University of Pretoria
A poacher in the Kruger National Park has just made a kill. He takes out his cellphone to call it in: his mission was a success. What he does not realise is that by using his phone he is transmitting a radio wave that can be picked up for many kilometres all around. We can intercept his signal and tell which direction it is coming from. Anti-poaching rangers can be sent in search of the culprit and catch him before he leaves the area. His private call is not that private, after all.
Our research team at the University of Pretoria, led by Prof Warren du Plessis, is developing a new direction-finding signal system that is cheaper than existing systems but just as accurate. This opens up uses for the technology which, to date, has been considered too costly for many applications. Direction-finding systems are useful in many military scenarios, locating illegal radio transmitters, and in areas of security such as anti-poaching efforts.
A direction-finding antenna array consists of multiple antennas working together to estimate where a radio wave is coming from. Traditionally, each antenna requires its own receiver – the part that captures the radio wave and converts it into a form that we can use. But this can quickly become expensive as receivers often cost much more than the antennas themselves.
Our design aims to reduce cost by combining or compressing the signals from the antennas, meaning that we need fewer receivers. While an accurate estimate of location requires the use of multiple antennas, there is really only one piece of information we are interested in – the direction that the radio wave is coming from. Conventional systems use many ears to listen to what is effectively very little being said. If we are only interested in the direction, we do not need as many receivers.
When multiple antennas are placed next to each other (as in a conventional array), a radio wave will typically not reach all of them at the same time. This means that there is a time delay between when the signal reaches each antenna, and using this delay, we can work out the direction the wave is coming from.
Imagine standing on a beach, where the waves lap onto the shore. A friend stands some distance away, their toes touching the shoreline. If a wave comes towards the shore, head-on, it will reach both of you at the same time. If the wave comes from your side, it will reach you first and then your friend. If you know the speed of the wave, you can tell its direction simply by timing the delay between when it first reaches each of you.
The time delays in a well-designed antenna array are a unique description of the incoming direction of the radio wave. In other words, the delays between the antennas result in a special “signature” for that direction which sets it apart from all other directions. This allows us to find the direction of a radio transmitter by comparing the received signals and seeing which signature it most closely resembles.
In the case of our new design, we combine the signals after they hit the antennas, but before the receivers capture them. Instead of the simultaneous shouting from all the antennas, the receiver will receive a special weighted combination of signals from all the antennas. Each signal is multiplied by a specially chosen number before being added to the rest of the similarly weighted signals to be sent to that receiver.
Now that there are fewer receivers, the job of listening to the antennas must be shared among them all. Each antenna shouts to the receivers and each receiver must listen to all the antennas, so that all the antennas are being heard to some degree. But the receivers cannot give the antennas the same amount of attention as they did to one antenna. Even so, each receiver pays more attention to some antennas than to others.
This attention takes the form of a weighting: each antenna signal is multiplied by a specific weighting before it is sent to the receiver, so that some are heard more loudly than others, depending on their location relative to the signal direction. Although there was much talk, the system knows that it is looking for only one piece of information, which helps it to make sense of the confusion.
The compressed direction-finding array also has special signatures for each direction, but using fewer receivers has the potential to result in signatures which are similar to each other – something which we must avoid. We must choose the compression weights carefully to ensure that each signature is as different as possible from all others.
In order to design the new direction-finding system we are developing software which chooses the compression weights for us, as it would take a very long time to solve this problem by hand.
So far our results are promising, but there are still many practical aspects to consider. The next stage in the design process is to build a prototype system and test it in real-world conditions. If the design performs better than a conventional system of similar cost, this will prove the usefulness of our design and hopefully encourage others to start following a similar approach in designing better direction-finding systems for real-world applications.