ALGORITHM OF COMPUTING THE AMBIGUITY FUNCTION IN THE TASK OF SIMULTANEOUS ESTIMATION
OF TIME-FREQUENCY PARAMETERS OF SIGNALS 

Loginov Aleksey Andreevich, Candidate of physical and mathematical sciences, senior stuff scientist, Physical-Technical Research Institute of Nizhny Novgorod State University named after N. I. Lobachevsky (23 Gagarin
avenue, Nizhny Novgorod, Russia), loginov@nifti.unn.ru
Marychev Dmitriy Sergeevich, Postgraduate student, Nizhny Novgorod State University named after Lobachevsky (23 Gagarin avenue, Nizhny Novgorod, Russia), dsmarychev@nifti.unn.ru
Morozov Oleg Aleksandrovich, Doctor of physical and mathematical sciences, associate professor, sub-department of information technology in physics, Nizhny Novgorod State University named after Lobachevsky
(23 Gagarin avenue, Nizhny Novgorod, Russia), loginov@nifti.unn.ru
Fidel'man Vladimir Romanovich, Doctor of engineering sciences, professor, Head of sub-department of information technology in physics research, Nizhny Novgorod State University named after Lobachevsky (23 Gagarin avenue, Nizhny Novgorod, Russia), loginov@nifti.unn.ru 

621.396 

Background. Among widely-known and promising methods of estimating different time-frequency parameters of signals are those based on computing and analysis of the ambiguity function (or the cross-ambiguity function). In different tasks of radio navigation and estimation of location of the radio-emitting engines such characteristics as time delay of arrival (TDOA) and frequency difference of arrival (FDOA) are extremely important. These appear when the signal spreads from the emitter to a number of spatially separated receivers. Both emitter and receivers might make arbitrary moves in the space as well. The ambiguity function in this case is used for simultaneous assessement of TDOA and FDOA. The TDOA is commonly used in the range-difference method of location of emitting sources, at the same time FDOA is used to estimate their velocities. The most serious problem in practical usage of the ambiguity function is a large amount of computations taken. The main purpose of the work is to develop an efficient method of computing the ambiguity function which would be suitable for implementing in modern graphics hardware.
Materials and methods. All the results presented in the paper have been obtained by computational modeling of the proposed algorithm on modern graphics processors and multi-core processors of general purpose.
Results. A method of computing the ambiguity function has been proposed. The method is based on parallel implementation of computing linear convolution. It has been proved that the method is quite applicable in practical cases for graphics hardware. Comparison of existing parallel approaches to computing the ambiguity function with the proposed one is also presented.
Conclusions. The results presented in the paper show a significant increase in computing the ambiguity function in case of using the proposed approach. The importance of the result enables to achieve adequate capacity without any special hardware like DSP being involved. This is especially important for multi-channel systems.

ambiguity function, radar, parallel computing, mutual time delay. 

1. Woodworth P. M. Probability and Information Theory with Applications to Radar. Pergamon Press, 1953.
2. Maks Zh. Metody i tehnika obrabotki signalov pri fizicheskih izmerenijah [Methods and technology of signal processing in physical measurements]. Moscow: Mir, 1983, vol. 2,256 p.
3. Levanon N., Mozeson E. J. Wiley & Sons, Inc New Jersey, 2004, 411 p.
4. Grishin Ju. P., Ipatov V. P., Kazarinov Ju. M. et al. Radiotehnicheskie sistemy: ucheb. dlja vuzov po spec. «Radiotehnika» [Radio engineering systems: university tutorial for “Radio engineering” specialty]. Moscow: Vyssh. shk., 1990, 496 p.
5. Loginov A. A., Morozov O. A., Soldatov E. A., Hmelev S. L. Izvestija vuzov. Radiofizika [University bulleting. Radio physics]. 2007, vol. L, no. 3, p. 255.
6. Loginov A. A., Morozov O. A., Hmelev S. L. Avtometrija [Autometrics]. 2010, vol. 46, no. 6, p. 40.
7. Kirk D. B., Hwu W. W. Morgan Kaufman Publishers, 2010.
8. Tolimieri R., Winograd S. IEEE Trans. Acoust., Speech, Signal Processing. 1985, vol. ASSP-33, no. 4.
9. Yatrakis C. L. Computing the cross ambiguity function – a review. Binghamton University, State University of New York, 2005, 131 p.
10. Johnson J. J. Implementing the cross ambiguity function and generating geometryspecific signals. Thesis, Naval postgraduate school, Monterey, California, 2001.
11. Ajficher Je., Dzhervis B. Cifrovaja obrabotka signalov: prakticheskij podhod: per. s angl. [Signal digital processing: practical approach: translation from English]. Moscow: Vil'jams, 2004, 992 p.
12. Stein S. IEEE Trans. Acoust., Speech, Signal Processing. 1981, vol. 29, June. pp. 588–599.
13. Auslander L. IEEE Transactions on. 1988, vol. 36, pp. 359–364.
14. Intel Integrated Performance Primitives for Intel Architecture. 2007, vol. 1: Signal Processing, 1352 p.