Set Phasers To Kill
A spectrum update
Seen in action above, DragonFire “cost about £100 million, or $126 million, to develop.” It is basically a laser cannon. “Firing it for 10 seconds is the cost equivalent of using a regular heater for just an hour, “ according to the Ministry of Defense. Whereas a US Navy SM-2 interceptor missile costs $2.4 million, the price of a laser blast “is typically less than $13 a shot.” Laser weapons have the advantage of ‘deep magazines,’ being limited only by available power.
This is what catching up looks like, though. Israel’s Iron Beam, now in deployment, has about three times the power output of DragonFire. US Navy ships are already carrying more powerful lasers than this one. There are other problems common to lasers, such as limited range due to atmospheric scattering. Laser weapon development has been underway for decades, but has yielded few tangible results. Engineering issues are mostly to blame, but also there was no tactical mission which made the laser economical. That cost-benefit equation has changed.
As I recently explained in an essay for premium subscribers, radio frequency jamming has enjoyed high success rates against the anti-ship missile threat for a long time. What has changed now is the drone threat, more specifically the drone swarm. Compared to “hard kill” systems like the famous Phalanx CIWS (close-in weapon system), lasers will be superior against swarming tactics with drones because they do not run out of ammunition. Perhaps we will see Ukraine test out this technology soon. After all, the “deep magazine” effect eliminates reliance on foreign partners for ammunition supply.
High-energy lasers (HELs) properly belong to the domain of electromagnetic warfare EW). So-called “soft kill” systems that jam the radar on anti-ship missiles, or the navigation and control signals of drones, are becoming more powerful and more common. For example, US Navy ships in the Red Sea today are equipped with AN/SLQ-32, an active electronic scanned array (AESA) which emits a “pencil beam” of focused radio energy at the target.
“Since we can move and steer beams at computer speed at literally very small portions of seconds, we can put multiple beams out simultaneously and we can hit multiple things at the same time,” Mike Meaney, Northrop Grumman VP in charge of the Surface Electronic Warfare Improvement Program (SEWIP) project, explained in 2021. “So we can jam multiple threats simultaneously, we can dynamically scan them [the beams] and move them around quickly and do multiple functions simultaneously, as needed.”
These advances are made possible by new, solid-state antennas capable of scanning a larger part of the spectrum than past generations of purpose-built antennas. Artificial intelligence (AI) and software-defined radio (SDR) interfaces allow humans to observe all that spectrum without missing important changes. Industry buzz calls this “full spectrum threat awareness” and “instantaneous bandwidth.” Developed by Israelis first, the technology is visible in the ‘cheek bulges’ of this brand-new EA-37B.
I wrote that post in March. At the time, the modified business jet was an EC-37B. Recently, however, the US Air Force redesignated this aircraft as EA-37B because of its “electronic attack” role, fueling persistent rumors that the plane has a microwave weapon in those cheeks. This form of directed energy (DE) weapon overloads the circuits of the electronics onboard the target. Guidance, navigation, and control systems simply fizzle out. Like a laser, or AESA jamming, this weapon uses a very narrow, focused beam of energy for each target. It has no physical magazine to replenish. These weapons could also defend Kyiv and Kharkiv and Odessa, freeing up air defense systems that use physical ammunition for direct combat. And perhaps they already are doing that.
Last week, Ukraine released a photo of a Russian copy of the Iranian Shahed-136 kamikaze drone, called Geran-2, that landed gently in a field after being hit with “an undisclosed electronic warfare system.” This is consistent with a spoofed guidance signal. (In fact, that is exactly how Iran brought an American RQ-170 Sentinel drone into their airspace and down to earth in 2011.) This was despite efforts to harden the navigation system:
The latest Shahed drones now incorporate Russian-made Kometa digital antennas in their navigation systems, a strategic adaptation aimed at reducing susceptibility to enemy jamming. This technology shares similarities with that used in Russia’s UMPK glide bombs.
Shahed-136 was specifically supposed to overwhelm (or “swarm”) air defenses with so many targets that some number of them will get through to their intended targets. However, all the Geran-2 drones used against Ukraine are guided by GLONASS, the Russian equivalent of GPS, without an inertial navigation system (INS) backup. Ukraine likely took advantage of this and use some sort of AESA-like system to focus their signal power.
Russian military planners are acutely aware of this vulnerability and seek to exploit it against western technologies, including the precision weapons supplied to Ukraine. The Kremlin recently tested out GPS jamming in the strategic Suwalki corridor, through Polish territory, from a jamming system located in the strategic enclave of Kaliningrad. In the event of World War III, that jammer should be a first target of NATO action in the first hour.
This is called a “denial of service” (DoS) attack. Russian forces in Syria have also used GPS DoS to deny airspace to the enemies of the Assad regime. It is distinct from “spoofing” with a false signal. Russians first tried GPS spoofing on the Black Sea in 2016, and they have tested it at various locations since then, usually forcing GPS receiver sets to report their location as the nearest Russian airport. At first, this practice seemed to be following Vladimir Putin around Russia, but it frequently interfered with civilian navigation. (You can read a fascinating report on all that from the University of Texas at Austin.)
A new arms race is now under way. Lasers still need a few seconds on target to destroy it. The lower the power, the longer the time-on-target has to be. Manufacturers will use ablative materials to increase the need for power and time on target. Even AESA and microwave weapons must focus energy on a target, or on separate targets, for some amount of time in order to misguide or destroy it. Materials scientists are already exploring potential shielding technologies against microwave pulse weapons; I have already observed a webinar demonstration.
In a future combat environment filled with drones, jammers, spoofing, and direct energy threats, including lasers that can destroy incoming artillery shells and mortar rounds and rockets, AI will be the new king of battle. It is already there. This is most true of all in the electromagnetic spectrum, where novel threats can appear, giving human operators only 60 seconds to react. AI is capable of “in-mission learning.” People working in the new field of “cognitive EW” discuss the need to evaluate systems for their ability to learn rather than their performance.
Indeed, the word “cognitive” refers to the way the AI makes decisions. Like lasers, cognitive systems are older than AI. See, for example, the “Kalman filter,” an algorithmic method for navigational tracking one’s position while on the move in an aircraft or other vehicle. If you have ever driven into a tunnel and your car’s built-in GPS continued to report a position, this was because the Kalman filter was estimating your position based on your speed.
The really new thing in the world is not AI. It is the AI training environment — that is, the AI simulation that teaches an AI to change the way it makes decisions. Training an AI is really about the art of simulation, giving the AI new “games” with new adversarial problems to solve. The call is out for AI “red teams” to create these challenges. Each scenario will likely teach programmers new behaviors that are unacceptable to human needs.
If an AI detects that a laser is being aimed at it, for instance, and the AI is programmed to avoid the laser, but then executes this imperative by flying a manned fighter into the ground, that behavior must be identified and excluded from happening in the real world.
At the moment, “low fidelity testing” is the only way to test an AI very quickly. More complex AI environments will be required, and a great many different ones must be applied to avoid “environmental overfit.”
However, we can expect the drone swarm to preoccupy this process. It is a “research hotspot,” potentially allowing drones to concentrate and disperse in unpredictable ways. Some AI researchers suggest that the western preference for delegated authority is a possible advantage over Russian and Chinese swarming AIs, which do seem to obey a “follow-the-leader” format. Only time, and war with direct energy weapons against drone swarms, will truly tell.
Polemology Positions is a reader-supported publication. Please like, share, subscribe, and consider a paid subscription to support my work