Ka-Band Link Optimization with Rate Adaptation Jun Sun*, Jay Gaot, Shervin Shambayatit, and Eytan Modiano* *Laboratory of Information and Decision Systems Massachusetts Institute of Technology Cambridge, MA 02139 Email: {junsun, modiano}@mit.edu tJet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Drive, Pasadena, CA 91109 Email: {Jay.L.Gao, Shervin. Shambayati}(jpl.nasa.gov Abstract-On-going development of Ka-band capability for in antenna gains at higher frequencies, an improvement in the Deep Space Networks (DSN) will radically increase the the signal strength by a factor of four can be expected bandwidth available to support advanced mission concepts in the transition from X-band to Ka-band [1. Moreover, envisioned for future robotic as well as human exploration of Mars and beyond. While Ka-band links can operate at much at I a-band deep sace communications is alMlHocated 500 higher data rate than X-band, they are much more susceptible MHz of bandwidth compare to the 50 MHz of bandwidth to fluctuating weather conditions and manifest a significant allocated to the X-band [L]; leading to even greater increase trade-off between throughput and availability. If the operating in throughput when using Ka-band. point is fixed, the maximum average throughput for deep space Although the throughput gain in Ka-band can be signif- Ka-band link is achieved at about 80 percent availability, i.e., weather-related outages will occur about 20 percent of ucant, Ka-band data transmissio is susceptible to weather the time. Low availability increases the complexity of space fluctuations. The deep space systems use receivers with mission operation, while higher availability would require very low noise temperatures of approximately 25 K. This additional link margins that lowers the overall throughput. makes them extremely sensitive to atmospheric effects [3], To improve this fundamental throughput-availability tradeoff, [2]. Weather events, such as rain and clouds, increase the data rate adaptation based on real-time observation of the channel condition is necessary. moisture in the atmosphere; and consequently the noise In this paper, we model the ka-band channel using a temperature. Markov process to capture the impact of the temporal The proposed data transmission scheme in Ka-band, correlation in weather conditions. We then develop a rate which is going to be demonstrated on the Mars Recon- adaptation algorithm to optimize the data rate based on naissance Orbiter (MRO), uses constant rate transmission. real time feedback on the measured channel conditions. Our algorithm achieves both higher throughput and link Speci.fically, duorsing a paotcular communicatioi n sessiomn, availability as compared to the constant rate scheme presently based on the statistics of the atmospheric noise temperature in use. and pass geometry, a link configuration is decided in advance to achieve a constant transmission rate. When the actual transmission takes place, the targeted, data rate can be TABLE OF CONTENTS achieved with acceptable reliability if the atmospheric noise I INTRODUCTION ................ . 1 temperature is below a certamn threshold, noise temperature. 2 LINK MARGIN ANALYSIS .......... . 2 If the noise temperature exceeds the threshold, significant 3 THROUGHPUT AND LINK AVAILABILITY. . 3 data corruption and loss of synchronization will occur, 4 ADAPTIVE RATE TRANSMISSION SCHEME. 4 making the link unusable. 5 COMPARISON: THROUGHPUT AND SER- An immediate consequence of using a pre-determined, VICE AVAILABILITY ............. . 6 fixed rate transmission scheme is the disruption of data 6 CONCLUSION .................. 6 continuity due to bad weather. Weather phenomena such REFERENCES ..................... 7 as microbursts can significantly increase the system noise BIOGRAPHIES . . . . . . . . . . . . . . . . . . . . . 7 temperature thus causing outages for a short period of time [1]. If a timely decision on whether to retransmit the lost 1. INTRODUCTION data cannot be made, such disruptions will greatly compli- Ka-band trnmiso is viwda . rm en cate mission operationas. fIn order to ensure the reliabi:lity for meeting the increasing demands for high data rate ofacmuiainln,adiinlpwrmri ilb added to combat the weather flLuctuation, at the expense of services of space explLoration missions. Due to the increase lin thoghu. 'O-7803-9546-8/06/$20OO0(e) 2006 IEEE In this paper we propose to u.se anl ad.aptive rate trans- 2IEEEAC paper ii 84, Version iL Updated October7 5 2005 missioln sc:heme to combat the weather effect associated