Northern Prairie Wildlife Research Center
Hugh J. Spencer, Australian Tropical Research Foundation, Cape Tribulation Tropical Research Station, PMB 5 Cape Tribulation 4873, Australia, Frank Savaglio, and Doug Maskell, Department of Electrical and Computer Engineering, James Cook University, Townsville, Queensland 4811, Australia
Because of the difficulties that we have encountered, both logistical and financial, in intensive radio-tracking of small animals, particularly of bats and birds, over any extended period, we have developed an automatic radio-location system. Its operation is based on determining the time-of-arrival of the transmitted signal to three receiving stations, an approach also known as "hyperbolic navigation."
As the speed of propagation of electromagnetic waves is virtually the speed of light, time measurements have to be made with considerable precision, as each 10 nS delay represents 3 m distance on the ground. Attaining this level of precision has been the primary difficulty encountered in the past by previous workers in translating this concept into an operating system, especially in the presence of the substantial noise and multipath signals characteristic of the real environment in which animal tracking studies are carried out.
Our system employs spread-spectrum technology to reduce the susceptibility of the received signal to interference, and hence reduce the transmitter power required for reliable transmission. All transmitters operate on a common carrier frequency and generate a pulsed signal which can have a repetition interval ranging from seconds to hours, depending on the temporal resolution of the animal's movements required. A central station processes the received signals which are relayed through line-of-sight microwave links. All transmitters transmit a common preamble code which is detected using a spread-spectrum receiver, and which provides a warning to the system that a transmission from that transmitter is arriving. The transmitter then transmits a unique pseudo-random code which, besides identifying the transmitter, is used to determine its times of arrival at the receiving stations.
From initial field tests of the prototype system, we have achieved accuracies of +/- 25 m or better, and this accuracy is maintained over the entire study area, which is bounded by the 3 receiving stations. This accuracy has been maintained at signal-to-noise ratios of -5 dB. We anticipate that these accuracies can be improved at least by a factor of 4. Provided the signal strength is adequate, the coverage area can range from 1 km2 to tens of km2, with no loss in positional accuracy. Transmitter radiated power required is of the order of 10 to 50 mW, and this can be readily produced by current transmitter technology, and the system is currently able to be used on animals capable of carrying a 25-g or larger package. Because the transmitted signals are pulsed with long inter-pulse intervals, the system can accommodate a large number of transmitters in the field.