Wireless communication has become ubiquitous and personalized. It is used more and more indoor as well as on the road. Measurements and network monitoring are essential for the operation of any cellular radio network. Terminal measurements and reports, as collected by the network, are a viable source that can fill in the gaps of drive and walk tests, especially since the latter are confined to roads and public places. Those reports, once enhanced by a location engine that attributes geographic coordinates to each call, provide a focused view on problematic areas and enable problem solution and network performance improvement based on actual user usage patterns. The combination of terminal reports together with a location engine for the purpose of network optimization is, what we call, a Virtual Drive Test. In this article the concept of the virtual drive test will be defined and explained, including the geolocation principles on which it is based. Network optimization using the virtual drive tests will be shown and demonstrated using existing cellular networks.

Modern communication systems are required to provide services based on high data rates burst-mode packet-data transmission, capable of operating at very low SNR conditions. Turbo codes enable the operation at low SNR, close to the Shannon limit. However, carrier frequency and phase synchronization, needed for optimal coherent performance of the receiver, still remains a problem in low SNR and short bursts conditions. This paper proposes a new carrier synchronization method, the Maximum-Likelihood Iterative-Soft-Decision-Directed (ML-ISDD), which uses the turbo-decoder soft decisions to improve the carrier synchronization performance at low SNR values. The ML-ISDD method operates iteratively and jointly with the turbo decoder, enhancing both the turbo-decoder and the synchronization performance. The ML-ISDD method has been shown by simulation to significantly increase the allowed initial frequency and phase uncertainty region, thus allowing the use of very short training sequences for initial carrier synchronization.

Turbo codes, which provide superior performance, close to the Shannon limit, have found their way into WLAN and mobile system standards. Synchronization of turbo-coded signals operating at low SNR conditions may be difficult, especially for short packet transmission. This paper proposes two techniques for using the turbo decoder information to improve the carrier phase and frequency offset synchronization for short packets communication systems. One of those techniques, which are based on soft-decisions produced by the Turbo decoder, can be used for tracking. The other technique is based on maximizing a newly invented cost function, the MSSO (/spl Delta/f,/spl Phi/) function, to acquire and track the carrier frequency and phase errors, with the cost of complexity. A combination of the two techniques is introduced as well to provide a wholesome solution.

Inverse synthetic aperture radar (ISAR) is a technique used to obtain an image of the radar target. It can be described as a spectral analysis problem. This paper discusses the application of the minimum free energy (MFE) spectral estimation algorithm for the purpose of ISAR imaging. The MFE algorithm is a regularization algorithm which achieves a better noise performance than other methods studied previously. The paper discusses the choice of parameters for the MFE technique, and presents ISAR images obtained using MFE for various observation times. MFE indeed achieves a 3-4 times higher resolution than the classic Fourier technique, without prominent spikes in the noise region. However, some of the details observed in the Fourier image are lost. Using diversity (incoherent) combining, it is shown that based upon the same RF data, MFE and the Fourier techniques produce images of similar quality, although MFE poses a less stringent requirement on the length of the coherent observation time. The MFE can be applied to ISAR imaging of close to radial targets.

Costas signals have an ambiguity function that approaches the ideal "Thumbtack" configuration. To increase the main lobe/side-lobe ratio without increasing the number of frequencies (signal dimension, N) it may be possible to stagger Costas signal pulses whose ambiguity sidelobe patterns do not coincide. We prove here that Costas signals of the same dimension, with completely different ambiguity sidelobe patterns, do not exist for N > 3.