capmimo.html

Continuous Aperture Phased (CAP) MIMO

Research Group Page

Group Members:

Faculty:

Akbar Sayeed
Nader Behdad

Students:

John Brady
Gi Hong Song
Hongzhou Zhang (undergraduate, helping Spring 2012)

Past Students:

Mudar Al-Joumayly
Pushkar Kulkarni

List of relevant main papers (MP):

A. Sayeed and N. Behdad, Continuous Aperture Phased MIMO: Basic Theory and Applications, Allerton 2010. Extended version.
A. Sayeed and N. Behdad, A Novel Hybrid Analog-Digital Architecture for Optimum AgileWireless Communication Using Discrete Lens Arrays, NSF EAGER proposal.
A. Babakhani, D. Rutledge, A. Hajimiri, Near-Field Direct Antenna Modulation, IEEE Microwave Magazine, Feb 2009.
P. Daly and J. Bernhard, Directional Modulation Technique for Phased Arrays, IEEE Trans. Antenna and Propagation, Sep 2009. Earlier conference paper.
P. Daly, E. Daly and J. Bernhard, Demonstration of Directional Modulation for Phased Arrays, IEEE Trans. Antennas and Propagation, May 2010.
IMAT white paper from Skycross.
D. Lawrence, Low Probability of Intercept Antenna Array Beamforming, IEEE Antennas and Propagation, Sep 2010.
H. Kogelnik and T. Li, Laser Beams and Resonators, Proceedings of the IEEE, October 1966. (Also in some cases Applied Optics, October 1966).
H. Friis, A Note on A Simple Transmission Formula, Proc. IRE, May 1946.
H. Friis, An Introduction to Radio and Radio Antennas, IEEE Spectrum, April 1971.

Other papers (OP):

S. A. Schelkunoff, A Mathematical Theory of Linear Arrays, Bell Sys Tech Journal, 1942-43
J. B. Keller, Geometrical Theory of Diffraction, J. Optical Society of America, Feb 1962
M. Kales and R. Brown, Design Considerations for Two-Dimensional Symmetric Bootlace Lenses, IEEE Antennas and Propagation, July 1965.
W. Kock, Metallic Delay Lenses, Bell Sys Tech Journal, 1948.
W. Kock, Path-length Microwave Lenses, Proc. IRE, August 1949.
D. McGrath, Planar Three-Dimensional Constrained Lenses, IEEE Transactions on Antennas and Propagation, Jan 1986.
R. Harrington and J. Mautz, Theory of Characteristic Modes for Conducting Bodies, IEEE Trans. Antennas and Prop., Sep 1971.
M. Cabedo-Fabres, E. Antonino-Daviu, A. Valero-Nogueira and M. Ferrando Bataller, The Theory of Characteristic Modes Revisited: A Contribution to the Design of Antennas for Modern Applications, IEEE Antennas and Prop. Magazine, Oct 2007.
E. Antonino-Daviu, M. Cabedo-Fabres, M. Gallo, M. Ferrando-Bataller, and M. Bozzetti, Design of a Multimode MIMO Antenna Using Characteristic Modes, European Conference on Antennas and Propagation, EuCAP 2009.
W. Rotman and R. F. Turner, Wide-angle microwave lens for line source applications, IEEE Trans. Antennas and Propagation, Nov. 1963, pp 623-632.
J. Dong, A.I. Zahloul, and R. Rotman, Phase-error performance of multi-focal and non-focal two-dimensional Rotman lens designs, Microwaves, Antennas & Propagation, IET, Dec 2012, Vol. 4 Issue 12

Related presentations, white papers and articles (RP):

R. Daniels, J. Murdoch, T. Rappaport, R. Heath, 60 GHz Wireless: Upclose and Personal, IEEE Microwave Magazine, Dec 2010.
T. Rappaport, M. Corsi, R. Heath, Broadband Wireless Personal Area Networks - 60 GHz and Beyond, IEEE Globecomm, Business Forum, Dec 2010.
WiGig white paper - Defining the Future of Multi-gigabit wireless communications, Wireless Gigabit Alliance, July 2010.
J. Wells, Faster than Fiber - the future of multi-Gb/s wireless communication, IEEE Microwave Magazine, May 2009.
A. Niknejad, Siliconization at 60GHz, IEEE Microwave Magazine, Feb 2010.
Vijay Venkateswaran and Alle-Jan van der Veen, Analog beamforming in MIMO communications with phase shift networks and online channel estimation, IEEE Trans. Signal Processing, August 2010.
T. Marzetta, Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas, IEEE Transactions on Wireless Communications, Nov 2011.
mm-wave beamforming in MIMO-OFDM for WPAN, IEEE Wireless comm, 2011.

Communication Modes (Optics) papers (CM):

D.A.B. Miller, Communicating with waves between volumes: evaluating orthogonal spatial channels and limits on coupling strengths,Applied Optics, April 2000.
A. Burvall, P. Martinsson, A. Friberg, Communication modes in large aperture approximation, optics letter, march 2007.
P. Martinsson, H. Lajunen, A. Friberg, Communication modes with partially coherent fields, JOSA, October 2007.
Shapiro et. al. Ultimate Capacity of Optilcal links.

Ray Tracing Papers

1. R. Brem and T. F. Eibert, Multi-Radiation Center Transmitter Models for Ray Tracing, IEEE Transactions on Antennas and Propagation, Vol. 60, no.7, July 2012 pp. 3382-3388.

Group meetings:

10-11-2010. Akbar presented his allerton talk on capmimo. ppt slides.
10-18-2010. Nader presented a talk on biomimetic array architectures.
10-25-2010. discussion of DM and NF-DAM papers in relation to cap-mimo. Papers: MP1, MP3, MP4

DM versus NF-DAM. both DM and NF-DAM are aimed at using the spatial dimension to create detectable modulation (QAM) in a desired direction while obfuscating the signal constellation in an undesired direction. Both can be used for arrays of antennas and rely on the differences in the signal path lengths from the array to a desired single-antenna receiver versus an undesired receiver at a differnet direction. A key difference is that NF-DAM uses switches within an single array element (of size lambda/2) to change its far-field radiation patterns and signal constellation as a function of location, whereas DM uses different phase patterns across the array to achieve the same objective. In this regard, due to the limitation of a single antenna element in producing changes in path lenghts/phases, DM seems more effective. Furthermore, NF-DAM is aimed at shorter links whereas DM can be used for longer links. Both are for LoS comm.
connection with CAP-MIMO. Certainly the idea of DM can be realized in CAP-MIMO in the beamspace. There may even be opportunities for doing a better job at secure communication by working directly in beamspace. Further, we can extend the idea to multipath propagation, wideband transmission (via frequency-selective surfaces) and full MIMO operation (higher rates) by sending multiple data streams towards the receiver.
Also discussed the spatial mode solution for a signal dipole as a function of feed location, presented by Nader. Its has a nice geometrical interpretation. Interesting aspects would be to extend the patterhs for larger length antennas and to explore further degrees of freedom in exciting different spatial modes.
Aside: Packers beat Vikings in a nail-biting (28-24) finish the night before. Favre almost worked his magic with a last minute touchdown to give Minnesota the lead, but the touchdown was ruled in incomplete reception!

11-1-2010. Discussed the CAP-MIMO WARF project. Three main tasks: 1) Detailed DLA phase profile design and full-wave simulation, 2) DLA fabrication and the system building (supports for focal feed elements at TX and RX), 3) Phase-coherent system measurements and data processing. Need to finish the tasks by June 2011. So, tentative deadlines for the three tasks. Task 1: Jan-Feb 2011. Task 2: Feb-March 2011. Task 3: June 2011. Some intermediate steps to keep in mind:

Purchasing computers and software for full-wave simulation.
Stitching together the full-wave simulation of the lens (which gives the field amplitude/phase profile as a function of feed excitation) with Ray tracing modeling of the propagation from TX to RX.
Preparing the students involved with both basic theory and modeling and full wave simulation and measurement process. Mudar will give a presentation on full-wave simulation and measurements next week.

11-8-2010. Akbar and Nader were out. (Akbar at Cowboys-Packers game; 45-7 blowout! Nader at a conference). Mudar discussed FSS-based lens design and associated software for simulation.
11-15-2010. 1) discussed the lens phase profile design. Akbar will lead the effort with help from John and Pushkar. Will share the matlab code. As a starting point we can formulate lens design as a n-dimensional optimization problem with n variables and n constraints (more details later). 2) Nader provided information on the full-wave simulation software choices. Will need both software and a dedicated machine. need to make that decision soon. will only simulate the lens design and then stitch with the overall system via ray tracing (matlab-based approach). will revisit the basic theory (periodic structures) behind FSS-design approach (Nader and Mudar) for a better understanding of the design and the resulting errors/iterations needed. 3) Akbar led a discussion on the directional modulation business. seems that having parasitic feed elements on the focal surface may be the best union of caltech and uiuc approaches. key conjecture: phase/amplitude shifting with parasitic feed elements should not impact the I/Q constellation much in the desired direction but significant differences in other (unwated directions). The other (parasitic) feeds for unwanted directions could be used for additional noise in unwanted directions. would like to include this aspect in APS paper (January deadline). Gi Hong and Pushkar will work with Akbar on this.
11-22-2010. Akbar discussed the connection between CAP-MIMO theory and the notion of spatial modes in Laser Beams and Resonators (MP 8 paper - thanks to Prof. Ron Vernon). This is also connected to the theory of characteristic modes that John had discovered (OP 7).
11-29-2010. Discussed the CAP-MIMO WARF project timeline through June 2011. On track for detailed prototype design by the end of semester (Milestone M1a). Allocated a rough timeline for the other two milestones: M1b (prototype hardware development for 1D arrays, including full-wave simulation of the DLA design - mid March), and M1c: system measurements and data processing (June 2011). Tentative assignment of responsibilities for graduate students: FSS-based DLA design and fullwave simulations -John Brady. Prototype hardware development, including the nature of feed antennas and focal arc assembly for the feeds - Pushkar. System measurents: all (John, Pushkar, Gi Hong). Data processing (Gi Hong). Mudar along with Akbar and Nader will be involved in all aspects. Need to finalize the purchase of software by end of semester. Further purchasing of hardware will be in Spring.
12-6-2010. Agenda: discussion of DLA phase profile design, including the choice of feed antennas, and the nature of full-wave simulations.
1-24-2011. Present: Akbar, John, Gi Hong, Puskhkar, and Mudar. Update on DLA beampattern measurement results from Mudar. Some issues regarding theoretical versus measurement results. John will replicate the theoretical results and compare with Mudar's.
1-31-2011. Present: All. Discussed three issues: 1) New GEMS software: whether its needed and when to purchase it. Consensus was to contact them for a trial offer for simulating a full lens and estimate its phase profile via the software and compare it to the designed phase profile. This would give us an idea of whether the software can handle the job (and subsequent ones) to make it a worthwhile purchase. will start a discussion with them and perhaps have the full-wave simulation of lens after Mudar-Nader have published their first paper on the design. John (and Pushkar) will lead the DLA full-wave simulation with guidance from Mudar and Nader. 2) Finalization of the DLA profile design and TX/RX DLA fabrication. John will lead this effort. Generate theoretical beampatterns to compare with Mudar's measurements and theoretical plots. Perhaps do additional measurement to isolate the source of possible errors (measurement vs inherent DLA design) before finalizing on the phase profile design prior to ordering fabrication. Timeline: to finalize the DLA phase profile design by end of Feb. 3) Discussed the system measurement set up and design for the prototype CAP-MIMO link. Starting point is the current set up Mudar has for the 27cm lens. Possible ideas: Increase the size of the insulator surrounding the lens aperture. Perhaps put an isolation box around the DLA to avoid leakage. Pushkar will lead this effort and we will go to lab to see Mudar's set in an upcoming meeting. 4) Gi Hong will lead the system measurement and data processing part. Create a detailed step-by-step protocol for system mesurements and processing.
2-7-2011. No meeting in celebration of Packers' 31-25 superbowl victory over the Steelers. Go Packers! CAP-MIMO tasks.
2-14-2011. Got an update from Brady and Gi Hong on the lens design and comparison with Mudar's measurement data. (Brady: ppt file here). went over the different tasks in the WARF project and deadlines.
2-21-2011. No meeting.
2-28-2001. updates and discussion of upcoming WARF interview on Mar 2.
3-7-2011. No meeting. Akbar met with Gi Hong and John for update on lens design.
3-14-2011. No meeting. Spring break.
3-21-2011. Got update from Gi Hong on the lens design capacity assessment. Brady making progress on physical design specification to send out for fabrication. Some issues: how much errors do we expect in the design process? (both design software as well as manufacturing tolerance). Tentative choice of feeds: dipoles antennas. Key issue here is to have an adequate SNR. Pushkar will do some calculations build a second dipole antenna if necessary to do some measurements.
3-28-2011. no meeting.
4-4-2011. John gave update on lens design; to do: accounting for bonding layer effects, and angle of incidence effects. Gi Hong gave an update: showed capacity results for random phase errors (upto 10 deg) in the optimum phase profile. While there was significant (about 20%) power loss compared to the optimum, an interesting result was that there were still 2 dominant eigenvalues (as in the no-error case), thereby making it feasible to demonstrate the capmimo gains over dish. Pushkar gave an update on feed design.
4-11-2011. Gi Hong updated that he got better results (less power loss) when he incorporated the smaller cell-size compared to lambda/2; needs to incorporate amplitude change using John's results. John updated that bonding layer stuff is taking more time. should have more results next week. but he has a coarse amplitude/phase profile that he could give to Gi Hong. currently focusing on 5th order filter design. Pushkar updated that he simulated the dipole feeds and found asymmetric beam patterns - needs to carefully read the papers and revisit it. Nader noted that we need to finalize on the cable orders soon since they take upto 6 weeks. Updated deadlines: lens design (John) asap, no later than end of April. Pushkar: end of april for feed design. Full measurement apparatus by end of May (design/dimensions to be finalized by Mid May). Akbar updated that he had talked to Booske about space.
4-18-2011. John updated on lens design - some problems with frequency response at the edges - will talk with Mudar - deadline May 15. Did a quick calculations due to angle of incidence at the edge of the lens- upto 10-15 degrees phase errors. Gi Hong updated on channel calculations with phase errors and smaller cell-size on phase profile - results look promising - even with 6 deg phase errors two dominant eigenvalues are present. Pushkar updated on feed design - a balanced dipole. The simulations indicated a tilt in teh beampatterns - fairly symmetric though. Nader suggested simulating hte components - the dipole and hte driving circuit separately as sanity checks to see if the circuit is behaving as expected. Akbar suggested going back to the beampattern analysis of a dipole to see what causes a tilt. That could be useful because tilting precisely corresponds to different beam directions, so this may be a clue to generating different beam directions. perhaps something to do with the phase difference between the two feeds to the dipole.
4-25-11. pushkar created two dipoles to act as feeds and receivers on the focal surfaces. did S21 measurements on the two at a distance of about 10feet (link length). Got a -60dB measurement for S21 between the two dipoles. The noise floor of the VNA is -90dB so with the expected gain for the lenses we should be okay. Finalized on hte order for two 20' feet cables with a 12dB loss. Each about $1000.
5-2-11 (Bin Laden killed!). Everything on track. discussed lens design sanity check with John with regard to transder function. Puskkar investigated the beam patterns of dipole - seems like cannot produce too much of a beam shift even with larger apertures due to the feed structure (this to be revisited).
5-17-11. met for a short meeting. Everthinng on track. Gi Hong will investigate optimization of feed locations (later results showed that the original positions are pretty good).
5-24-11. John got a quote on the lens manufacture and also on materials. makes sense. discussed metamaterials, power and information with Nader. The combined gain of the feed and the lens is a tricky calculations - will have to revisit.
5-31-11. Lens material arrived. John accounting for angle of incidence in design.
6-7-11. Pushkar ordered the absorbers. costly!
6-21-11. gi hong gave a presentation on off broadside channels. interesting results.
6-28-11. John finished lens design! Should receive the lenses next week. Pushkar made significant progress on prototype design and building. Should be done by the end of the week. Akbar and Nader will be at APS symp. in Spokane next week.
7-5-11. No meeting. Akbar and Nader at APS. Beautiful country north of Spokane.
7-12-11. Akbar, John and Gi Hong met. Reviewed the prototype construction. Only one lens arrived. Will test the first lens and then order the fab of second lens by next Monday. The lens looks beautiful.
7-19-11. Meeting cancelled to escape the heat and the lack of A/C in EH.
7-26-11. Gi Hong, John, Pushkar, and Akbar met to discuss prototype status and initial measurements. The prototype is nearly finished. The second lens is here and final steps on finishing the building. First set of measurements, without metal sheet or absorbers, expected by the end of the week or early next week. The measurement results would demonstrate the final milestone of phase 1 of WARF project.
8-2-11. The first set of measurements without metal sheet showed asymmetry but still demonstrated two eigenmodes. So, the team put metal sheets and redid the measurements for a 1D 2x2 system. The results are close to what we expect theoretically and demonsrate the cap-mimo technology (Phase 1 of WARF project).
8-9-11. Additional measurements were done on a 4x4 (2D) system and the results are again comparable to theory and demonstrate the existence of 4 dominant 2D modes. Theoretically, we expect upto 9 modes. Pushkar is optimizing feed design (flat dipole, patch, or waveguide). The next steps for John and Gi Hong are P2MP operation (and lens design). Also mutual coupling. The linear aspects of mutual coupling (at the feeds) can be readily incorporated in the CAP-MIMO theory. Akbar did some preliminary calculations and saw that mutual coupling can actually provide further increase in performance (contrary to conventional wisdom). Gi Hong will likely take the lead on the mutual coupling aspects.
8-17-11. Met with Pushkar and Nader to discuss feed designs. Two commercially viable ones: square and circular waveguide feeds (both give about 9db gain). Pushkar is checking on physical specs and cost. In terms of homemade design, between fat dipole and patch, the latter looks more promising. Fat dipole issues: sensitivity to feed location and the construction of ground plane. Pushkar will further investigate the patch design. Some issues: size of patch, size of ground plane (no larger that .9lambdax.9lamda), gap between patch and ground plane, and the dielectric constant of the gap
8-18-11. **** Met with Brady and Gi Hong. Discussed the plan for Gi Hong: performance assessment of CAP-MIMO for three TX/RX structures with or without channel estimation errors. The three structures: 1) Benchmark system corresponding to n dimensional SVD of H with p modes. 2) Beamspace channel with ideal DFT (H_o) and with lens (H_L). Ideal DFT (U_o), Lens: U_L. based on p_eff dimensional SVD. Relationship between p_eff and p to be determined; e.g based on capturing the same fraction (.98?) of total channel power. 3) Beamspace/Lens channel with independent transmisson on p_eff feed locations (beams) and MMSE receivers. Do capacity comparison (spectral efficiency vs energy efficiency (Eb/No)) for all three systems and also P_e analysis for uncoded (e.g., QPSK) transmission. Then Repeat above analysis with channel estimation errors. Channel estimation errors could be due to errors in channel estimation. But also consider other forms of errors (e.g., orientation of antennas). Brady will create an outline of the lens paper. Also gave him pointers for a few things: 1) Does U_L have a kronecker structure? If so, may reduce computation. 2) Next set of measurements. 3) Viewing the different mode propagation to TX aperture, RX aperture, RX focal surface. This would also allow us to estimation fraction of power that couples to receiver.
8-23-11. John updated on computational framework for DLA and system modeling. Showed contour plots for different beams at the RX aperture. Pushkar updated on waveguide choices.
8-30-11. Pushkar updated on costs for circular, rectangular (standard), and square (custom) waveguides. Akbar discussed plans for Phase II to be presented at the next WARF meeting on Sep 21.
9-21-11. Presented Phase I results to the WARF catalyst group, including plan for Phase II and a tour of the prototype lab. Very useful and informative meeting.
10-11-11. Updates from Brady on new lens measurements. Gi Hong updated on MMSE and SVD transceivers and impact of channel estimation errors. The MMSE results look very promising. The error analysis needs to be updated.
10-18-11. No meeting. Students busy with midterms.
10-25-11. Gi Hong updated results on MMSE and SVD transceivers. 2D-MMSE looks very promising. Still some problem with SVD. John updated on teh cost of various waveguides and transitions for hte next prototype.
10-31-11. Phase II proposal finally finished and submitted to WARF.
11-1-11. Discussed initial plans for Phase II and got updated from Gi Hong and Brady. Gi Hong's results looking very promisng. fixed the problem with SVD-based transcievers.
11-8-11. Discussed the first telecon with Samsung (Akbar and Brady could make it).
11-15-11. Discussed Phase II plans and also Samsung presentation.
11-23-11. Telecon with Samsung: Akbar, Nader, Jerry Pi, Farooq Khan, and Chirala.
11-29-11. Discussed the NSF proposal.
12-6-11. Further discussion of the planning for the NSF proposal.
2-6-12. NSF proposal submitted.
2-20-12. Charlie Zhang (Samsung) visited and gave an interesting presentation.
2-21-12. Discussed the various forms of technology for phased arrays. And power capabilities. Also, ongoing discussions with Samsung. Also discussed the recent visit by Charlie (JZ)
2-28-12. Discussed power calculations for the prototype to account for any losses and for reconciling theory with measurements. In particular, the importance of power not coupled between teh TX and RX and the use of ratios of coupled to total power. Also, first principle continuous approaches.
3-16-12. Brady discussed his paper plan, now that he is done with the qualifiers. In particular doing some modifications for the prototype in view of upcoming measurements.
3-27-12. Discussed Gi Hong's results on MF receiver performance. Brady's paper preparation and measurement schedule. Basically, detailed P2P measurements (NB), and BW measurements only for broadside. For P2MP, do broadside measruements for 0-60 deg, with 4-5deg increments. Then do spot measurments 2x2 measurements and BW measurements for 30, 60. Add as necessary. Akbar discussed Friis formula and some modifications that may be needed in our context.