A.B. Natalin - Ph. D. (Eng.), Associate Professor, Department of Radio Engineering Theory, Electrotechnical University «LETI» (Saint Petersburg) E-mail: abnatalyin@yandex.ru A.A. Samolyanov - Engineer, Scientific Laboratory of Radios Systems and Signal Processing, Electrotechnical University «LETI» (Saint Petersburg) E-mail: samol@mail.ru A.B. Sergienko - Ph. D. (Eng.), Associate Professor, Department of Radio Engineering Theory, Electrotechnical University «LETI» (Saint Petersburg) E-mail: sandy@ieee.org

Signals with large time-bandwidth product (TBW) are widely used in radar and radio navigation systems. Their popularity is due to the fact that matched filtering at the receiver can simultaneously provide good resolution along time and frequency and also solve a number of electromagnetic compatibility problems. But if wideband signal generators are based on analog technology matched filtering become a problem as parameters of analog transmitters are greatly varied from sample to sample, also they have a drift because of a number of environment factors. In this article an approach is proposed for the design of wideband digital receiver based on adaptive matched filtering. After startup the receiver should have an initial approximation for the filter impulse response. The article presents an algorithm that implements initial search with subsequent estimation of the impulse actual shape. The chirp impulse is used as an illustrative example. After obtaining estimations of the actual signal parameters the receiver is able to perform matched filtering adapting to slowly varying signal parameters. In this article we propose adaptation method that can be implemented digitally. Such implementation provides power loss under 1 dB comparing to ideal matched filter. The article also presents a method to improve accuracy of time of arrival estimate in the case when the number of signal samples per output (correlation) peak is extremely low (about 3). This method is interpolation-based, its loss relative to Cramer-Rao bound is about 0,5 dB. The proposed algorithms have been experimentally verified with laboratory setup of digital adaptive matched filter described in the article. Based on the review of the state-of-the-art digital components the upper bounds are derived for TBW and bandwidth of signals that can be processed by proposed algorithms.

 

1. Kuk CH., Bernfeld M. Radiolokacionnye signaly. Teorija i primenenie. M.: Sov. radio. 1971. 568 s.
2. Ipatov V.P. SHirokopolosnye sistemy i kodovoe razdelenie signalov. Principy i prilozhenija. M.: Tekhnosfera. 2007. 488 s.
3. Grishin JU.P., Ipatov V.P., Kazarinov JU.M. Radiotekhnicheskie sistemy: Ucheb dlja vuzov po spec. «Radiotekhnika» / Pod red. JU.M. Kazarinova. M.: Vysshaja shkola. 1990. 496 s.
4. Sergienko A.B., CHekunova I.S. Opredelenie polozhenija maksimuma signala pri interpoljacii po trem tochkam // Izvestija VUZov. Radioehlektronika. № 4. 2005. S. 51−55.
5. Sergienko A.B., CHekunova I.S. Ocenka polozhenija korreljacionnogo pika pri trekhtochechnojj interpoljacii kompleksnogo signala // Dokl. 8‑jj Mezhdunar. konf. «Cifrovaja obrabotka signalov i ee primenenie» (DSPA‑2006). Moskva. 29−31 marta 2006 g. S. 168−172.
6. Kreveckijj A.V., Melnikov A.D. Razreshenie-obnaruzhenie signalov na baze soprjazhennykh soglasovannykh filtrov // Radiotekhnika. 2007. № 4.