MDA’s Space-Based Kill Assessment (SKA) Experiment (August 9, 2016)

The Missile Defense Agency (MDA) plans to deploy a Space-based Kill Assessment (SKA) system by about mid-2017. The SKA system, which MDA describes as an experiment, will consist of small sensor packages deployed on a number of commercial satellite hosts. It is intended to demonstrate a capability to rapidly determine whether or not an interceptor has hit and killed its intended warhead target.  Other than two Space News articles, the SKA system has received little public attention.[1]

MDA argues that a kill assessment capability can reduce the cost and improve the efficiency of a missile defense system by eliminating the need to fire additional interceptors at a target that has already been destroyed.  “The faster we can determine a threatening missile has been eliminated, the fewer the number of interceptors are need in the fight.”[2] The MDA goes as far as arguing that the SKA experiment “has the potential to change the economics of the defense of the American homeland from enemy ballistic missiles.[3]  This approach to reducing the number of interceptors fired — known as shoot-look-shoot – may or may not be possible depending on the timelines involved.  However, at a minimum, it is clear that this approach requires a rapid capability to assess the outcome of an intercept attempt.

Is There a Need for Additional Kill Assessment Capabilities?

The task of kill assessment is closely related to that of target discrimination.  Effective discrimination – the capability to identify the actual warhead from among other possible threatening objects such as deployment debris, rocket stages or decoys – is universally recognized as essential to the effective operation of a missile defense system.  And the current U.S. GMD national missile defense system is officially claimed to be very effective. But if a defense system is very effective at identifying the warhead from among other objects, shouldn’t it also be able subsequently to determine if the warhead has been hit and destroyed?  Why then is an additional kill assessment capability needed?

In principle, there are several scenarios in which a capability such as that provided by the experimental SKA system might have at least some utility.

One possibility is that the current discrimination capability may not be nearly as good as the current claims of high GMD effectiveness seem to imply.  The effectiveness of the GMD system is assessed against a defined official threat – all details of which are classified.  However, it is very likely that this official threat is quite simple, perhaps no more than a warhead accompanied by several objects quite different from the warhead.  At the time that the MDA stopped releasing information about the decoys is was using in tests, the only decoys it was using in these tests were spherical balloons quite different in size from the conical warhead target. Against such a simple defined threat, accurate and effective discrimination might be possible, and a high kill probability could then be achievable by firing multiple interceptors (to overcome the interceptor’s currently low reliability) at the identified target.

However, against even a slightly more difficult target set – for example a final booster stage sliced up using cutting cords into warhead-sized pieces with the warhead itself disguised so as not to appear conical – the defense could be confronted with a large number of targets, any one of which could be the actual warhead.  In this case, the GMD current discrimination sensors — the Sea-Based X-band (SBX) radar and the infrared seekers on its interceptor’s kill vehicles – may have little capability to pick out the real warhead before an intercept attempt is made. Even if the interceptor strikes an object, the radar may not be able to determine whether this object was the real warhead.  (In 2004, the Israeli defense company Rafael, arguing for an optical kill assessment system, possibly based in space, stated that the breakup of the Iraqi Scuds in the 1991 Gulf War formed clusters of debris such that “Identifying the warhead among these clusters is practically impossible [by radar], making effective warhead kill assessment practically impossible.”[4])   As discussed below, the experimental SKA system can make no contribution to picking out the target before the first intercept attempt.  However, there is at least the possibility that it might be able to contribute to an assessment of whether or not an intercept attempt destroyed an actual warhead.

As another scenario in which a SKA capability might be useful, consider an intercept attempt by the Aegis Ashore (AA) facility in Romania against an Iranian missile launch against a target in Western Europe.  In this scenario, the warhead itself might never be detectable by the AA radar given it is limited range.  Instead, interceptor(s) would be launched based on data from the forward-based TPY-2 X-band radar in Turkey (a process known as engage-on-remote).  However, the intercept attempt(s) could well take place outside of the field of view of the TPY-2 radar.  In this case, something like the SKA system might provide the only kill assessment capability other than that of the seeker on the interceptor.  However, whether or not enough time would remain to able to make use of such information in launching additional interceptors would be scenario dependent.

Aside from scenarios such as these, it is apparent that several kill assessment related issues have arisen in past tests of the GMD system.  In intercept test IFT-06 held on July 14, 2001, the primary radar observing the intercept attempt, the Ground-Based Radar – Prototype (GBR-P), incorrectly reported the outcome of the intercept attempt.   According to the Director for Operational Test and Evaluation (DOT&E), “The only objective not satisfied in IFT-6 was real-time hit assessment by the GBR-P, which incorrectly reported a MISS.”[5]  In intercept test FTG-02, held on September 1, 2006, MDA reported the test as fully successful.  Several years later, however, DOT&E revealed that it had assessed the intercept attempt as “a hit but not a kill,” since it only achieved a “glancing blow” that would not have destroyed the target.[6]  In FTG-06, held on January 31, 2010, the Sea-Based X-band (SBX) radar that was tracking the target and guiding the interceptor shut down unexpectedly shortly before the intercept attempt due to “chuffing” of rocket motor fuel out of the interceptors final booster stage.[7]  The SBX, which was the only GMD system radar observing the intercept attempt (except for the California PAVE PAWS Upgraded Early Warning Radar, which observed part of the test in an off-line mode) was therefore unable to provide an immediate kill assessment of the intercept attempt, which failed due to a kill vehicle malfunction.


What Is The SKA System?

The FY 2014 National Defense Authorization Act Stated that the MDA should develop “…options to achieve an improved kill assessment capability for the ground-based midcourse defense system that can be developed as soon as practicable with acceptable acquisition risk, with the objective of achieving initial operating capability by not later than December 31, 2019…”[8]

In April 2014, MDA started the SKA program using funds left over from the cancellation of the Precision Tracking Space System (PTSS).  Each of the SKA sensor packages (see Figure 1 below) includes three single-pixel commercially-available photodiode detectors and has a mass of “approximately ten kilograms” (22 pounds) and can tilted around two axes.  The packages are being developed and built by the Johns Hopkins University Applied Physics Laboratory (APL), which will also initially operate the SKA system before turning it over to the MDA.



Figure 1. The Space-based Kill Assessment sensor.  MDA image[9].

MDA has not stated which commercial satellites will host the SKA packages, although it has been speculated that they will be Iridium Next communication satellites.[10]  According to FY 2017 MDA budget documents, the SKA packages will be deployed using three launches in the 3rd and 4th quarters of FY 2017.  This is consistent with the current Iridium Next launch schedule, in which the first launch of ten satellites is scheduled for September 16, with the goal of completing the system (66 satellites + four on-orbit spares) with six additional ten-satellite launches by the end of 2017 (additional spares would be launched subsequently).[11]  Since the SKA packages will be included in three launches, they could be deployed on up to 30 satellites, although the actual number could be lower. [Added 08/10/2016: According to Admiral Syring response to a question, Senate Appropriations Committee on April 13, 2016, SKA packages will be deployed on 22 satellites.]

The Iridium Next satellites will be in in 780 km altitude near-polar (86.4 degree inclination) orbits, in six orbital planes of eleven satellites each.  Each satellite has space dedicated to hosted payloads, with up to 50 kg mass, 30x40x70 cm volume, and average power of up to 50 W (200 W peak), allocated for hosted payloads.[12] The roughly 10 kg SKA packages would seem to easily fit within these parameters.  Multiple hosted payloads can be accommodated on each satellite as long as they fit within the overall limits. The hosted payload can be mounted facing either directly downwards (nadir direction) or along the satellite’s velocity vector (ram direction).

What Can the SKA System Do?

Although the MDA describes the SKA as intended to “best leverage intellectual capital investment in the PTSS program,” the SKA system will use much simpler and less capable sensors. The Precision Tracking and Space Surveillance System (PTSS) would have deployed 9-12 satellites, each with a telescope equipped with cryo-cooled charge-coupled device focal plane arrays.  Each satellite would have been capable of tracking a ballistic missile and its warhead throughout almost it entire flight (although it would have required cueing for initial detection) using several different infrared (and possibly visible) spectral bands.  Its information could be used guide interceptors, and although its resolution would have been too poor for effective discrimination, it may have also provided some information useful for kill assessment.

In contrast, the passively cooled, single pixel SKA sensors will have little or no tracking capability.  Instead they will apparently rely on information provided by the missile defense command and control system about the location of the expected intercept point to position their sensors in advance to observe the visible and infrared light produced by the high-speed collision of the intercept.  Thus they certainly cannot provide any pre-intercept discrimination information.  According to MDA budget documents, the data from the SKA sensors will be used for “flash detection and analysis; hit/miss/kill/glancing blow assessment.”[13]

A 2010 paper by researchers at the Johns Hopkins Applied Physics Laboratory (APL), which is developing the SKA sensor packages, describes their “technology effort [that] led to the foundation of a technical approach to develop space-based kill assessment sensors.”[14]  They state that the kill assessment technologies they were developing needed to be able to answer questions such as: “Did the interceptor hit the target and did it hit the intended target?  What type of payload (for example, nuclear, high-explosive, chemical or biological) did the target contain? Did the interceptor render the payload non-lethal?”  This information was to be extracted from the visible and infrared light produced by the impact of the hit-to-kill interceptor, including both an impact flash as well as radiation emitted (and in the case of a sunlit impact, reflected) by the expanding cloud of impact debris. According to a U.S. Army Space and Missile Defense Command (SMDC) fact sheet on kill assessment, key kill assessment observables are “ “cloud” expansion rate, intensity-time history and spectral signatures.”[15] (Although since the SKA uses non-imaging sensors, it cannot directly measure the cloud size or expansion rate.)  The interpretation of this visible and infrared light would be based on extensive modeling (for example, using “Re-entry vehicle Intercept Signature Kill assessment” (RISK) models) as well as experimental observations.[16]

The authors of the 2010 APL paper state that APL has played a major role “in support of the Missile Defense Agency’s (MDA) Kill Assessment Technology Program, an ongoing effort begun in 2001 to develop critical kill assessment technologies,” including “preflight predictions and postflight analysis for a number of BMDS [Ballistic Missile Defense System] intercept flight test missions during the past seven years.” According the SMDC fact sheet, data has been collected on Patriot, Aegis BMD and GMD intercept tests.  An example of such data is shown in Figure 2 below (from the APL paper).

SKA2                                     Figure 2.  Three infrared time-intensity profiles collected during an intercept test with non-imaging, single polarization sensors.[17]

It seems plausible (to me) that an SKA-type sensor could differentiate between a direct hit on a warhead and a hit on a light-weight decoy.  On the other hand, it less obvious that one could distinguish between an impact on a heavier decoy (such as a cut-up portion of a booster stage) and an impact (particularly a glancing one) on a warhead.

However, the MDA and the APL sound confident that the SKA system will work. MDA budget documents state that: “Nine years of testing using the “Kill Assessment Sensor Package” sensor on the Aegis Ballistic Missile Defense program indicated that the physics of the kill assessment problem was well understood and that expensive and risky technology development was not need for space-based kill assessment.  This sensor testing on the Aegis Ballistic Missile Defense program also showed that an electro-optical/infrared sensor was the optimal sensor to observe and intercept and record data in the frequency bands most advantageous for kill assessment”[18]  The APL researchers concluded that their “technology now provides the basis for designing and implementing kill assessment sensors. If implemented, such a kill assessment system will result in a savings on interceptors and will provide the situational awareness that will be required by senior leaders during operation of the BMDS.”[19]


[1] Mike Gruss, “MDA Kill Assessment Sensors Would Be Commercially Hosted,”, March 20, 2015. Online at; Mike Gruss, “U.S. Missile Defense Agency’s Hosted Payload Delayed until mid-2017,”, April 21, 2016. Online at

[2] Missile Defense Agency, “Frequently Asked Questions: Space-based Kill Assessment.”  Online at

[3] U.S. Department of Defense, Fiscal Year 2016 President’s Budget Submission,  Missile Defense Agency, RDT&E, Vol. 2a, February 2015, p. 2a-482.  Online at:

[4] “EO Sensors Could Provide ATBM Kill Assessment,” Jane’s Missiles and Rockets, April 2004.  Note that this scenario involves intercepts in the atmosphere, which should make both discrimination and kill assessment easier.  In the 1991 War, after a period of confusion, Patriot operators learned to target only the fastest falling object emerging from a missile breakup.  However, this still involved a salvo-firing strategy in which all the interceptors were launched before the first intercept attempt took place.

[5] Director, Operational Test and Evaluation, “National Missile Defense (Ground-Based Midcourse Defense),” 2001 Annual Report, February 2002, p. VI-5.  Online at

[6] See my blog post of October 18, 2012, “Ballistic Missile Defense:  A Significant Advance in Missile Defense Criticism Evasion Technology.”  Online at

[7] Amy Butler, “Diverted Attention,” Aviation Week and Space Technology, April 12, 2010, p. 26.

[8] National Defense Authorization Act for Fiscal Year 2014, Public Law 133-66, December 26, 2013, sec. 237.  Online at

[9] Image from MDA, “Fiscal Year (FY) 2017 Budget Estimates – Overview.” Online at

[10] Gruss, “U.S. Missile Defense Agency’s.”

[11] Peter B. deSelding, “Iridium’s SpaceX Launch Slowed by Vandenberg Bottleneck,”, June 15, 2016.  Online at

[12] Robert E. Erlandson, Michael A. Kelly, Charles A. Hibbits, C.K. Kumar, Hugo Darlington, Lars Dyrud and Om P. Gupta, “Using Hosted Payloads on Iridium Next to Provide Global Warning of Volcanic Ash,” in Proc. SPIE 8371, Sensing Technologies for Global Health, Military Medicine, Disaster Response and Environmental Monitoring II, and Biometric Technology for Human Identification IX, May 01, 2012.

[13] FY 2016 President’s Budget, p. 2a-484

[14] Robert E. Erlandson, Jeff C. Taylor, Christopher H. Michaelis, Jennifer L. Edwards, Robert C. Brown, Pazhayannur K. Swaminathan, Cidambi K. Kumar, C. Bryon Hargis, Arnold C. Goldberg, Eric M. Klatt and Greggory L. O’Marr, “Development of Kill Assessment Technology for Space-Based Applications,” Johns Hopkins APL Technical Digest, Vol. 29, No. 3 (2010), pp. 289-297. Online at

[15] U.S. Army Space and Missile Defense Command, “Kill Assessment Program,” fact sheet (undated).  Online at

[16] Erlandson, et al., “Development of Kill Assessment,” pp. 291-293.

[17] Erlandson, et al., “Development of Kill Assessment,” p. 295.

[18] FY 2016 President’s Budget, p. 2a-481.

[19] Erlandson, et al., “Development of Kill Assessment,” p. 295.

Leave a comment


  1. In the description of the use of a TPY-2 in forward-based mode with an Aegis Ashore site, the article states that the target warhead may never be detectable by the AA SPY-1 and thus would require launch-on-remote. I believe this should be engage-on-remote since launch-on-remote still requires the SPY-1 to acquire and track the target post SM-3 launch (ie. the launch is done using TPY-2 track data but post-launch the SPY-1 will acquire and track the target for the remainder of the engagement). Whereas engage-on-remote utilizes the remote radar (TPY-2) for the entire target track engagement sequence, with SPY-1 only performing missile uplink functions.

  2. Yes, you are correct. My mistake. I will correct the post.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: