mmWave Wireless

In this project we are developing a new millimeter-wave MIMO transceiver architecture — continuous aperture phased (CAP) MIMO — for emerging applications at cm-wave and mm-wave (10GHz-300GHz) frequencies. CAP-MIMO combines the directivity gain of traditional antennas, the beam-steering capability of phased arrays, and the spatial multiplexing gain of MIMO systems to realize multi-Gbps capacity potential of mm-wave technology as well as the unprecedented operational functionality of electronic multi-beam steering and data multiplexing (MBDM). It is based on the concept of beamspace MIMO communication — multiplexing data onto orthogonal spatial beams — for optimal operation. CAP-MIMO leverages beamspace MIMO theory through two key architectural elements:

  1. a lens-based front-end antenna for analog beamforming
  2. a multi-beam selection network

that enable joint hardware-software optimization for taming the transceiver complexity of the high-dimensional MIMO systems at such high frequencies. A key characteristic of cm-wave and mm-wave propagation channels that facilitates complexity reduction is their expected sparsity in beamspace. The hybrid analog-digital CAP-MIMO architecture enables optimal and efficient access to the low-dimensional communication subspace induced by the sparse MIMO channel.

capmimo arch

Fig 1: CAP-MIMO Transceiver: The front-end lens antenna is excited by the beamspace feed antenna array spanning the coverage area. p precoded data streams are mapped to O(p) beams via the mmW beam selector.   The selected beams span the sparse communication subspace of the high (n) dimensional channel.

Project Scope and Prototyping: The project involves the development of basic theory, communication and signal processing algorithms, computational modeling and simulation, and prototype development. We have recently achieved initial proof-of-concept demonstration of CAP-MIMO with a 10GHz prototype that can support four spatial channels in a point-to-point (P2P) link for achieving 1Gbps data rate. We have recently developed a second generation (Gen 2) prototype at 28GHz and achieved a major milestone of real-time demonstration of MBDM capability in a point-to-multipoint (P2MP) setting. We are currently building on the success to develop a CAP-MIMO testbed network at 28 GHz.

rf hardwareP2P LinkP2MP link
Fig 2: Gen 1 10GHz CAP-MIMO Protototype: RF + digital hardware; point-to-point link and initial measurements; point-to-multipoint link.

The prototype development effort is enabling three key outcomes:

  1. Proof-of-concept demonstration of CAP-MIMO technology
  2. Unprecedented multi-beam MIMO channel measurements
  3. A state-of-the-art wideband MIMO transceiver for experimentation and development of larger scale testbeds

The CAP-MIMO transceiver promises significant advantages over current and emerging competing technologies, including:

  • Significant improvements in data capacity, power efficiency, and functionality
  • Optimum performance with the lowest transceiver complexity — number of transmit/receive (T/R) modules or transceiver chains and DSP complexity.

CAP-MIMO also offers a broad application footprint:

  • P2P and P2MP network operation
  • Operation in Line-of-sight (LoS) and/or multipath (MP) propagation environments
  • Electronic MBDM capability for channel discovery and alignment, and for creating dynamic mobile P2MP links.

5G Applications: The CAP-MIMO architecture is applicable to three key Gigabit applications of cm-wave and mm-wave technology:

  1. backhaul mesh networks
  2. P2MP networks for last-mile wireless access
  3. mobile access networks.

It is expected to deliver performance and operational capabilities that are critical to emerging cm-wave and mm-wave wireless broadband applications in future Gigabit wireless networks (5G and beyond), but beyond the reach of current systems.

Technology Transfer: We are very interested in collaborating with industrial partners for further development and commercialization of CAP-MIMO technology.

Industrial partners: Support from industrial partners is also gratefully acknowledged: Analog Devices, Altera, 4DSP, MI-WAVE, Microwave Dynamics, Reactel.

Funding Sponsors: Funding from the Wisconsin Alumni Research Foundation (WARF) and the National Science Foundation (NSF) is gratefully acknowledged.