How does the virtual aperture work in co-located MIMO?

Each transmitter radiates a unique waveform (orthogonal to all others). Each receiver applies matched filters for all transmit waveforms, producing N_tx matched filter outputs per receiver. The k-th transmitter and l-th receiver pair has an effective phase center at the midpoint of their physical locations. With proper placement (e.g., transmitters spaced at N_rx times d and receivers spaced at d, where d is the half-wavelength element spacing), the N_tx times N_rx virtual phase centers form a uniform linear array with half-wavelength spacing. For example, 4 transmitters and 8 receivers create a virtual array of 32 elements, with angular resolution equal to a 32-element conventional array but using only 12 physical antennas. The aperture savings factor is N_tx times N_rx divided by (N_tx plus N_rx), which is maximized when N_tx equals N_rx.

What are the advantages of co-located MIMO over phased arrays?

Co-located MIMO offers four key advantages. First, improved angular resolution: the virtual aperture of N_tx times N_rx elements provides finer angular resolution than a phased array with N_tx plus N_rx elements. Second, waveform diversity: different transmit waveforms illuminate the target from the same direction but with different spatial signatures, providing independent observations of the target radar cross section. This reduces target scintillation losses (Swerling fluctuation) by up to 10 dB through diversity combining. Third, enhanced parameter identifiability: the additional degrees of freedom enable joint estimation of target angle, range, and Doppler with better accuracy than a phased array of equal physical aperture. Fourth, lower sidelobe levels: the virtual array can be designed with tapered weighting to achieve sidelobes 10 to 20 dB lower than the physical array. The main disadvantage is reduced coherent processing gain: each transmitter radiates only 1/N_tx of the total power, reducing per-target SNR by 10*log(N_tx) dB compared to a phased array that focuses all power in one direction.

Radar & Defense

Co-Located MIMO

Q: What waveform orthogonality methods are used?

Three primary methods achieve orthogonality between co-located MIMO transmitters. Time-division multiplexing (TDM) transmits from each antenna sequentially, simple to implement but reduces the effective PRF by N_tx, limiting Doppler coverage. This is the dominant approach in automotive MIMO radar (TI AWR series, NXP S32R). Frequency-division multiplexing (FDM) assigns each transmitter a different frequency offset (delta_f = 1/T_chirp for FMCW radar), maintaining full Doppler coverage but requiring wider total bandwidth. Code-division multiplexing (CDM) applies orthogonal phase codes (Hadamard, Gold, or optimized binary codes) across pulses, providing simultaneous transmission from all antennas with full Doppler coverage. CDM is used in advanced military MIMO radars and some 5G NR positioning systems. Hybrid approaches combining TDM and CDM (e.g., alternating slow-time codes across TDM cycles) offer practical compromises between complexity and performance.

/koh-loh-kay-ted my-moh/

Co-located MIMO radar uses multiple transmit and receive antennas at one site, each transmitter radiating an orthogonal waveform. Matched filtering creates a virtual aperture of N_tx×N_rx elements from only N_tx+N_rx physical antennas, providing angular resolution equivalent to a conventional array of N_tx×N_rx elements. Also provides waveform diversity for reduced target scintillation (up to 10 dB), enhanced parameter estimation, and lower sidelobe levels.

Category: Radar & Defense

Virtual elements: N_tx × N_rx

Physical antennas: N_tx + N_rx

Understanding Co-Located MIMO Radar

Traditional phased array radar uses a single waveform transmitted coherently from all elements, forming a beam that scans across space. All elements contribute their power to one direction at a time, maximizing SNR for a single target. Co-located MIMO takes a fundamentally different approach: each transmit element radiates a different waveform simultaneously, illuminating the entire field of view at once. The price is reduced per-direction power (each transmitter contributes only 1/N_tx of total power), but the reward is a virtual array with N_tx×N_rx elements that can simultaneously estimate target angles across the entire illuminated space.

The virtual array concept works because each transmit-receive pair has a unique spatial phase relationship. When transmitter k at position x_k and receiver l at position y_l observe a far-field target, the round-trip phase depends on x_k+y_l. By choosing transmitter and receiver positions carefully (e.g., minimum redundancy arrays), the set of all x_k+y_l sums forms a virtual array with more elements and wider aperture than either physical array alone. For the simplest case of a uniform linear array with N_tx transmitters at N_rxd spacing and N_rx receivers at d spacing (d = λ/2), the virtual array is a filled ULA of N_tx×N_rx elements at d spacing. This is exactly the principle used in automotive MIMO radar: a 3Tx/4Rx chip creates 12 virtual channels for angle estimation.

Co-Located MIMO Equations

Virtual Array Size:
N_virtual = N_tx × N_rx

Angular Resolution:
Δθ = λ / (N_virtual × d × cosθ)

SNR Comparison (MIMO vs Phased Array):
SNR_MIMO = SNR_PA - 10 log(N_tx) (per virtual element)

Where d = element spacing (λ/2 for half-wave), θ = look angle. Example: 4Tx/8Rx = 32 virtual elements, Δθ at broadside = λ/(32×λ/2) = 3.6° vs 15° for 12 physical elements in phased array mode.

MIMO vs Phased Array Radar

Parameter	Phased Array	Co-Located MIMO	Advantage
Angular resolution	λ/(Nd cosθ)	λ/(N_txN_rxd cosθ)	MIMO (N_txx finer)
Per-target SNR	N² (coherent)	N_tx×N_rx	PA (N_tx dB more)
Sidelobe control	Physical array taper	Virtual array taper	MIMO (more DOF)
Target scintillation	Single observation	N_tx diversity obs.	MIMO (up to 10 dB)
Simultaneous beams	Limited (DBF)	Full FOV	MIMO

Common Questions

Frequently Asked Questions

How does the virtual aperture work?

Each Tx-Rx pair has phase center at (x_k+y_l)/2. With N_tx transmitters at N_rxd spacing and N_rx receivers at d spacing, virtual phase centers form a filled ULA of N_tx×N_rx elements. Example: 4Tx/8Rx = 32 virtual elements from 12 physical antennas. Aperture savings: N_tx×N_rx/(N_tx+N_rx).

What are the advantages over phased arrays?

Finer angular resolution (N_txx improvement), waveform diversity (reduces Swerling scintillation by up to 10 dB), full-FOV simultaneous coverage, and more sidelobe control DOF. Disadvantage: per-target SNR reduced by 10log(N_tx) dB since power is spread across waveforms instead of focused.

What waveform orthogonality methods are used?

TDM (sequential Tx, simple but reduces effective PRF), FDM (frequency offsets, full Doppler but wider bandwidth), and CDM (phase codes like Hadamard, simultaneous Tx with full Doppler). TDM dominates automotive MIMO (TI AWR, NXP S32R). CDM used in military systems. Hybrids (TDM+CDM) offer practical compromises.

Related Terms

← Co Channel Interference Satellite Co Pol Pattern →