New Radiometer Concepts For Ocean Remote Sensing

Transcript

New Radiometer Concepts for Ocean Remote Sensing: Description of the Passive Advanced Unit (PAU) for Ocean Monitoring A. Camps, J. F. Marchan-Hernandez, I. Ramos-Perez, X. Bosch-Lluis, R. Prehn UPC Campus Nord, Building D3, tel. +34+934054153, e-mail: [email protected] E-08034 Barcelona, Spain Abstract— Sea surface salinity can be remotely measured by means of L-band microwave radiometry. However, the brightness temperature also depends on the sea surface temperature and on the sea state, which is probably today one of the driving factors in the salinity retrieval error budgets of SMOS and Aquarius/SAC-D mission. This work describes the architecture design of the Passive Advanced Unit (PAU) for Ocean Monitoring, its subsystems and its main characteristics. PAU combines in a single instrument three different sensors: an L-band radiometer with digital beamforming to measure the brightness temperature of the sea, a GPS-reflectometer also with digital beamforming to measure sea state, and an infrared radiometer to provide sea surface temperature estimates. The key characteristic of this instrument is the fact that both the Lband radiometer and the GPS-reflectometer share completely the RF front-end, and to be able to track the GPS-reflected signal is not possible to chop the antenna signal as in a Dicke radiometer, therefore a new radiometer topology has been devised. I. INTRODUCTION It is well known that sea surface salinity can be remotely measured by means of L-band microwave radiometry [1]. At present, two space-borne missions are planned to be launched in the near future with this purpose: ESA’s SMOS mission, using a Y-shaped synthetic aperture radiometer, and NASACONAE AQUARIUS/SAC-D mission, using a tree beam push-broom radiometer. Sea surface salinity can be indirectly measured through the variations of the brightness temperature due to the change of the sea water dielectric constant with respect to temperature and salinity. However, the brightness temperature also depends on the sea surface roughness, which cannot be simply parameterized in terms of the wind speed, the significant wave height or any other currently available parameter. Therefore, despite the field experiments performed in the past years to improve our understanding of this effect [2], the sea surface roughness correction still remains one of the most critical corrections needed to retrieve the salinity with the required accuracy, and inaccuracies in the brightness temperature direct model may induce significant errors (biases) in the retrieved salinity [3]. In the SMOS mission, as first proposed in [4], the multiangle observation capabilities allow to simultaneously retrieve not only the surface salinity, but in addition, the surface temperature and an “effective” wind speed that minimizes these errors. In AQUARIUS an L-band scatterometer will be used to perform the necessary sea state corrections. The potential use of GNSS-R opportunity signals for altimetry [5] and sea state determination in terms of the mean square slope (mss) has been already tested from ground-based [6-9], airborne [10-17], and spaceborne [18,19] experiments. However, the underlying science still needs further refinements to extract meaningful physical quantities that can be successfully used in the remote sensing and oceanographic communities. Probably, one of the main advantages of this technique lies in the capability to obtain simultaneous and collocated sea state information, which is not possible with other auxiliary data sets [3]. A performance study of the capabilities and requirements of this technique was recently presented in [20], applied to the SMOS case. The Passive Advanced Unit (PAU) is a new concept radiometer that was proposed in 2003 to the European Science Foundation (ESF) within the frame of the EURYI program, and was funded in 2004 [21]. It consists of a suite of three instruments operating synergetically: 1) PAU-RAD: an Lband radiometer to measure the brightness temperature of the sea surface, 2) PAU-GNSS-R: a GPS-reflectometer to measure the sea state, and 3) and PAU-IR: and IR radiometer to measure the sea surface temperature. The L-band radiometer is an array of 4 x 4 dual polarization receivers [22] integrated behind a patch antenna, whose outputs are digitized and then properly calibrated and combined to produce several beams using a digital beamformer [23]. The GPS-reflectometer uses the same hardware as the L-band radiometer, but the digital beamformer synthesizes beams pointing towards the specular reflection points of the GPS signals [24]. Finally, the third element is a commercial 8-14 µm thermal IR radiometer. In order to be able to re-use the same receivers for both the radiometer and the GPS-reflectometer, a new radiometer topology –to author’s knowledge- has been devised. Instead of connecting the antenna output to the radiometer receiver, it is connected to a Wilkinson power splitter that divides the signal in two in-phase signals: sa1 and sa2=sa1. However, at the same time, the 100 ohm resistance of the Wilkinson power splitter adds to the previous signals two noise signals that are 180º out of phase: sn1 and sn2=-sn1. This way, the signals at the input of the two channels of the radiometer are sa1 + sn1 and sa1 – sn1, that once properly amplified, down-converted, sampled, and combined with the ones from other receivers to digitally form 0-7803-9510-7/06/$20.00 (c) 2006 IEEE 1 the beam(s), are finally cross-correlated to lead -, which is proportional to Ta–Tph, being Ta the antenna temperature, and Tph the physical temperature of the Wilkinson power splitter resistor. That is, the system output is the same as the one of the Dicke radiometer, but the input signal is not chopped, so that it can be used to track the GPSreflected signal. This article presents and describes the PAU instrument concept at system level and the predicted performance in terms of angular resolution and radiometric sensitivity. Three companion articles [22-24] describe in detail the receiver topology, the implementation of the digital beam-former and the calibration strategy, and the implementation of a real-time Doppler-Delay Map (DDM) generator. of RJ-45 grade 5 cables and applying the band-pass sampling technique, the in-phase and quadrature signals are obtained. Fig. 1. Schematic of the PAU-RAD units Fig. 2. Receiver block diagram II. INSTRUMENT DESCRIPTION A. Receiver Front-End One of the technological goals of the PAU instrument is to demonstrate the feasibility of combining in a single hardware two types of receivers: the radiometer (PAU-RAD) which, for stability reasons cannot be a total power radiometer, and the GNSS-Reflectometer (PAU-GNSSR). In order to be able to use the same receivers for both the radiometer and the GPSreflectometer, a new radiometer topology has been devised (Fig. 1). Instead of connecting the antenna output to the radiometer receiver, it is connected to a Wilkinson power splitter that divides the signal in two in-phase signals: sa1 and sa2=sa1. However, at the same time, the 100 ohm resistance of the Wilkinson power splitter adds to the previous signals two noise signals that are 180º out of phase: sn1 and sn2=-sn1. In addition, for calibration purposes two calibration signals are foreseen: uncorrelated noise, generated by a matched load at each input channel, to compensate for instrumental biases, and 2 levels of correlated noise from a common noise source to compensate for phase and amplitude mismatches among receivers (Fig. 2). In order to produce a beam with good properties for a microwave radiometer: a 20º 3 dB beamwidth with a main beam efficiency (MBE) larger than 95 % must be synthesized for pointing angles up to ±20º from the boresight direction (45º from nadir: range of incidence angles from 25º to 65º). These three requirements are satisfied with a 4 x 4 array of elements spaced 0.63 wavelengths (Fig. 3) at the GPS L1 frequency (1.57542 GHz), with “triangular” weighting function [1 2 2 1]. The inter-element spacing determines the maximun size of the receiver to which the antenna is connected: 7 cm x 11 cm x 3 cm. The first three circuits assembled and tested are shown in Fig. 4: the elementary element antennas are microstrip patch antennas, which are very convenient at this frequency (top), and the RF (left) and IF (center) boards, which are interconnected by means of 4 semi-flexible cables, folded, and integrated in a 7 cm x 11 cm x 3 cm box (right). The isolation between channels is better than 40 dB, suitable for the measurement of the four Stokes parameters, and the input power is -110 dBm. The output IF signals are transmitted at a central frequency of 4.3 MHz through a pair Fig. 3. Array of 4x4-elements pointing to the ocean. The array boresight will be 45º off nadir. Fig. 4. Elementary element antennas microstrip patch antennas (top) and RF (left) and IF (center) boards that are interconnected by means of 4 semi-flexible cables, folded, and integrated in a 7 x 11 cm box (right). B. Analog-to-Digital Converters The 8 bit samplers are formed by an array of 16 ADC cards whose inputs are the outputs of each receiving unit: 4 channels per receiving unit. The throughput of each ADC card is: 4 channels x 8 bits x 5.745 MHz = 183.84 Mbps, and the total input rate at the FPGAs is thus 16 times higher, approximately 3 Gbps. This hughe amount of data is entered into the FPGA by using nearly all available I/O pins, but time multiplexing by 4 is needed, that is 4 signals sampled at 5.745 MHz are input at 22.98 MHz. C. Digital Beam-Forming (DBF) Processor The design has been split in two parts and fit in two Altera FPGAs (Field Programable Gate Array) Stratix EP1S10F780C6E (without asymmetrical load) to implement PAU-RAD using the VHDL-93 digital description language. The first one is called Control Unit and manages the system. The control unit has an external communications module using the TCP/IP protocol in order to provide a standard interface to remote control the system. Furthermore, it calculates the corrections to be applied to the digital I/Q signals for calibration. These functions have been developed using the Altera VHDL Nios microprocessor, which makes the design of the control module easier and more versatile, since it is programmed using Ansi-C. The second part of the system is called ALU (Arithmetic-Logic Unit), where the algorithm is implemented. Signals are digitally processed in this block to obtain the four Stokes parameters with maximum accuracy. The PAU-RAD design includes other VHDL blocks as well that are not directly related to the radiometer sub-system, but 0-7803-9510-7/06/$20.00 (c) 2006 IEEE 2 to the PAU-GNSS Reflectometer. Simulated results are shown in Fig. 5. a) b) Fig. 5. Simulated synthetic beam at: a) the boresight direction, and b) 20º off boresight, including the radiation pattern of the elementary element. D. Real-time Delay-Doppler Map (DDM) Generator To generate a DDM from a stream of GPS data downconverted and sampled it is necessary to generate simultaneously several different frequency carriers along with a number of shifted versions of the corresponding PRN code. With the aim of obtaining real-time DDM waveforms a DDM generator has been implemented in a FPGA using VHDL digital logic description language. To simplify the system, a commercial GPS receiver is added to the system to provide some ‘a priori’ information on which satellites are visible and which are their respective delay and Doppler offsets for the direct signals. This information is fed into the DDM generator along with the stream of phase-quadrature (I/Q) GPS samples. Every clock cycle a new pair of I/Q samples are processed: for every Doppler and delay coordinate a carrier and PRN code samples are obtained, then they are combined with the I/Q samples, and finally they are sent to their corresponding accumulator block. At the end of the integration period, which is related to the PRN chip rate and to the coherence time of the sea surface, the result is dumped, thus obtaining a complex DDM point. This reflectometer computes all these DDM points at the same time, so after the selected integration time a whole DDM is available. Since the sampling frequency of the system is higher than the chip rate of the PRN codes, each chip is represented by more than one sample. To generate a PRN code with a noninteger chip offset first this offset is split into integer and noninteger smaller that 1 chip values. The integer shift is obtained by applying a delay mask to the status registers of the two G1 and G2 linear feedback shift registers (LFSR) used to generate the GPS codes. The non-integer shift is generated by adjusting the initial value of a counter that outputs the driving pulse for the G1 and G2 LFSR every time that it exceeds a certain top value. The count step value and the above mentioned top value are related to both the sampling frequency and the chip rate. The implementation of the digital oscillator is based on a linear correspondence between the phase and the amplitude in the first wave quarter, plus a level correction. Every clock cycle a counter is increased in an amount related to the desired frequency to synthesize. This counter value is equivalent to the wave phase value, and its maximum value corresponds to a phase of 2π. Thus, the most significant (MSB) bit indicates whether the sine amplitude is positive or negative, and the second MSB indicates whether the wave quarter increases or decreases. The remaining bits are used to determine the absolute value of the amplitude. The DDM generator is the main peripheral of an embedded reflectometer system. Such a system has an UART RS-232 interface to receive the ‘a priori’ information from the commercial GPS receiver, a microprocessor, a memory to host the program code, the DDM generator attached to a common bus, and a buffer peripheral for the incoming sampled GPS data, that arrives in a continuous fashion to the reflectometer from the PAU-RAD system [24]. The program run by the microprocessor receives the data packets with information from the direct signals and extracts from them the necessary input parameters for the DDM generator. To do so it selects some of the available satellites taking into account a given position mask (elevation and azimuth) and their respective power levels. It also tells the PAU-RAD, which is in charge for the beamforming (DBF), which is the corresponding beam that best points to each satellite. Then the code sets the DDM parameters into the DDM generator, commands the raw data buffer peripheral to start dumping the I/Q data and the DDM generator to start processing and accumulating the result. Once the new DDM is ready an interrupt indicates the processor to read the obtained points and to store or display them. At the moment this reflectometer is in its implementation stage. Fig. 6. Receiving unit connected through two RJ45 cables to an ADC card, which is connected to the master DBFN FPGA. Fig. 7. Master and slave Altera FPGAs interconnected by a MICTOR cable and sharing the same clock signal. The design has been verified and the digital description has been largely simplified by developing a whole software emulator (simulating noise samples) using Matlab. Simulation results have been satisfactorily compared with the theoretical predictions and preliminary results using only one analog receiver and one antenna (Figs. 6 and 7) are shown in Figs. 8 and 9 for the DDM generation. The four Stokes parameters are also calculated and verified, and an accuracy analysis of the gain and the phase estimation is performed. Fig. 8. a) Testing the system with 1 receiver at the Barcelona Olympic Village Fig. 9. DDM signal after processing the data collected in an experiment (Fig. 8). 0-7803-9510-7/06/$20.00 (c) 2006 IEEE 3 E. Infrared Radiometer Finally the IR radiometer is a thermal infrared commercial radiometer covering the spectral range 8-14 µm that will be connected to the serial port of the control PC. MODTRAN simulations are being carried out to determine the feasibility of using model results to properly correct for the atmospheric effects (clouds and rain) that will affect the measurements. Depending on this analysis a second radiometer may be needed to measure the downwelling radiation. [7] [8] [9] [10] III. CONCLUSIONS This work has summarized the objectives, topology and main characteristics of the PAU instrument which is currently under development. It is expected to integrate and test the first PAU instrument will be during the second half of 2006. During 2007 a second PAU instrument using aperture synthesis techniques will be integrated and tested, and they will be deployed in the field during the end of 2007 or the first half of 2008. The PAU instrument concept and design have been patented. [11] ACKNOWLEDGMENT This work, conducted as part of the award “Passive AdVanced Unit (PAU): A Hybrid L-band Radiometer, GNSSReflectometer and IR-Radiometer for Passive Remote Sensing of the Ocean” made under the European Heads of Research Councils and European Science Foundation EURYI (European Young Investigator) Awards scheme in 2004, was supported by funds from the Participating Organisations of EURYI and the EC Sixth Framework Programme. [14] REFERENCES [18] [1] [2] [3] [4] [5] [6] Swift, C. T., and R. E. McIntosh, Considerations for microwave remote sensing of ocean surface salinity, IEEE Trans. Geosci. Remote Sens., 21(4), 480–491, 1983. Camps, A., J. Font, M. Vall-llossera, C. Gabarró, I. Corbella, N. Duffo, F. Torres, S. Blanch, A. Aguasca, R. Villarino, L. Enrique, J. Miranda, J. Arenas, A. Julià, J. Etcheto, V. Caselles, A. Weill, J. Boutin, S. Contardo, R. Niclós, R. Rivas, S.C.Reising, P. Wursteisen, M. Berger, and M. Martín-Neira, The WISE 2000 and 2001 field experiments in support of the SMOS Mission: Sea Surface L-Band Brightness Temperature Observations And Their Application to Multi-Angular Salinity Retrieval, IEEE Transactions on Geoscience and Remote Sensing, Vol. 42 (4), pp. 804-823, April 2004 Sabia, R., A. Camps, N. Reul, M. Vall-llossera, J. Miranda, “WP 1400: Towards Best SSS Level 2 products,” final report, December 2005, Synergetic Aspects and Auxiliary Data Concepts for Sea Surface Salinity Measurements from Space, ESTEC ITT 1-4505/03/NL/Cb. ftp://ftp.estec.esa.nl/ftp/pub/vr/VRO/GS_Study_18176_FR.pdf Camps, A., I. Corbella, M. Vall-llossera, N. Duffo, F. Torres, R. Villarino, L. Enrique, J. Miranda, F. Julbé, J. Font, A. Julià, C. Gabarró, J. Etchetto, J. Boutin, A. Weill, V. Caselles, E. Rubio, P. Wursteisen, M. Berger, and M. Martín-Neira, “L-band sea surface emissivity: Preliminary results of the WISE-2000 campaign and its application to salinity retrieval in the SMOS mission, RADIO SCIENCE, VOL. 38, NO. 4, 8071, doi:10.1029/2002RS002629, 2003. Martín-Neira, M., A Passive Reflectometry and Interferometry System (PARIS): Application to Ocean Altimetry. ESA Journal 1993, vol. 17. pp 331-355. Martin-Neira, M. P. Colmenarejo, G. Ruffini, C. Serra, Ocean Altimetry using the Carrier Phase of GNSS Reflected Signals, Proceedings of Ocean Winds 2000, Plouzane, France, CERSAT News, Issue 11, November 2000 [12] [13] [15] [16] [17] [19] [20] [21] [22] [23] [24] Belmonte Rivas M. ; Martin-Neira M.; Coherent GPS Reflections From the Sea Surface, Coherent GPS Reflections From the Sea Surface; IEEE Geoscience and Remote Sensing Letters: (in press), DOI 10.1109/LGRS.2005.855617 Soulat, F., M. Caparrini, O. Germain, P. Lopez-Dekker, M. Taani, G. Ruffini, Sea state monitoring using coastal GNSS-R Geophys. Res. Lett., Vol. 31, No. 21 Ruffini, G., Caparrini, M., Chapron, B., Soulat, F., Germain, O., Ruffini, L., Oceanpal: an instrument for remote sensing of the ocean and other water surfaces using GNSS reflections, Proceeding of EuroGOOS, 3-6 December 2002, Athens, ISBN 0-444-51550-X (Elsevier). Garrison, J.L., Katzberg, J.L., Effects of sea roughness on bistati- cally scattered range coded signals from the Global Positioning System, GRL, vol 25, n. 13, 1998. Komjathy, A., V. Zavorotny, P. Axelrad, G.H. Born, and J.L. Garrison, GPS signal scattering from sea surface: Wind speed retrieval using experimental data and theoretical model. Rem. Sens. Env., 73:162–174, 2000. Garrison, J.L.; Komjathy, A.; Zavorotny, V.U.; Katzberg, S.J.; Wind speed measurement using forward scattered GPS signals; IEEE Transactions on Geoscience and Remote Sensing, Vol. 40, Issue 1, Jan. 2002 pp. 50 – 65, DOI 10.1109/36.981349 Cardellach, E., G. Ruffini , D. Pino , A. Rius , A. Komjathy , and J.L. Garrison 2003 Mediterranean Balloon Experiment: ocean wind speed sensing from the stratosphere, using GPS reflections Remote Sensing of Environment Volume 88, Issue 3 , 15 December 2003, Pages 351-362 Lowe, S., C. Zuffada, Y. Chao, P. Kroger, J.L LaBreque, and L.E. Young, 5-cm precision aircraft ocean altimetry using GPS reflections,GRL, (29):4359–4362, 2002. Rius, A., J.M. Aparicio, E. Cardellach, M. Martín-Neira, and B. Chapron, “Sea surface state measured using GPS reflected signals”, Geophys. Res. Lett.,29(23), 2122, 2002. Ruffini, G. F. Soulat, M. Caparrini, O. Germain , M. Martin-Neira, The Eddy Experiment: accurate GNSS-R ocean altimetry from low altitude aircraft, Geophys. Res. Lett., Vol. 31, No. 21, 2004 Germain, O., G. Ruffini, F. Soulat, M. Caparrini, B. Chapron and P. Silvestrin, The Eddy Experiment II: GNSS-R speculometry for directional sea-roughness retrieval from low aircraft Geophys. Res. Lett., Vol. 31, No. 21, 2004 Lowe, Stephen T.; LaBrecque, John L.; Zuffada, Cinzia; Romans, Larry J.; Young, Larry E.; Hajj, George A., First spaceborne observation of an Earth-reflected GPS signal, Radio Science, Vol. 37, No. 1, 07 February 2002. Gleason, S.; Hodgart, S.; Yiping Sun; Gommenginger, C.; Mackin, S.; Adjrad, M.; Unwin, M.; Detection and Processing of bistatically reflected GPS signals from low Earth orbit for the purpose of ocean remote sensing; IEEE Transactions on Geoscience and Remote Sensing, Vol. 43, N. 6, June 2005, pp. 1229 – 1241, DOI 10.1109/TGRS.2005.845643 Camps, A., M. Caparrini, R. Sabia, G. Ruffini, “Sea Surface Salinity Retrieval from Space: Potential Synergetic Use of GNSS-R Signals to Improve the Sea State Correction and Application to the SMOS Mission,” Microrad 2006, San Juan, Puerto Rico, February 28th – March 3rd, 2006. http://www.esf.org/generic/2079/AwardeeCamps.pdf Ramos-Perez, I., X. Bosch-Lluis, A. Camps, J. F. Marchan-Hernandez, R. Prehn, “Design of a compact dual-polarization receiver for pseudocorrelation radiometers at L-band,” proceedings of the IGARSS 2006, Denver, Colorado, July 31-August 4, 2006. Bosch-Lluis, X., A. Camps, J. F. Marchan-Hernandez, I. Ramos-Perez, R. Prehn, “FPGA-based implementation of a polarimetric radiometer with digital beamforming,” proceedings of the IGARSS 2006, Denver, Colorado, July 31-August 4, 2006. Marchan-Hernandez, J. F. , I. Ramos-Perez, X. Bosch-Lluis, A. Camps, R. Prehn, “FPGA-based implementation of DDM-generator for GPSreflectometry,” proceedings of the IGARSS 2006, Denver, Colorado, July 31-August 4, 2006. 0-7803-9510-7/06/$20.00 (c) 2006 IEEE 4