Bats make their living on the wing, racing in and around often-complex habitats – and they do it in the dark. That makes studying their behavior and travels very difficult. But our ability to monitor bat behavior has thankfully evolved in recent decades beyond simple banding studies, in which you mark a captured bat with a numbered band and hope to catch it again later. Today’s high-tech tracking techniques include GPS tags (which identify precise geographical coordinates), satellites, Doppler radar and radiotelemetry. And still bats are tough to track.
So the idea (suggested by Sebesta) of tracking bats with a low-cost “unmanned aerial vehicle” – a deftly modified model airplane – was simply too powerful to ignore. We jumped at the chance, and a BCI Student Research Scholarship helped launch this unique effort.
The result is the Automated Tracking and Localization Aerial System (ATLAS): a collaborative effort by engineers and biologists to automate radiotelemetry using off-the-shelf model airplanes and affordable electronics. Although still in its early stages, we believe this project will ultimately make it possible to track radiotagged bats using a fully autonomous airborne platform that costs less than $500 in parts.
And no, we aren’t kidding.
Traditional radiotelemetry is a rather straightforward process, at least at first glance. First, you capture a bat and attach (usually with a short-lived glue) a tiny radio transmitter that sends out faint radio pulses at regular intervals. Then you establish at least three separate ground stations – stationary, on foot or in cars – with radio-receiving gear. The receiving antennas are rotated until they find beeps emitted by the bat’s transmitting backpack. Then the field crews follow those beeps through the night as if their lives depended on it. This continues for days, for as long as the transmitter is viable.
If that sounds exhausting, you are getting a feel for the challenges of radiotracking. Bats pay no attention to roadways, regularly fly over mountains, rivers and highways and show no interest in property rights. Weary human field crews, however, must contend with all these factors and many others while driving, collecting data and checking maps. And then there’s the frequent need to defuse misunderstandings with local residents or police patrols. In a word, bat tracking is hard core.
Life for radiotrackers gets a bit easier if you are lucky enough to have access to an airplane, which sails over roads and provides superior signal quality since radio pulses are rarely disrupted by foliage or terrain. The drawback, of course, is the extreme cost of aircraft. The simplest of airplanes costs more than $150 an hour to rent, plus fuel and pilot fees ($40 to $80 per hour). An entire season of bat-tracking with airborne telemetry is prohibitively expensive.
This brings us back to ATLAS, which, we hope, will provide the best of both worlds: the quality of airborne tracking with a ground-level price tag. With this system, low-cost ground crews will be uplinked to what amounts to a self-guided, flying antenna that’s in almost continuous contact with a flying bat, while calculating its own position and that of the bat using an onboard GPS receiver.
ATLAS includes three components that comprise the ears with which it listens to signals from the bat’s radio backpack. The outer ear consists of two external directional antennas strung between the wings and tail of the airplane, one aimed 45 degrees to the left of the nose and the other to the left of the tail. This means antenna coverage is exclusively to the plane’s left side.
ATLAS’ middle ear is a pair of tunable “software-defined radio” (SDR) chips that process signals from the antenna. The SDR chips send this pulse into the inner ear, a tiny Raspberry Pi computer that runs signal-processing software developed by colleagues in Luxembourg. The final result – a directional interpretation of the signal pulse – is logged to an onboard SD memory card and transmitted to the onboard flight-control system.
Once ATLAS hears a bat, it needs to maneuver to the bat’s location. This requires an autopilot. It turns out there’s a good reason why the first autopilot system was developed in the 1920s: autonomous flight is surprisingly easy. So long as your aircraft maintains velocity and attitude, it will stay in the air and, in general, be quite stable. ATLAS uses OpenPilot hardware and software to guide its flight. This powerful open-source flight-control system maintains flight stability by using onboard gyros to determine the plane’s orientation, then adjusting flight surfaces to compensate for any instability.
OpenPilot can be loaded with a predetermined flight trajectory that the aircraft will follow when it receives certain inputs. This means that we can launch the airplane autonomously at the push of a button. It will follow a search pattern to hunt for a signal from a radio-tagged bat, lock onto the detected signal and circle it, then return home for battery recharge and data offload at the push of another button. OpenPilot is ATLAS’ brain.
And within that brain is the flight-control system that really puts ATLAS to work. While searching for a bat flying low through a forest, ATLAS flies at an altitude of 350 feet (106 meters) as it listens for the soft radio blips from below. Once a signal is detected, the system puts ATLAS into an orbit around the source, a flight pattern that permits a very effective triangulation technique. After a few seconds, ATLAS should be able to locate the source – the radio-tagged bat – to within a reasonable distance, thus localizing a moving or stationary bat at least as well as a conventional radio-tracking ground crew.
So far, every component of ATLAS works independently. And the complete ensemble works in simulation. We have a model airplane that flies autonomously and an efficient signal-processing unit. Now these two components must be connected and tested in a working aircraft. We are working on this connection and plan to be tracking bats autonomously by the summer of 2013.
Our hope is that ATLAS will one day change the way wildlife radiotelemetry is conducted. It should be able to track almost any elusive organism from the air at a fraction of the cost of traditional radiotelemetry techniques.
This would mean, for example, that far- and fast-flying species, such as Mexican free-tailed bats or migratory tree bats, could be tracked much more effectively than current technology allows. Migratory birds or even insects such as monarch butterflies might be tracked and their paths analyzed to assess the danger of wind turbines and pesticide application. A detailed understanding of the pre- and post-hibernation movements of hibernating cave bats should help us understand how diseases such as White-nose Syndrome are spread.
The sky really is the limit for what ATLAS may accomplish in tracking small airborne organisms. Our real goal is to make wildlife tracking a lot less difficult – but still hard core.
NATHAN W. FULLER is a Ph.D. candidate in biology at Boston University. KENNETH D. SEBESTA, Ph.D. is a visiting scholar in the Department of Mechanical Engineering at Boston University.