System Parameter Design of Multimodal Small Satellite SARs Operating in Scan Mode and Transmit Power Optimization for Marine Scenes

Multimodal small satellite synthetic aperture radar (SAR) is a new radar system under development that integrates an SAR, altimeter, scatterometer, and spectrometer. When applied to marine scenes, this system can be used to measure both marine targets and marine dynamic environments with high precision. This study addresses system parameter design for multimodal small satellite SAR operating in scan mode, and this design provides system parameters such as antenna size, signal bandwidth, pulse repetition frequency, scanning wavenumber and wave position for system simulation in scan mode. A method for optimizing the transmit power when illuminating a marine scene based on the wind speed and wind direction above the sea surface is studied. The goal is to fully use the characteristics of strong sea surface microwave scattering under a suitable wind speed and wind direction to reduce the required transmit power, thus improving the available data sampling time per orbit of a multimodal radar when working in SAR mode. Various simulation experiments were conducted, and the system parameter design results are given under scan mode. Furthermore, imaging simulation results of ocean scenes are also given under conventional and decreased power after optimization. The results show that good ocean scene imaging results are obtained when the designed system parameters are used for system simulation. In addition, the simulation results also verify that when the sea surface wind speed is relatively high and the wind direction is suitable, an acceptable ocean scene imaging result can still be obtained by using reduced transmit power.


I. INTRODUCTION
Multimodal small satellite synthetic aperture radar (SAR) is a new radar system under development [1] that simultaneously takes full advantage of the characteristics of two-dimensional phased array antennas such as flexible beam scanning and multiple beam generation [2] and the flexibility of recon-The associate editor coordinating the review of this manuscript and approving it for publication was Gerardo Di Martino . figurable electronic devices introduced by software reconfigurable technology [3]- [5]. This system integrates remote sensor functions including an SAR, altimeter, scatterometer and spectrometer by a software-reconfigurable radar receiver and by steering the antenna side looking with a medium incidence angle, nadir looking, rotating 360 degrees horizontally with a medium incidence angle and switching between two different incidence angles, and rotating 360 degrees horizontally with a small incidence angle, respectively.
More specifically, the system can work in different modes including the SAR, altimeter, scatterometer, and spectrometer under a time-sharing mechanism by receiving remote control instructions from ground stations. When applied to marine scenes, SAR mode has obvious advantages in detecting marine targets [6]; altimeter mode measures sea surface height with high precision; scatterometer mode is good at retrieving sea surface wind field information with high precision [7], [8]; and spectrometer mode has unique advantages for retrieving wave spectrum information and extracting wave parameters [9]. Because each operating mode has its own advantages, this new radar system can both detect marine targets and measure marine dynamic environments with high precision; thus, it has good development prospects. In addition, a laser communication terminal can be installed on the multimodal small satellite SAR to enable multiple satellites to communicate, thus achieving synchronous multistatic observations to improve the remote sensing performance such as achieving high-resolution wideswath microwave images and obtaining high-precision information of marine targets and marine dynamic environments simultaneously, or achieving network observations through satellite constellations to improve the temporal resolution. Furthermore, to deal with the limited processing resources of a small satellite, the real-time processing capability of a small satellite can be effectively improved through intersatellite parallel computing with the support of high-speed inter-satellite communication. Due to the rapid development of electronic equipment technology, the size and cost of phased array antennas and laser communication devices are continuously decreasing [10], [11], making it possible to manufacture the multimodal small satellite SARs described in this paper.
As a new radar system under development, multimodal small satellite SAR needs to be simulated by computers, which is an important means for carrying out early research. This paper focuses primarily on imaging simulations and transmission power optimization of ocean scenes when multimodal small satellite SAR is working in scan mode. In SAR system simulations, operating parameters such as antenna size, signal bandwidth, pulse repetition frequency (PRF), scanning wavenumber and wave position should be determined first [12]. Due to the size and load limitations of small satellites, the parameters of traditional spaceborne SAR cannot be directly used in determining the operating parameters of multimodal small satellite SAR; therefore, specific designs are required. After the design results for the operating parameters of the multimodal small satellite SAR system are obtained, ocean scene imaging simulations can be further carried out. Due to transmit power limitations and power consumption, the available data sampling time of spaceborne SAR is usually only a few minutes per orbit [13], and the available data sampling time of multimodal small satellite SAR is even shorter when working in SAR mode. A multimodal small satellite SAR working in scatterometer mode can obtain the wind direction and wind speed of the irradiated sea area with high precision, and can transmit that information to other small satellites that are synchronously observing the scene. When the acquired wind speed is high and the wind direction is suitable, the microwave scattering intensity of the irradiated sea area will be strong. Here, a suitable wind direction means directions apart from those that deviate from the radar line of sight by approximately ±90 degrees since the microwave scattering is relatively weak when the wind direction is perpendicular to the radar line of sight [14]. In this case, the emission power of the small satellite SAR can be decreased, thus increasing the available data-sampling time per orbit of SAR mode.
There are numerous studies on the design of system parameters for SAR working in strip and spotlight modes. In [15] the design process of system parameters in strip mode was introduced, and [16] discusses the design of key system parameters for a high-resolution wide geosynchronous orbit (GEO)/low earth orbit (LEO) bistatic SAR system. The authors of [17] introduced software for parameter design and simulation of a spaceborne strip SAR system. The software can complete the parameter design and perform echo and imaging simulations for the strip SAR system. In [18], a design method for system parameters in spotlight mode was introduced that analyzes the factors influencing system design from several aspects, such as large distance migration, PRF selection, and antenna pattern design. However, there are relatively few studies on system parameter design for SAR working in scan mode. The details and a comprehensive discussion on the performance parameters and design constraints of scan SAR were provided in [19], including the scanning period, resolution, number of wave positions, antenna area, swath width, average transmit power and echo data rate, but that study did not reveal how to determine the position and time parameters of different subswaths, nor did it consider transmit power optimization.
In this paper, a method for designing system parameters, imaging simulations of ocean scenes and a method for optimizing the transmit power based on sea surface wind speed and direction are studied for multimodal small satellite SAR working in scan mode. Part II introduces the details for small satellite SAR system design in scan mode. Part III introduces the method for optimizing the transmit power based on the wind speed and direction above the sea surface when the multimodal small satellite SAR illuminates marine scenes. Part IV introduces the simulation process and the imaging results for small satellite SAR over ocean scenes operating in scan mode, including the adopted sea surface simulation method, echo simulation and imaging processing method, system parameter design results using the proposed method, and ocean scene simulation results under both conventional power and decreased power after using the proposed method for transmit power optimization. Part V concludes this paper. VOLUME 8, 2020

II. PARAMETER DESIGN METHOD FOR SMALL SATELLITE SAR SYSTEMS OPERATING IN SCAN MODE
When designing the system parameters for small satellite SAR in scan mode, system operating parameters such as antenna size, signal bandwidth, PRF, scanning wavenumber, positions and the time relation of each subswath need to be determined [20], [21]. To determine the above operating parameters, other parameters such as the range resolution, azimuth resolution, swath width, satellite orbit height, satellite speed, carrier frequency, view angle range, noise figure of the receiver, and pulse width are usually known or given according to the user requirements.
First, the number of scanning wave positions is determined. The number of scanning wave positions is related to the scanning time of each subswath (i.e., the full aperture imaging accumulation time corresponding to the beam width) and the beam dwell time (i.e., the continuous imaging time of a burst within a certain subswath).
Let k represent the index of a given subswath; then, the scanning time and beam dwell time corresponding to the k-th subswath are respectively, where λ is the wavelength, R k is the distance from the antenna to the center of the given subswath, v g is the satellite's velocity relative to the ground, l a is the azimuth antenna length, v s is the orbit speed of the SAR, ρ a is the azimuth resolution, and N L is the effective multiple view number. Dividing the scanning time corresponding to the nearest subswath by the beam dwell time corresponding to the farthest subswath is used to calculate the number of wave positions, N B : where R N and R F represent the nearest and farthest slant ranges, respectively. In equation (3), for a given ρ a and a given N L , the product N B l a will be a constant, and a compromise should be made between l a and N B during the design process.
The second step is to determine the position of each subswath. The relation between the view angle, the incidence angle and the geocentric angle of a given subswath center is where R e is the radius of the earth, H is the height of the satellite, γ is the view angle, α is the incidence angle, and ϕ is the geocentric angle. From equations (4) and (5), we can see that we need to know only one of the angles to calculate the other two. For the first subswath, the view angle γ 1 is known, and the geocentric angle ϕ 1 can be calculated. Let ϕ a1 and ϕ b1 denote the geocentric angles corresponding to the nearest distance R n1 and the farthest distance R f 1 of the first subswath, respectively. These values can be calculated as follows: where A 1 B 1 is the width of the first subswath, whose value can be estimated by the user requirements of the imaging swath and N B . After obtaining ϕ a1 and ϕ b1 , the nearest distance R n1 and the farthest distance R f 1 corresponding to the first subswath can be calculated as follows: The beam width θ 1 of the first subswath is calculated by The widths of the first and second subswaths should generally overlap by one tenth. Thus, where A 2 B 1 is the intersection width between the first and second subswaths.
Assuming that the initial value of the view angle of the second subswath isγ 2 , the width of the overlap region A 2 B 1 can be calculated by where R x is the width in the slant range direction, and its relationship with whereα 2 is the incident angle corresponding toγ 2 . Obviously, the estimated view angle of the second subswath to meet the accuracy requirement can be obtained by using the iterative method. Thus, the position of the second subswath can be further determined. The positions of the other subswaths can be calculated similarly.
In step 3, the signal bandwidth B r can be determined according to the user requirement of the range resolution ρ gr by where C is the speed of light, and α a1 is the incident angle corresponding to ϕ a1 (i.e., the incident angle at the nearest end).
The 4th step involves calculating the minimum antenna area: where R f ,max is the farthest distance and θ f ,max is the farthest incident angle.
Step 5 determines the scanning time T FK , the dwell time T DK and the regression time T RK of each subswath. Here, the regression time T RK refers to the time interval between two bursts of a certain subswath, and T FK and T DK are calculated by (1) and (2), respectively, while T RK is calculated by In the 6-th step, the PRF value and the number of PRF samples in each subswath are determined. To reduce azimuth ambiguity, the PRF of the k-th subswath should satisfy where PRF k represents the PRF value of the k-th subswath, B p is the Doppler bandwidth and B p = v s /ρ a . To reduce range ambiguity, the PRF of the k-th subswath should meet the following requirements: where m represents the number of pulses between the scene echo and the transmitted pulse and R nk and R fk represent the nearest and farthest distances corresponding to the k-th subswath, respectively. After obtaining the PRF range, the timing diagram can be drawn. The X axis of the timing diagram is the PRF range and the Y axis is the range of the incident angles. Among these, to avoid interference from the emission pulse, the following requirements must be met: where the function Frac() represents obtaining the fractional part of the variable, the function Int() represents obtaining the integer part of the variable, T r is the pulse width, PRI k is the pulse repetition interval corresponding to the k-th subswath, and τ g is the receiver protection time. To avoid interference from the nadir echo, it is necessary to satisfy: where l represents the number of pulses between the nadir echo and the transmitted pulse. After drawing the timing diagram, the PRF value suitable for each subswath can be selected. The number of azimuth samples in each burst of each subswath can be obtained by dividing the beam dwell time by the value of PRI.
Step 7 verifies the range ambiguity to signal ratio RASR k and the azimuth ambiguity to signal ratio AASR k of each subswath. The formula for RASR k is where i is the serial number of the range cell in the main beam, j is the serial number of the ambiguity region, R ki is the slant range corresponding to the i-th range cell in the main beam, and R kij is the slant range corresponding to the j-th ambiguity region of the i-th range cell in the main beam, and S ki and S kai are the signal power and the ambiguity power of the i-th range cell of the k-th subswath, respectively. G kij and θ kij are the antenna gain and the incident angle corresponding to the j-th ambiguity region of the i-th range cell of the k-th subswath, respectively, n 1 and n 2 are the lower limit serial number of the near-end ambiguity region and the upper limit serial number of the far-end ambiguity region respectively, and φ kij is the angle between the beam center direction and the j-th ambiguity region of the i-th range cell of the k-th subswath. Here, l r is the antenna size in the range direction, and N k is the number of range cells in the k-th subswath. The formula for AASR k is where f is the Doppler frequency.
In the 8th step, the required average transmit power P av is given by where P t is the peak power and PRI min is the minimum value of all PRI for each subswath.
Step 9 verifies whether the normalized equivalent noise figure NEσ 0 meets the user's requirements. According to VOLUME 8, 2020 the radar equation, the signal-to-noise ratio (SNR) of SARs operated in scan mode can be written as where σ 0 is the normalized radar cross section (RCS), R max is the farthest range between the radar and the scene, B k is the Boltzmann constant, T 0 = 290 K , F n is the receiver noise figure, L s is the system loss, G is the antenna beam center gain, and ρ gr is the ground range resolution. Therefore, the corresponding formula to verify NEσ 0 is Finally, whether the data rate f D meets the requirements is verified by where Q is the quantization bit of each sample, ρ r is the range resolution, and k s is the oversampling coefficient. Fig. 1 shows a flowchart of our proposed parameter design method for a small satellite SAR system in scan mode, which summarizes the previous design steps. The proposed system parameter design approach for a small satellite SAR operating in scan mode is applicable for both land scenes and marine scenes.

III. TRANSMIT POWER OPTIMIZATION METHOD BASED ON SEA SURFACE WIND SPEED AND DIRECTION INFORMATION
First, the influence of the wind speed and the wind direction above the sea surface on the backscattering coefficient is discussed. Next, the proposed transmission power optimization method for marine scenes is introduced.

A. INFLUENCE OF WIND SPEED AND WIND DIRECTION ABOVE SEA SURFACE ON BACKSCATTER COEFFICIENT
The sea surface can be seen as a kind of composite rough surface composed of slightly rough surface superimposed on a large wave structure. The large wave structure causes the slightly rough surface to tilt, forming the double-scale model. The small-scale capillary wave component and the large-scale long wave component of the double-scale model correspond to the two types of sea surface backscattering mechanisms under high-frequency electromagnetic waves: Bragg scattering at oblique incidence and specular reflection at near-zero incidence, respectively. For SARs and scatterometers, the incident angle is relatively large; therefore, Bragg scattering is the main scattering component [22], [23]. The Bragg scattering component is expressed as follows: where w k is the wavenumber, θ is the incidence angle, g ij (θ) is the first-order scattering coefficient related to the polarization mode, and S PM is the two-dimensional Pierson-Moskowitz (PM) spectrum [24], which is a typical representative wave spectrum. The PM spectrum is a type of gravity spectrum in which long and short gravity waves play a major role when the wavenumbers are different [25]. It is obtained by fitting a large number of ocean observation data after spectral analysis. The PM spectrum model in the main wave direction is: where α = 0.0081, β = 0.74, g is the gravitational acceleration, and U 19.5 is the wind speed 19.5 m above the sea surface. When U h1 , which is the wind speed at the height of h 1 above the sea surface, is known, the wind speed at the height of 19.5 m above the sea surface can be obtained by the following conversion: The two-dimensional PM spectrum can be expressed as: where ψ is the angle between the current direction and the reference direction and ψ m is the angle between the wind direction and the reference direction. For radar systems operating in HH polarization mode, there are: For radars operating in VV polarization mode, there are: where ε r is the relative dielectric constant of seawater. According to equations (28) and (31), it can be found that the backscattering coefficient of the sea surface has a strong relationship with the wind speed and direction above the sea surface. As the wind speed increases, the backscattering coefficient increases according to the exponential function term in equation (31). The backscattering coefficient is largest when the observation direction is consistent with the wind direction and smallest when the observation direction is perpendicular to the wind direction according to the cosine function term in equation (31).
The mature CMOD5 geophysical model function [26] can also obtain the relationship between the backscattering coefficient and the wind speed and direction. Fig. 2 shows the relationship between the simulated backscatter coefficient and the wind speed, while Fig. 3 shows the relationship between the simulated backscatter coefficient and the wind direction. Figs. 2 and 3 are plotted using the CMOD5 model and verify the characteristics of the relationship between the backscattering coefficient and wind speed and direction summarized above.

B. OPTIMAL CALCULATION OF TRANSMIT POWER
Multimodal small satellite SARs make full use of the flexible scanning and the ability to generate multibeam coverage of phased array antennas and software-reconfigurable technology. By receiving remote control commands from ground stations, remote sensors can be operated in different modes, including the SAR, altimeter, scatterometer and spectrometer.   When the multimodal small satellite SAR is working in scatterometer mode, highly precise information regarding the wind direction and wind speed of the irradiated sea area can be obtained. The microwave scattering intensity on the sea surface is relatively strong when the wind speed in the irradiated sea area is relatively high and the wind direction is appropriate. When other small satellites synchronously observing the sea scene receive information that the wind speed is relatively high and the wind direction is appropriate, they can optimize their own emission power, thus improving the available data sampling time. Fig. 4 shows an example of the above application. Small satellite A is in SAR mode and small satellite B is in scatterometer mode. Both satellites are observing the same sea area at the same time. Small satellite B transmits the retrieved information, including the wind speed, the wind direction, the average backscatter coefficient of sea VOLUME 8, 2020 surface, etc., to small satellite A. If the wind speed above the irradiated sea area is relatively high and the wind direction is suitable, small satellite A can optimize its own emission power based on the average backscatter coefficient information of the sea surface transmitted by satellite B. The working modes of different satellites at a given time are arranged by receiving remote control instructions from ground stations when satellites pass through the coverage areas of the ground stations. A task scheduling technique needs to be developed to arrange appropriate working modes for different satellites at a given time according to the requirements of the tasks.
According to equation (26), the following can be derived: where P t is the normal signal peak power and PRI max is the maximum PRI among all subswaths. Therefore, the following formula for optimizing the transmission power can be derived: where P t is the optimized signal peak power,σ 0 is the average backscatter coefficient of the sea surface under calm sea conditions, andσ 0 is the average backscatter coefficient of the sea surface measured by a scatterometer synchronously observing the sea surface. In SAR systems, the movement of a platform inevitably deviates from the nominal motion. These uncertainties of the platform's motion degrade the image quality during imaging [27], [28]. However, the platform's motion uncertainties have little influence on equation (35) since this equation is related to the transmitted signal power rather than the signal phase. Obviously, when the wind speed above the irradiated sea area is high and the wind direction is appropriate,σ 0 is notably larger thanσ 0 , and the required peak signal power can be significantly reduced. Even under the circumstance that a small satellite operating in SAR mode moves toward the wind direction, i.e., the case that the direction of the radar line of sight is perpendicular to the wind direction since SARs normally operate in side view, it is possible to optimize the transmit power as long as the wind speed is high, resulting in thatσ 0 is greater thanσ 0 .
In summary, the system adaptively operates with the appropriate transmit power according to the scattering intensity of the illuminated sea. A small satellite working in SAR mode receives information such as the wind speed, the wind direction, and the average backscatter coefficient of the sea surface transmitted by another small satellite that synchronously observes the same area and operates in the scatterometer mode. Then, the optimized transmit power is calculated using equation (35) according to the information that the small satellite receives, and the small satellite transmits at the calculated power.

IV. SIMULATION AND RESULTS OF OCEAN SCENE IMAGING FOR SMALL SATELLITE SARS IN SCAN MODE
Several simulation experiments were conducted. The main goals included two aspects: (1) verifying the rationality of the proposed system parameter design method; and (2) verifying the correctness of the proposed transmission power optimization method. First, the sea surface simulation method, point target echo simulation and imaging processing method are introduced. Second, the design results of system parameters of the small satellite SAR system in scan mode are provided. Third, the imaging simulation results of ocean scenes under conventional transmission power are given to verify the rationality of the parameter design results. Finally, the simulation results of ocean scene imaging with optimized transmit power are reported to verify the correctness of the proposed transmit power optimization method.

A. SEA SURFACE SIMULATION METHOD ADOPTED
To simulate ocean scene echo for small satellite SAR systems, a sea surface simulation is the first requirement. At present, the main method for sea surface simulation is to generate the wave spectrum and then simulate the sea surface according to the generated wave spectrum. The wave spectrum is the power density spectrum of the sea surface, which is the Fourier transform of the correlation function of sea surface height fluctuation. Common wave spectra include PM spectrum, JONSWAP spectrum and Apel spectrum. In this paper, the PM spectrum, which is mathematically the most concise, is used to simulate the sea surface.
After obtaining the two-dimensional PM spectrum given in equation (31), the two-dimensional sea surface simulation can be completed using the following two-dimensional inverse Fourier transform: where x m and y n represent the range and the azimuth coordinates of the two-dimensional sea surface, respectively, L x and L y represent the length of the two-dimensional sea surface in the range and azimuth directions respectively, M and N represent the number of sampling points in the range and azimuth directions respectively, and S PM K m k , K n k represents the PM spectrum in the two-dimensional frequency domain.

B. POINT TARGET ECHO SIMULATION AND IMAGING PROCESSING METHOD ADOPTED
At present, the main methods for simulating sea echo include the time domain and frequency domain algorithms [29]. The frequency domain algorithm has higher computational efficiency, while the time domain algorithm has higher accuracy. In this study, the time domain algorithm [30] was chosen for simulation. According to the principle of scan SARs, the echo signal of a point target can be expressed as follows: where τ is the range time, t is the total time, σ is the backscattering coefficient of the point target, rect() represents the rectangle function, T D is the beam dwell time, R(t) is the instantaneous slant distance, and k r is the modulation frequency of the transmitted signal. After simulating the height change of the two-dimensional sea surface, the backscattering coefficient of each cell can be calculated by using the wind speed, wind direction, incidence angle, etc. After calculating the instantaneous slant range from each cell on the sea surface to the radar, the echo data of the sea surface can be obtained by superposing the echo signals of each cell according to the scattering coefficient matrix and the impulse response function of the system. After the echo matrix is obtained, the SPECAN imaging algorithm [31], which is suitable for scan mode, is adopted for imaging to obtain the sea surface SAR image. Table 1 gives the technical indicator requirements for small satellite SAR in scan mode, and Table 2 gives the known technical parameters. Note that the parameters listed in Tables 1 and 2 refer to the parameters to be adopted in an ongoing development plan. According to these parameters, using the proposed system parameter design method, the system parameter output results shown in Table 3 are obtained. It is noted that the nearly 80 W average power listed in Table 3 is acceptable for minisatellites (100-500 kg) [32]. Table 4 lists the calculated results of the spatial position and beam pointing angle of each subswath. Table 5 shows the residence time, the regression time and the scanning time for each subswath.

C. SYSTEM PARAMETER DESIGN RESULTS OF SCAN MODE
Combinations of different wave positions and PRF were selected for verification; then whether the combination of different wave positions and PRF meets the requirements of indicators such as ambiguity to signal ratio and data rate   can be judged. The verification results are shown in Table 6, which shows that the calculated results all meet the requirements of the index. Table 7 shows the calculated PRF values and the number of PRF samples for each subswath. Fig. 5 plots the corresponding timing diagram for selecting the proper PRF.

D. SIMULATION RESULTS OF OCEAN SCENES AT CONVENTIONAL TRANSMITTING POWER
Using the two-dimensional PM spectrum model, the twodimensional sea surface was simulated, and the results are shown in Fig. 6. The wind direction angles corresponding to Fig. 6 (a)    and roughness become larger. Comparing Fig. 6 (a) with Fig. 6 (c), it can be found that when the wind direction angle is 90 • , the sea surface fluctuation is slightly lower than when the wind direction angle is 45 • . The above phenomena coincide with the theoretical law.
After the two-dimensional simulated sea surface was obtained, a sea surface simulation imaging experiment was carried out according to the parameters of the satellite platform and the radar. Based on the echo simulation and imaging processing methods described in Section IV.B, the echo signal of the whole sea surface was obtained first; then, the sea surface image in scan mode was obtained. Fig. 7 shows the sea surface imaging results. The wind direction angles corresponding to Fig. 7 (a), (b), (c), and (d) are 45 • , 45 • ,  As shown in Fig. 7, good sea surface imaging results can be obtained under different wind speeds and wind directions when using the designed radar parameters, and clear sea wave stripes can be seen. In addition, simulation experiments of ship imaging both with and without sea clutter were also conducted. We artificially set an area with a very low RCS in the upper-left corner of the simulated marine scene to facilitate the calculation of the SNR values. Noise was added according to the bandwidth and the noise figure of the receiver listed in Tables 2 and 3,  respectively. Fig. 8 shows a scatter diagram of a simulated ship. Fig. 9 shows the result of the ship imaging simulation without sea clutter, and Fig. 10 plots the results of the ship imaging simulation in a sea scene. The wind direction angles corresponding to Fig. 10  From Figs. 9 and 10, it can be seen that the experiment using the designed radar parameters achieved good imaging results under different wind speeds and directions, thus verifying the rationality of the designed radar parameters.

E. SIMULATION RESULTS OF OCEAN SCENES WITH OPTIMIZED TRANSMIT POWER
The average backscattering coefficient of the sea surface in a calm sea state was set to −23 dB. According to the parameters set in the simulation and equation (24), the calculated average emission power is 79.4 W. When the wind direction angle is 45 • and the wind speed is 10 m/s, the average backscattering coefficient of the sea surface is −14 dB, and the average transmitting power of the radar can be reduced to 10 W. When the wind direction angle is 45 • and the wind speed is 15 m/s, the average backscattering coefficient of the sea surface is −12 dB, and the average transmitting power of the radar can be reduced to 6.3 W. When the wind direction angle is 90 • and the wind speed is 10 m/s, the average backscattering coefficient of the sea surface is −17 dB, and  sea surface is −18 dB, and the average transmitting power of the radar can be reduced to 25 W. Fig. 11 shows a composite imaging result of the ship and the sea surface after the  transmission power is optimized. The wind direction angles corresponding to Fig. 11  Comparing Fig. 10 (a)-(d) with Fig. 11 (a)-(d), respectively shows that although the image quality decreases slightly after optimizing the transmission power, the wave stripes and the ship outline in the image are still relatively clear. Table 8 presents the SNR and the signal-to-clutter ratio (SCR) related to Figs. 10 and 11. It can be seen that the SCR values are not related to the transmitted signal power and that the SNR values after the power optimization are still high enough. These phenomena conform to the theoretical expectations because the signal power and the clutter power are in direct proportion to the transmit power and because the normalized RCS of a ship's scatterer is generally far greater than the normalized equivalent noise figure of SAR systems. That is, the SNR value after the transmit power optimization is still high with the relatively small reduction of the transmit power. Table 9 gives the image entropy and the image contrast of Figs. 10 and 11. It can be seen that the values of the image entropy and the image contrast after power optimization are roughly equal to the values before power optimization. In other words, the image quality has not decreased significantly after power optimization.

V. CONCLUSION
Multimodal small satellite SAR is a new type of radar system under development that can work in SAR, altimeter, scatterometer, and spectrometer modes in a time-shared manner. When applied to marine scenes, this approach can achieve measurements of both marine targets and marine dynamic environments with high precision. The ability to simulate multimodal small satellite SAR using computer systems is an important foundation for related research into this new system. In this study, an imaging simulation of ocean scenes and the optimization of emission power for multimodal small satellite SAR in scan imaging mode were investigated-primarily the system parameter design method for multimodal small satellite SAR in scan mode, the imaging simulation of ocean scenes and the optimization method of emission power based on wind speed and wind direction information of the sea surface. Various simulation experiments were carried out. The simulation results showed that when the designed system parameters were used for an imaging simulation of scan mode, good imaging results of ocean scenes were obtained. In addition, the simulation results also verify that when the sea surface wind speed is relatively high and the wind direction is suitable, a relatively good imaging result of the ocean scene can be obtained even when using reduced emission power.