Beginning in the mid-1990s with the initial success of the General Atomics Predator system, growth of remotely piloted aircraft has exploded and continues to do so at ever-increasing rates. Assisted by the remarkable advances communication technologies have experienced over the same time frame, RPAs and their smaller cousins in the so-called “drone” market are poised to displace traditional pilots as the predominant form of military aircraft. There are already more RPA pilots in the USAF than in any other USAF flying community, and RPAs are flying more hours than entire MAJCOMs while achieving the highest mission readiness rates. While there will likely always be use cases where an on-board pilot is the preferable configuration, technological advances, the world’s on-going arms race, and growing fiscal restraints will invariably lead to the removal of pilots in virtually all military applications. This article will illuminate three technologies that are pushing military aviation to a remotely piloted or wholly unmanned future.
Active Electronically-Scanned Arrays
For a quick primer on AESA systems, reference section 3-7 of Naval Air Warfare Center’s Electronic Warfare and Radar Systems Engineering Handbook. Current General Atomics aircraft, the most prolific large-scale unmanned aircraft, use parabolic antennas to maintain over-the-horizon communication links with the crew in the ground control station. The finite rate at which the aircraft’s gimbals can move the antenna and the fact that these antennas are only capable of tracking a single satellite at a time has hampered RPA performance thus far. Aggressive maneuvering by the pilot, often required in close air support and other dynamic missions, can result in momentary disruption of that link due to excessive roll rates or excessive bank angles. These are colloquially known as “link hits” with extended periods resulting in a condition known as “lost link.”
AESA systems have the ability to steer their “beam” near-instantaneously which would eliminate issues with roll rate and would improve or even eliminate bank limits. Furthermore, AESA systems would allow remotely piloted systems to track multiple satellites. This would provide redundant communications links to keep an aircrew in control of their system in contested, degraded, or operationally limited environments and could do so with a low probability of intercept and with greater resistance to jamming. And if you don’t want to track multiple satellites, you could operate on multiple frequencies/transponders on a single satellite enabling higher bandwidth for advanced on-board systems, higher resolution video feeds, and so on.
Aside from their relative expense when compared to a simplistic parabolic antenna, AESA systems are superior in almost every way. This technology is well-understood, mature, and already dominant in the worlds of air-to-air and surface-to-air radar systems. To the maximum extent possible, AESA systems can and should be integrated into future remotely piloted aircraft.
Application-Specific Integrated Circuits
Current RPAs experience a 1.0-1.5 second-long delay from command input through execution and response. Approximately 0.3 seconds of this delay is due to the time required for commands to travel from the GCS to the aircraft and back, but the remaining is due to signal processing at each node along the network path. ASICs function by bringing signal processing logic out of software and into hardware eliminating a whole series of issues associated with general purpose processing. I’ll spare you the details of these issues, but suffice it to say that ASICs do one thing, do it well, and do it at “line speed“. They can be expensive to develop, but once designed, and mass-produced, this cost will diminish. We could also integrate GPUs which are many times faster than general purpose processors and less expensive than ASICs, but these still induce some latency. How low the latency needs to be is a matter of mission requirements and integration with other technologies such as AI, but so long as we desire to have a human being at the controls of the aircraft, we need minimal latency… particularly when it comes to communications.
This is a relatively minor problem when working an ISR tasking with little to no chance of things becoming more dynamic and we should be awed that the delay is as low as it is. There are situations, however, where even this latency will be detrimental not only to an RPA crew, but to all other players in the area. Let’s consider a CAS mission in a contested environment with a lot of radio traffic and multiple other assets all trying to use the same frequency. There may be a 2-ship of MQ-9s, a 4-ship of F-16s executing low altitude tactics to avoid an enemy SAM threat, 2-4 Wild Weasels, and a JTAC. This may also require a radio relay between the JTAC and any low-level fighters effectively doubling the amount of radio traffic. Needless to say, there would be a lot of information to be shared in as quick a manner as possible. In environments like this, communication flows quickly and each pilot is looking for any break in radio traffic within which he can pass a transmission. Given their proximity to the fight, the manned assets will experience near zero radio delay. The RPAs, meanwhile, will experience delayed awareness of that frequency (de)congestion and, during that delay, a more proximate asset will likely have already begun speaking. If the RPA pilot begins his radio call, there is a significant chance that he will be “stepping on” some other radio transmission effectively jamming that frequency and detracting from the overall effectiveness of the CAS team.
The simple solution here is to push the RPAs to another tasking or another frequency. More and more, this is becoming unacceptable as RPAs bring significant capabilities to a fight like this which dramatically improves the effectiveness of manned aircraft. We need to continue integrating manned and unmanned aircraft for the foreseeable future… at least until we get to an RPA-dominant force. If the RPA is to be taken seriously in combat environments like this one, we can and must reduce network latency to a minimum. This problem will not be easy to overcome given the sheer number of connections/processors between any given GCS and the aircraft it is controlling, but it is certainly possible to upgrade the military-owned nodes and begin the process of bringing this latency down to the sub-second realm.
Computer vision is an interdisciplinary field that deals with how computers can be made to gain high-level understanding from digital images or videos. Google and Facebook engineers currently use this technology to enable reverse image lookup and facial recognition with user photos. There is a smorgasbord of other public, private, commercial, and educational organizations working on computer vision technology with advances being made at regular intervals. Engineers at NASA, Carnegie Mellon University, Oxford, and more are working on CV algorithms specifically for aviation applications. With the maneuvering advantages mentioned above, on-board passive sensors or off-board cueing, and sufficiently high-fidelity imaging systems, unmanned aircraft will be able to close to visual range, identify critical components of the enemy aircraft, and destroy it with a well-aimed burst of bullets. Indeed, manned aircraft are already losing air-to-air engagements with simple, purpose-built unmanned aircraft… i.e. surface-to-air and air-to-air missiles. A multi-role unmanned aircraft with improved sensors, faster decision making, predictive AI, and better maneuverability would be even more lethal.
Unmanned Aviation is Just Hitting Its Stride
Remote and fusion warfare is poised to massively disrupt all Industrial Age militaries across the world. In addition to those listed above, there are a host of other technologies enabling the continued growth of remote and unmanned aviation… artificial intelligence, airborne mesh networks, advanced tactical data links, swarm tactics, automated refueling, advances in material science, and automatic takeoff and landing to name just a few. As with most transformative military technologies, initial progress is slow. Improvements in robustness and reliability are usually required before a nation can make that technology a major tenet of its national defense. So though Air Force culture, fiscal limitations, and philosophical debate slows the acquisition and employment of more remote aircraft, the future of airpower with remote and autonomous systems as the lynchpin is already clear.
For a more in-depth discussion on the rise of the remotely piloted aircraft, the technologies discussed in this article, and the resulting institutional changes underway, we recommend reading the Air and Space Power Journal’s paper titled Nightfall: Machine Autonomy in Air-to-Air Combat and associated papers referenced on that page.