example of what this paper will term a traditional launch. Traditional launches are characterized by single satellites launched on one booster or multiple identical satellites launched on one booster and then distributed in mean anomaly. In this case, 6 store-and-dump communications payloads were launched together. [1] The JAWSAT launch of 2000 demonstrates a new type of common launch in which numerous, distinctive payloads are launched on a single booster. As shown in Figure 1, the payloads are a variety of shapes and sizes. Even the differences in the developer organizations can be noted due to the different units used to specify sizes. One of the key aspects of this launch was the need for radar tracking of the OPAL payload to support communications with that payload. The operators required good element sets within days to accomplish the primary scientific goals related to releasing the Aerospace picosats. [2] Launches such as this pose a challenge to space surveillance operations tuned toward support of traditional launches. Several factors have contributed to changes in satellite launches. First, as use of space matures, experience with and technology for launch has led to an overall reduction in risk and effective cost. This includes launch facilities as well as rockets. Second, electronics, materials and manufacturing technology have led to the ability to create smaller, cheaper and more capable components for satellites at a lower cost. Computing and communications capabilities have also been enhanced. All of this technology has been pushed down to academic and small-scientific payloads of differing shapes, sizes, functions and capabilities. The following are some of the key features of the modern scientific small-sat launch: a. Multiple, distinctive payload shapes and sizes b. Slow separation velocities c. Reliability and lifetime traded out for cost/weight/size d. Limited or minimal ground support for payload operations This work sponsored by the Air Force under Air Force Contract No. FA8721-05-C- 0002. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the United States Government.
ASUsat 24.5 cm x 32 cm diameter (14 sides) Opal 23.5 x 21 cm (6 sides) FalconSat 17 x 18 inch box OCS 3.5 meter diameter JAWSAT 35 x 35 x 42 inches [2] Figure 1 Payloads from JAWSAT Launch The JAWSAT launch provides a good example of these features. The separation velocities in the first Minotaur launch ranged from 0.15 m/s to 0.8 m/s [3]. The first two deployed picosats were designed for very short lifetimes and reached too-low power within fifteen days of launch. Communications with the OPAL payload were limited and not very robust, despite the requirement for time-critical support [4]. To conduct space surveillance operations under the approach used for traditional spacecraft, networks were dealing with medium to large spacecraft and periods of weeks to months to for the spacecraft to become operational. The communications capacity available for dissemination of space surveillance data limited the plans for data distribution. The available sensors include powerful radars and high-quality optical systems. In these cases, using metric data to separate objects was sufficient and efficient. To conduct space surveillance with the modern small spacecraft used for science and technology research, sensors are dealing with spacecraft of sizes on the order of the wavelength of a satellite tracking radar. Separation velocities are sub-meter/second, yet good identifications and orbits required are required by payload ground stations within hours or days. This is in the face of multiple payload types along with similar-shaped objects with different missions and owners. Sensors with good sensitivity and metric data are required, but net-centric operations that are correctly employed can enable a space surveillance network to make better use of the limited number of capable sensors available to it. This paper proposes several concepts of net-centric operations that will enhance the ability of a space surveillance network to deal with the launches of small multiple payloads. Section 2 provides a brief explanation of net-centricity and an overview of
options. Section 3 describes the challenges of the exemplar launch. Section 4 describes the net-centric operations concepts for addressing these launches. Section 5 reviews how some of these concepts were applied to the JAWSAT launch. Section 6 discusses a potential path ahead for the MIT sensors and Section 7 provides a summary. 2. Overview of Net-centric Operations Options The concept of net-centricity is provided in this definition drawn from a report on netcentric warfare made by the Department of Defense to the US congress.- Links sensors, communications systems and weapons systems in an interconnected grid that allows for a seamless information flow to warfighters, policy makers, and support personnel. [5] The following paragraphs summarize some concepts for applying net-centricity to space surveillance. Foster simultaneous cooperative operations through sharing of real-time data. For example, by sharing pointing information in real time, net-centricity can enable simultaneous tracking of objects with poorly known orbital elements. With direct sensorto-sensor hand-off of tracking data, there is positive control that the two sensors are tracking the same object. Figure 2 Simultaneous operations and direct sensor hand-off Provide contextual data to support rapid identification and tracking of objects of interest. For example, element sets based on early tracking data of one object may have sufficient uncertainty that another object in the vicinity may be accidentally acquired and assumed to be the object of interest. Information such as relative order of objects in a train and signature data can help sensors, in early stages, to collect on the desired object. Passing of relative order information on objects can help to assure consistent tracking of a specific object when absolute metric data has errors. Often relative position can be distinguished and captured even when the absolute measurements have large uncertainties. Provide data to support discrimination between objects to enable rapid development of precision orbits. Early element set uncertainties can allow data from another closelyspaced object to appear to be data on the first object. For distinctive objects, signature data can support consistent tracking of specific objects, which then allows better filtering of data to be used in forming orbital element sets, which then reduces the positional error.
Figure 3 Relative errors within train of satellites 3. JAWSAT Launch and Its Challenges The JAWSAT launch had 5 primary payloads (and several payloads deployed later). Figures 4-8 show the payloads with their dimensions along with sample RCS vs time plots from Millstone Hill Radar. [2] The solid line is the principal polarization and the dashed line the orthogonal. Note that the first three payloads are very distinctive in character, with ASUsat and OCS having much greater separation of principal and orthogonal polarization than OPAL, and the OCS being much larger than either. In addition, the principal return is much more stable for OCS than the others. The RCS vs Time plot for OCS is very characteristic of a sphere and quite recognizable to an experienced radar operator. FalconSat is of a size consistent with ASUsat and OPAL, tending toward larger than either. JAWSAT is larger than ASUsat, OPAL and FalconSat, but smaller than OCS. It also has much less separation between principal and orthogonal polarization than OCS. -10 RCS dbsm -30 ASUsat 24.5 cm x 32 cm diameter (14 sides) 22:32:15 TIME 22:32:30 Figure 4 ASUsat and RCS vs Time Plot at L-band [2]
-10 RCS dbsm OPAL 23.5 x 21 cm (6 sides) -30 00:12:30 TIME 00:15:00 Figure 5 OPAL and RCS vs Time Plot at L-band [2] +10 RCS dbsm OCS 3.5 meter diameter -30 12:07:00 TIME 12:08:00 Figure 6 OCS and RCS vs Time Plot at L-band [2] OCS -10 RCS dbsm -30 FalconSat 17 x 18 inch box 22:36:15 TIME 22:36:45 Figure 7 FalconSat and RCS vs Time Plot at L-band [2]
+10 JAWSAT 35 x 35 x 42 inches RCS dbsm -30 00:19:30 TIME 00:22:00 Figure 8 JAWSAT and RCS vs Time Plot at L-band [2] The predicted difference in mean anomaly between various payloads and OPAL, expressed as time, is shown in Figure 9, as provided to MIT Lincoln Laboratory by The Aerospace Corporation [2]. Note that OCS and ASUsat stay within seconds of OPAL even several hours after launch, and FalconSat returns in proximity. This complicates separation of observations by metric data alone. Time Residual Plot 0.100 0.080 0.060 Time residual (min) 0.040 0.020 ASUSat OCS FalconSat JAWSAT OPAL Minotaur 0.000 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00-0.020 Figure 9 Time residual plot Time since launch (hours)
The biggest challenge of the launch was producing accurate orbits on each object (including correct identification) within several days of launch, despite the low separation rates. The possibility of mis-tagging observations due to close proximities raised the possibility of corrupted element sets, which then lead to a greater possibility of misassociating an observation with an element set. The rapid identification and precision orbit determination was required to support communications between the SRI 150 ft dish and the OPAL to command deployment of the Aerospace picosats. Once OPAL was deployed, the picosats only had approximately 60 hours of battery life while operating in OPAL. The task of the surveillance sensors was to locate OPAL as soon as possible and provide acquisition element sets to USSPACECOMMAND as soon as possible, avoiding cross-tagging from pass to pass. [2] [4] [6] As an example of some of the challenge of using metric data alone to correlate observation to observation, theoretical measurement errors were calculated for the Millstone Hill Radar. Table 1 shows the parameters of the radar and Table 2 shows the estimated errors, assuming a 0.3m 2 target at 1000km and using a typical tracking bandwidth of 1 MHz [8]. These errors are estimated based on classic techniques such as those found in Skolnik. The measurement error alone for an object the size of FalconSat at typical ranges is approximately 80 m. The error for OCS will likely be less and for OPAL will be greater. Table 1 Millstone Hill Radar L-band Parameters [7] Center Frequency 1295 MHz Maximum Bandwidth 8 MHz Peak Power Pulse Width PRF 3 MW 1 ms 40 Hz Table 2 Millstone Hill Radar Theoretical Measurement Errors σ angular 4.80 mdeg (~ 80 m @ 1000 km) σ range 2.83 m 4. Concepts for Net-centric Operations to Support Multi-payload Launches The first set of net-centricity concepts can be referred to as shared site net-centricity. They are used when there are two or more sensor sites in close proximity, such as shown in the photo of the Lincoln Space Surveillance Complex (LSSC) in Figure 10. This site includes the Millstone Hill Radar, the Haystack Radar and the Haystack Auxiliary (HAX) Radar, and these sites use some of the shared site concepts already. Table 3 shows some of the basic operating parameters. In this case, there are radars at three different frequencies, and the operators can leverage the broader beam of the Millstone Hill Radar
for search along with the better small object detection capability of Haystack and HAX. In addition, Haystack and HAX can produce Range-Doppler images of the tracked objects, providing a much more natural correlation between data and physical shape than possible with a narrowband signature. Table 3 Basic Operating Parameters of Radars at LSSC Millstone Haystack HAX Beamwidth (deg) 0.44 0.05 0.1 Frequency 1295 MHz 10 GHz 16.7 GHz Haystack Haystack Auxiliary Radar ~ 1 km Millstone Hill Radar Figure 10 Lincoln Space Surveillance Complex Shared sites have the advantage of being able to have multiple sensors tracking the same object simultaneously. This can be enabled by sharing pointing data in real time. This simultaneous tracking then enables the ability to collect signatures over nearly identical satellite aspects. It also enables real-time cooperative operating strategies. Lincoln had enabled many net-centric concepts at the LSSC at the time of the Minotaur JAWSAT launch, and has enabled more since. This shared site provides the basis for real-world examples in discussions of shared site concepts. In the situation of a closely-spaced train of objects, LSSC operators use a process locally referred to as tag-team operations. Millstone, with its broad beam, searches for an object in the train. Millstone collects both metric and narrowband signature data on a found object and also provides pointing information to the other two sensors in real time.
The other sensors can choose whether to acquire or not, based on the operations plan. These sensors may collect metric, signature and/or image data. Once a sensor that is planning to track an object has acquired the object and the acquisition is verified (by voice communications or other means), Millstone can be freed to search along orbit for the next object in the train. The data from the other sensor(s) can be used to help verify the identity of the objects, and tracks by all sensors are given consistent object identity tags. A sensor like Haystack may use the acquired object as a reference point to search for other smaller objects. This approach was used in the detection and tracking of the ODERACS spheres in the mid-1990s [9]. The OCS in the JAWSAT launch made an excellent reference object, being both large and of a shape easy to distinguish from other satellites in both narrow-band radar and image data. When there is a distinctive reference object like a sphere or another object that can be confidently identified, sensors sharing a site can pass on information such as ahead of or behind a reference object (effectively a mean anomaly difference). This supports a search along an orbit. Such relative-spacing information can be passed to sensors at different sites with future tracking opportunities, with the understanding that differences in orbital altitude can lead to changes in relative positions. A new capability, developed since the Minotaur JAWSAT launch, is a joint control facility where all three of the LSSC sensors are operated. This joint control room was enabled by net-centric infrastructure that supports remote operation of the individual sensors. In this current control facility (Lincoln Space Situational Awareness Center), the operator of one radar can observe the data screens of the other radars, and operators can easily discuss what is happening with multiple radars with the screens for all of the radars in front of them. In addition, there are several radars not at the LSSC whose screens can be observed in real-time at the LSSAC. These are not controlled at the LSSAC, but the data is readily available to operators there. For most launches, the same-site cooperative activities are paired with a search strategy that leverages the nature of low-velocity deployments. In such cases, most of the velocity applied to separate the objects is applied in the plane of the orbit. For low separation velocities, it is practical to assume that all objects are in nearly the same orbital ellipse and separated in mean anomaly. This assumption allows the use of a one-dimensional search that we refer to as a time-search. The concept of a time search is to assume that an object of interest follows an orbit defined by a specific orbital ellipse (including plane) and within a span of mean anomaly values. A sensor operator identifies the earliest possible rise time implied by the range of mean anomaly values (perhaps with some additional padding on mean anomaly) and positions the radar at the horizon at an azimuth consistent with detecting an object at the earliest rise time. LSSC operators often select a rise time that is 300 seconds prior to the rise time of a nominal reference element set. For mean anomalies corresponding to later rise times, the azimuth where the satellite would rise is different than the azimuth for the earliest rise, but in a predictable way. It is possible to steer smoothly at the horizon along the predicted rise azimuths for the advancing mean anomalies to have the radar in the right place at the right time to detect an object with any given mean anomaly within the range. The width of the beam allows for some tolerance in predicted location and
steering, as well as some altitude/plane differences. Figure 1.1 illustrates some of the concepts of the time search. This search strategy is extremely effective when combined with shared-site cooperative operations. While a single sensor operating alone may not be able to search and also collect adequate metric or characterization data during the time the object train is within view, the use of multiple sensors allows for collection of additional metric and characterization data during the period when the search sensor is acquiring the next object. TIME BIAS - 30 seconds Predicted elset Actual elset range dimension Target rises early and is detected at a lower than expected range Radar horizon Figure 11 Time-Search [2] The key benefits of the proposed shared site strategies are summarized here: a) Metric data on multiple objects within train Use of the tag-team process allows controlled sequence of data collection on various objects in the train. Sensors not occupied in search can spend longer collecting metric data, while allowing the search sensor to perform the initial detections. b) Characterization in two phenomenologies Multi-phenomenology characterization provides an improved chance of correct identification over single-sensor data, especially when something such as range-doppler imaging is available. In addition, the multiple frequency/multiple phenomenology data may help other sensors tracking the object with discrimination and identification. For example, if the tag-team search enables Haystack to collect X-band signatures on all
objects in a train, then an X-band sensor with a later pass (but no search support) may use the prior X-band data to help quickly identify the object of interest. c) Careful cataloging of relative position with revisit opportunity The relative position data, particularly for closely-spaced objects, can help one sensor guide a second sensor to the correct object. This aids the process of characterizing the spacecraft using multiple phenomenologies. It can also be passed to sensors with future track opportunites, particularly in the near term, to aid in correct identification. d) Small object search A reference object can be used as an anchor to search for small objects. If a sensor like Millstone tracks the reference object continuously, a sensor like Haystack can search in the vicinity of the reference object, matching pointing with Millstone on a routine basis to recalibrate the basis for the search. Once the high-sensitivity sensor has found the object, it may then be possible to direct the other sensors to the object by sharing pointing information. A single phased array radar can achieve some of the benefits of a shared site arrangement. A phased array may be able to focus its horizon search more narrowly by using the time search process than when using its typical fence. This may provide benefit in detection likelihood by increasing the number of pulses potentially hitting the target and/or allowing for multi-pulse integration. Since a phased array radar can easily track multiple objects, it can follow an orderly process of searching for and tracking the multiple objects in the train and also provide relative separation information. Depending on the track versus characterization capabilities, it may be able to characterize some objects while tracking others. It also could provide comparative cross section vs time data on the objects in the train. The one main disadvantage over the multi-sensor shared site model is the inability to leverage the different strengths of the different types of radars. The second set of net-centricity concepts can be referred to as site-to-site hand-off. These concepts are used when sensors at one site collect metric and signature data and provide information to sites that will have passes not long after the first site. The process is beneficial when the nature and timeliness of the information allows the site with the later pass to do a better job of detecting, tracking, identifying and characterizing the objects than possible without the pre-pass information. The hand-off data include; time difference from nominal element set, relative ordering of objects, position data on reference object (e.g., large calibration sphere), element sets, and signature data or statistics. Under the assumption that an object will be in a similar orbit to what is expected or modeled, one of the easiest pieces of information to exchange is the time difference between the actual and predicted mean anomaly. This information supports determination of predicted rise time as well as time searches. Related to this is the relative ordering and spacing of objects in a train of objects. As long as the orbital mechanics have not caused a swapping of apparent order (which happens over time with orbits of different altitudes), this ordering can help with consistent tagging of the identity of objects. The other benefit is that the later sensor sites can use the hand-off to reduce the time to find a specific object of interest. This is useful when the later sensor has
limited search capability and has to perform a specific task with respect to a specific object. The evolution of orbits and the apparent change in relative position (illustrated previously in Figure 9 with the blue/purple cross-over and the yellow cross-over of pink and blue) can cause problems with use of this hand-off data. Early data can be used to estimate separation velocities, which can then be used to predict orbits and changes in relative positions. Unfortunately, the surface-to-mass ratio is larger for small objects than for typical spacecraft, and this can cause difficulty in modeling the effects of drag on the objects. Figures 4-8 showed that the RCS vs time data for the different objects in the JAWSAT launch was distinctive. If one sensor can select a reference object that will have a distinctive signature over multiple sensor operating bands, then later sensors can clearly identify the object as the reference object. This provides additional benefit to relative position information, since positions of all objects can be specified with respect to a single easily identifiable object. This object can also become the anchor for searches at the new sensor. If signature data or key signature statistics are combined with metric data in a joint correlation algorithm, this new data can help provide a higher likelihood of correctly associating metric data on the same object and filtering out metric data on other objects. Average RCS alone may be insufficient, but information such as estimated tumble rates, differences in response as a function of polarization, and the variability of the cross section can provide parameters for discrimination. The third set of net-centricity concepts are based on algorithms that leverage multisensor data, and use of multi-sensor signature data has been alluded to previously. The growth in computational power of computers, as well as extensive algorithm development, has enabled the possibility for powerful data fusion techniques. Researchers at Lincoln Laboratory have begun exploring some of the possibilities, but this paper will just identify several interesting concepts. These concepts represent a small subset of the algorithm concepts that could be useful in space surveillance when leveraging multi-sensor data. The first set of algorithm concepts involve signature prediction and matching. With published information on objects being launched, it would be possible to make reasonable predictions about the radar or optical signature of an object as a function of frequency, attitude, and (as appropriate) illumination angle. As a sensor prepares for a pass, the predicted signature of objects for the pass could be selected and compared with actual signatures to support identification. Once a sensor has collected signature data on an object, that signature data can be used as the basis for matching new signature data. As mentioned previously, inclusion of signature matching in correlation processing has the potential to reduce the incidence of false correlations. A more advanced capability would be to predict signature for future collection based on a past collection. Since the future collection will likely be at a different attitude and/or illumination angle and may also be at a different frequency, this would likely involve some prediction of object shape based on signature data. Such estimates could then be used to predict signatures at different aspects or for sensors at different frequencies.
Another capability would be to be able to compare two signatures and determine whether they are on the came object or not. This will be more complex as attitude angles and frequencies differ. The second interesting area of exploration on concepts is to be able to construct an image or three-dimensional representation of an object based on signatures from multiple sensors, separated in distance. This would be most useful if it could be done rapidly. 5. Applying Net-centric Concepts to JAWSAT Launch LSSC operators used many of the shared site concepts presented earlier in support of the JAWSAT launch. In particular, the time search and tag-team operations were very effective. Operators also used multi-site hand-off procedures providing metric data, relative position data, and signature data directly to the ALTAIR radar site in the Marshall Islands. There was limited advanced algorithm support, but the team had very capable operators who could recognize signatures and signature similarities. These operators also had experience working with the sister sites and looking at data from those sites. The time search was very effective in finding the train of objects and then enabling the tracking of various objects in the train. The OCS was an excellent object to use as a positional reference object, since the sphere produced a distinctive signature at all radar frequencies, and showed very clearly as a sphere in images. By handing off pointing directly, operators guaranteed that the site was consistent, sensor-to-sensor, in the identification of the objects on which each sensor collected data. This facilitated multiphenomenology characterization. It also allowed operators to set up a search anchored on the OCS. Hand-off of relative positions, particularly with a predicted set of relative positions available, let the two separated sites collect data and tag it consistently from site to site. 6. Evolving Net-centric Operations at MIT Lincoln Laboratory As described previously, Lincoln now has a joint control room for operation of multiple sensors. This joint control room allows operators to view other sensor activities real-time. There is direct communications among sensor operators, and the shared room enhances cross-sensor familiarity. In addition, operators in this facility can do remote viewing of ALTAIR and the other radars sharing the ALTAIR site. This enables real-time viewing of sensor activities and cross-sensor familiarity over multiple sites. They say a picture is worth a thousand words seeing the activities is much more informative than reading a report. Remote viewing also provides the best possible planning time from tracking at one site to tracking at the other, since operators view the other site operations as they happen rather than waiting for descriptive reporting. One future possibility might be a joint control room for multiple sites. Care should be taken to still use trained and experienced operators, since they are key to dealing with unusual situations.
Lincoln is developing a net-centric infrastructure to use as a testbed for net-centric space surveillance operations. The operations team is actively working on the innovative concepts to improve operations. The testbed will also provide the infrastructure for the US space surveillance community to explore concepts for data distribution and for serving new applications. In addition, Lincoln researchers are exploring algorithm concepts of a) Model-based signature prediction, b) Statistical signature matching, and c) metric/signature matching are being explored. These can be tested using the net-centric infrastructure under development at Lincoln Laboratory. 7 Summary Looking toward the future, we anticipate a variety of changes in the space operating environment. Of particular interest are the following: a) Space will become more available for small-science applications b) Multi-payload launch and microsats will be enablers for small-science efforts in space c) A non-trivial subset of payloads will be short-lived d) Separation velocities for small-science launches will typically be slow In terms of enabling net-centric concepts, current communications allow for large data transmissions. The importance of high capacity communications is highlighted in the following three areas: a) Distribution of auxiliary information (like signatures) to all participants b) Remote viewing of sensor operations c) Command and control linkages for centralized sensor operations d) Fusion of more complete sensor data this has the secondary benefit of bringing the data together for development of new algorithms. Also enabling net-centric concepts is the continued growth of computing power. This enables: a) Detailed modeling that can be executed within practical time spans b) Complex statistical matching algorithms c) Complex data fusion algorithms With the growth of both communications and computing power, the concept of netcentric operations has the potential to enable us to conduct space surveillance in new ways. This paper presented several concepts for using a net-centric infrastructure to address the challenges posed in supporting launches of multiple small scientific payloads. These concepts are summarized below. Disseminate information from one sensor site to all participants. The time differential from nominal, the relative order of objects within a train of objects, signature data or statistics, and element sets (particularly on reference objects) can help reduce incidents of cross-tagging. This can then improve the speed of fitting quality orbits. In addition, such data can help to reduce the search efforts of sensors with tasks related to specific objects and limited search capabilities. Shared sites, particularly those with sensors of different types can leverage cooperative operations to significantly improve the success of detection, tracking and characterizing
newly launched payloads over what can be achieved by sensors operating independently. Cooperative efforts enable simultaneous multi-phenomenology characterization of an object while ensuring that all sensors are actually collecting on the same object. Cooperative operations also allows multiple sensors to search to their strengths (e.g., area coverage, detection of small objects) while providing the benefits of their finds directly to other sensors that are poorer with that type of search. Remote viewing of sensor operations allows operators at one site direct and immediate familiarity with the activities of and data collected by another site. This is without the time-lag or information filtering inherent in providing such information in a report. While the expertise of operators at each site is distinct and critical, and while expert interpretation of raw data may be required for the information to be fully useful, remote viewing fosters operator-to-operator interchange and provides greater opportunity to request information and interpretations beyond what operators at a site might naturally provide in a report. Advances in computing can enable operational use of signature prediction and matching algorithms. Signatures models based on pre-launch information on payloads can be used to seed signature matching algorithms. Signatures can be compared pass-topass on objects. If the object correlation algorithms are a marriage of metric comparison and signature comparison, the likelihood of cross-tagging can be reduced. It may be possible to fuse signatures from sensors operating at different frequencies to estimate features on a spacecraft. By providing a broader dissemination of data to other sensor and processing sites, including the dissemination of data that normally does not leave a sensor site, there is greater potential for enhancing the knowledge about an event through fusion of multiple types of data. Another area alluded to previously is leveraging non-traditional sources such as owner/operator information. For example, models from the owner could be used to support signature prediction. Predictions of separation trajectories could aid sensors in searches. Data from the operators of one satellite that positively identifies their satellite in a train and includes metric data could enable the identification of other objects and allow for anchoring of a search. This would be particularly useful in a launch for which there are many objects of similar shapes but with different functions and owners (e.g., a launch of cubesats). In summary, easier access to space requires timely ways to deal with launches with subtle operations. Net-centric operations concepts and infrastructures provide means to make broader, better use of existing capabilities. Lincoln Laboratory has demonstrated gains from extensive multi-sensor sharing and coordinated operations. This paper has presented several proposals on ways to expand these concepts to a broader space surveillance network.
Bibliography 1 Encyclopedia Astronautica, Mark Wade, 1997-2007 http://www.astronautix.com/ 2 Bougas, W, Cott, T.A., and Andrews, S.E. Unique Technologies of Millstone Hill Radar Detection and Tracking of the DARPA PICOSATS, Core Technologies for Space Systems Conference, Colorado Springs, CO November 29, 2001 3 Tang, W Flight Plan: SSN Support to DARPA/Aerospace PICOSAT Mission [Prepared for USSPACECOM/J33] September 5, 1999 4 Aerospace Corporation, Picosatellites Complete Mission http://www.aero.org/news/newsitems/complete-021400.html 5 Network Centric Warfare Department of Defense Report to Congress, 27 July, 2001 6 SRI International SRI International: The Dish Antenna Facility http://www.sri.com/esd/dish/ 7 Stone, M.L. and Banner, G.P. 2000. Radars for the Detection and Tracking of Ballistic Missiles, Satellites, and Planets. Lincoln Laboratory Journal 12, 217-244. 8 Gaposchkin, E.M. 1985. Metric Calibration of the Millstone Hill L-Band Radar. MIT Lincoln Laboratory Technical Report 721 (ESD-TR-85-173). Approved for Public Release; distribution unlimited. 9 Andrews, S.E., Hall, D. Sridharan, R. Searching for Satellite Ejecta with Groundbased Radars, Proceedings of the Second European Conference on Space Debris, ESOC,Darmstadt, Germany, 17-19 March 1997 (ESA SP-393, May 1997). 10 eoportal OPAL(Orbiting Picosat Automatic Launcher) http://directory.eoportal.org/pres_opalorbitingpicosatautomaticlauncher.html 11 Cebrowski, A.K. and J.J. Garstka, Network-Centric Warfare: Its Origin and Future United States Naval Institute Annapolis Seminar Proceedings, January 1998