A Low-cost Mechanically-Steered Phased-Array Polarimetric Doppler Weather Radar


Project based at the
Fraunhofer-Institut für Hochfrequenzphysik und Radartechnik FHR

Fraunhofer-Institut für Hochfrequenzphysik und Radartechnik FHR: Stefano Turso (PI)

Project Goals
Dense networks of inexpensive short-range weather radars might yield a fundamental advancement for the understanding of weather phenomena and the timely reaction to extreme events. Due to the Earth curvature, Figure 1a, about 70% of the troposphere below 1 km cannot be observed by radar means. Consequently, traditional long range weather radars (up to about 200 Km range)are unable to provide coverage of the lower part of the atmosphere where most of the interaction with weather phenomena actually takes place.Composition of short-range weather maps, Figure 1b, appears therefore a viable solution to improve sensing of the lower troposphere, enhance spatial resolution and achieve a revisit time shorter than one minute [1].



Fig. 1a. Earth curvature effect Fig. 1b. Short-range weather maps composition concept.

The Fraunhofer Institute for High Frequency Physics and Radar Techniques (FHR)in partnership with the Institute for Geosciences and Meteorology at the University of Bonn is designing and developing a novel Phased-Array Radar (PAR)to be operated within the framework of PROM. The system will be deployed close to the twin X-band Doppler radars at the JOYCE Core Facility and extensively tested for data quality assessments. Sustainable deployment of a dense network of weather radars strictly requires low unitary and ownership costs. To this end, electronic steering is limited to a single dimension (to scan the zenith angle) while a rotor provides mechanical steering in azimuth, as shown in Figure 2.


Fig. 2a. Flat panel concept, front.


Fig. 2b. Flat panel concept, back.

While a simple mechanical rotation can provide full azimuth coverage, at least electronic beam steering in elevation is necessary to achieve timely volumetric sounding of the troposphere. This allows to overcome the major limitation of current operational polarimetric weather radars, which typically require up to five minutes to mechanically steer a large dish in two dimensions and provide complete sensing of the hemisphere.

Technological advancements like the availability of highly integrated MMICs promise to lower fabrication costs to a point of competition with mechanically scanned weather radars [2]. Distributed power generation removes the need of a costly power amplifier and rotary joint, reduces losses on transmit and grants graceful degradation. The entire back-end circuitry, including the first stage raw data processing and reduction, can be located right behind the panel (following the receiver over-elevation approach) to improve the receiver noise figure. Preprocessed datasets can be transmitted to a server via radio-link for further processing.

A new generation of RF chipsets featuring high level of functional integration allows for reduction of cost, complexity and development risk at X-band. A few integrated solutions for electronic steering are commercially available on the civilian market. Quad-core controllers able to implement electronic steering over four channels can drive fully integrated front-end chipsets offering complete transmit and receive functionalities. The resulting functional block, the Active Row Module in Figure 3a, offers a modular and scalable implementation of electronic steering functions and constitutes the building block of the entire aperture (Figure 3b)


Fig. 3a. The active row module.


Fig. 3b. Electronically steerable panel.

Current Status, Front-end
A set of chipsets offering high scale of integration is selected. The Anokiwave AWS-0103 quad core controller, Figure 4a, can handle the core AESA functionalities and support four independent branches. Each branch features 6 bits phase shifting (a phase step of 5.6°) and gain adjustments of 0.5 dB. Transmit and receive gain blocks up to 21 and 7 dB are available to provide flexible tapering options. Fast logic programming and a weights caching mechanism offer sufficient performance for common beam steering tasks. Transmit and receive blocks are as well packaged within a single unit (the Anokiwave AWMF-0106, Figure 4b) offering switched front-end functionalities. Specifically, the complete transmit chain is integrated on-chip (with power amplification up to 30 dB linear gain) as well as the receiver stages (limiter and low noise amplifier up to 24 dB linear gain). Time critical transmit/receive switching is based on analog levels. Temperature and emitted power monitoring are integrated to support essential in-line calibration functions. This so called Medium Power Front End (MPFE) provides up to 4 Watt output power, features a noise figure as low as 2.8 dB in receive mode and offers about 4 GHz of bandwidth to support frequency sweeping.



Fig. 4a. Quad-core prototype PCB. Fig. 4b. Quad front-ends prototype PCB.

PCB development, interfacing and testing are ongoing in partnership with the chipset manufacturer to ensure optimal use of the chipsets capabilities.

Current Status, Polarimetric Antenna
Matching the same data quality of bi-axial mechanical solutions strictly requires unbiased estimation of polarimetric moments for consistent classification of hydrometeors over the entire scanning volume [3]. Consequently, one of the most stringent requirements specific to weather radars is the achievement of a cross-polarization discrimination (XPD) in excess of 30 dB [4]. Therefore, investigations on methods to improve the polarization purity have been carried on to ensure feasibility within cost constraints. As an outcome, an antenna tile has been engineered in the form of a scalable sub-array, Figure 5, and finely tuned to reach the required polarization purity.



Fig. 5. Antenna sub-array.

Ongoing validation mostly supports theoretical expectations concerning the polarimetric performance, also for beam directions off-broadside. Threedimensional radiation patterns and XPD plots in (𝑢, 𝑣) coordinates, Figures 6, show good agreement in between simulations and measurements, with XPD levels above 30 dB at a beam steering angle of 45°.



Fig. 6a. Row module radiation diagram and (u,v) XPD plots, simulations.
Fig. 6b. Row module radiation diagram and (u,v) XPD plots, measurements.

Development of a suitable low-cost AESA front-end is a crucial step towards sustainable deployment of short-range netted weather radars able to generate composite maps with high space and time resolution, and finally improve monitoring of the lower troposphere. Early results from this recently started weather radar project at Fraunhofer FHR are summarized, showing a potential for novel concepts and designs to meet strict polarimetric requirements and generate valid weather radar observations within cost-constrained solutions.


References
[1] “Atmospheric Sciences: Entering the Twenty-First Century,” p. 181, National Academy Press, 2101 Constitution Avenue, NW Washington, DC 20418 USA, 364 pp. 1998, ISBN 0-309- 06415-5.

[2] S. Turso, T. Bertuch, M. Jäger, S. Stanko, P. Knott, S. Trömel, C. Simmer, “A low-cost mechanically-steered weather radar concept,” in 2018 IEEE Radar Conference (RadarConf18), Oklahoma City, OK, 2018, pp. 1491–1494.

[3] D. S. Zrnic and R. J. Doviak, “System Requirements for Phased Array Weather Radar”

[4] G. Zhang, R. J. Doviak, D. S. Zrnić, R. Palmer, L. Lei, and Y. Al-Rashid, “Polarimetric PhasedArray Radar for Weather Measurement: A Planar or Cylindrical Configuration?,” in J. Atmos. Oceanic Technol., 28, pp. 63–73, 2011, doi: 10.1175/2010JTECHA1470.1.