Tuesday, June 25, 2013

Why my PhD Research in Anti-Jamming Beamforming Antenna Arrays is Important

I am researching anti-jamming beamforming antenna arrays optimized with stochastic / evolutionary algorithms. The genetic algorithm (GA) is the main algorithm that I have investigated to date. Beamforming antenna arrays (i.e., beamformers) are a type of electronic countermeasures in the sense that they can be used to focus electromagnetic / radio frequency energy on a Signal of Interest (SOI, the good guy) while simultaneously minimizing energy in the direction of interferers / jammers (i.e., the bad guys). Today, I want to explain why my PhD research is important for both commercial and military applications. I've included a YouTube video explaining the importance of my research below.


People love to use wireless, mobile devices because they can go with you wherever you go. Sure, laptop computers are great, but they're bulky compared to smart phones and tablets like the iPhone and iPad, etc. Ericsson Corporation also noted that the number of mobile subscriptions (assuming one device per subscription) exceeded the five billion mark in July 2010, and they expect this number to increase 10 fold by 2020. What this means is that the frequency spectrum will get saturated, and interference is bound to occur. Frequency sharing / scheduling is an option, but it will not completely solve the problem because the useable frequency spectrum is limited. Wireless telecommunications companies can buy spectrum from regulating agencies like the FCC. However, this is very expensive.

A solution that my PhD advisor proposed (when I began my PhD) is to evolve an anti-jamming beamforming antenna array using a genetic algorithm. There were several reasons we followed this route. First, he successfully optimized antennas using genetic algorithms with built antennas whose behavior matched that of simulations. The GA found antenna shapes that met project requirements. We believed that we could adapt hardware settings to evolve an array in situ with hardware. In this sense, the evolvable hardware was in the settings of the array and not the physical shape of the array. Second, although beamforming array technology was not new, it used hardware versions of gradient descent to form the array beam (i.e., electromagnetic radiation pattern in space). This resulted in systems that were contained in large chassis and cost on the order of $1 million. Such a system would not be feasible for commercial use, and we believed that we could reduce the hardware cost by shifting the optimization algorithms from hardware to software. I built a four antenna array with phase shifters, step attenuators, and controller hardware as shown in Figure 1. I've discussed my array in a previous post.

Figure 1: Picture of My First Beamforming Array Prototype (Copyright Jonathan Becker)
As I mentioned in my video, I've published several papers showing that the GA successfully adapts the array to focus energy on an SOI while simultaneously minimizing energy in jammer directions. My first prototype array thwarted two jammers very well and had moderate performance with three jammers. As you can tell from Figure 1, this prototype is certainly not ready for commercial use. In fact, it is at a Technology Readiness Level (TRL) of 1 or 2, and the cost of the hardware was roughly $5000. This is much better than $1 million, but it is not good enough for a commercial application.

Here's where I stand in my PhD research. I want to build an eight antenna array with surface mount components and printed antennas (such as patch antennas). I believe that eight printed antennas would be sufficient to thwart the number of signals seen in a typical wireless environment, and I can design a printed circuit antenna array that could be integrated with a laptop computer. I am also investigating other stochastic algorithms such as Particle Swarm Optimization (PSO), and I'm considering using wideband antennas. I've read research papers with PSO simulations that indicate that PSO performs better than the GA in terms of convergence time and solution quality, and I want to verify that I see such improvements in hardware. I'm also considering wideband antennas, so one can use the array on multiple frequency band. Imagine using one array for WiFi, WiMax, and 4G wireless communications. In essence, I am aiming for a TRL of 5 or 6. The second prototype might not be ready for commercial use by the time I'm done with my PhD, but I hope that it will be pretty close to being ready.

The importance of my research for commercial application is clear: My PhD research has the goals of creating a small formfactor antenna array that is inexpensive and can quickly adapt to changing environments and changing mobile signals (both desired and interference). I also believe that this technology will be ubiquitous in the sense that it could be integrated into your laptop, and you wouldn't know that it's there and operating. I would love to see this technology integrated into smartphones and tablets. However, it is currently limited in size because antennas take up bulk of the array's physical size. One would need to make sacrifices in terms of number of antennas used, electrically small antennas with low efficiencies, and so forth in order make it ready for use with smaller mobile devices.

In addition, this technology has multiple military applications. First, the array could be mounted into unmanned aerial vehicles (UAVs) to maintain secured communications between the UAV and mission control. The beamformer would need to be small and lightweight in order to operate inside a small UAV. There was the case a couple years ago of the stealth UAV that was supposedly highjacked in Iraq and forced to land in Iran. Assuming that the Iranian engineering told the truth, he jammed the UAV's wireless control signal and sent it a spoofed control signal that tricked it into thinking Iran was its homebase in Afghanistan. If the stealth UAV had an anti-jamming beamforming array mounted inside it, this situation would not have been possible. Second, the beamformer could be mounted into various military vehicles to ensure that there is secure and uninterrupted communications between the vehicle and homebase. In military field operations, the enemy wants to jam our wireless communication signals to create confusion, disorder, loss of property, and loss of life. It is to the military's advantage to make sure that this doesn't happen. Third, the Department of Defense (DOD) has been under pressure to cut budgets, and this is very true today considering the sequester's effects. A low cost beamformer would be desirable as a result.

In summary, this is why my research is important. I've also explained what research I have done, and what I hope to accomplish in the near future. I hope that you enjoyed reading this article.


Sincerely,

Jonathan Becker
ECE PhD Candidate
Carnegie Mellon University