Module-4 BEC703 Wireless communication systems
Multiple Input Multiple Output Wireless Communications: Introduction to MIMO
Communications, MIMO system Model, MIMO Zero Forcing Receiver, MIMO MMSE
Receiver, Singular Value decomposition of MIMO Channel, SVD and MIMO capacity,
Alamouti and Space-Time Codes, Nonlinear MIMO receiver: V-Blast, MIMO
Beamforming.
[Text1:6.1,6.2, 6.3, 6.4, 6.5, 6.6, 6.8, 6.9, 6.10]
1. With a neat diagram explain MIMO system model. (10 marks)
2. Explain MIMO zero forcing Receiver. (10 marks)
3. MIMO MMSE receiver. (10 marks)
4. Explain singular value decomposition of MIMO channel. (10 marks)
5. Note on a) SVD and MIMO capacity
b) Nonlinear MIMO receiver: V-Blast, MIMO Beamforming.
1. With a neat diagram explain MIMO system model. (10 marks)
Consider a wireless MIMO a wireless system with multiple antennas at both the transmitter
and the receiver.
• t: Number of transmit antennas.
• r: Number of receive antennas.
• This is called an r × t MIMO system.
The Transmit Side
x₁, x₂, ..., xₜ: These are the individual symbols (pieces of data) being sent.
Each symbol xᵢ is transmitted from its own dedicated antenna (the i-th transmit
antenna).
These symbols are grouped together into a Transmit Vector, x:
This vector x is a compact mathematical representation of what is being sent from all
antennas at a given moment.
The Receiver Side
y₁, y₂, ..., yᵣ: These are the individual symbols received at each antenna.
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Each receive antenna hears a combination of all the transmitted signals, altered by the
wireless channel.
This vector y represents the signals observed at all receives antennas.
Let the complex channel coeffient hij between ith transmit channel to j th receive
channel.
The net MIMO channel can be represented as
y=H x + n
The signal at the first receive antenna y1 is:
The signal at the first receive antenna y2 is:
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Case1: SISO (single input and single output antenna.
Case2: SIMO (Single input and multiple output)
Case3: MISO (Multiple input and single output)
Case4: MIMO (Multiple input and Multiple output)
The covariance matrix of the noise R n of the noise vector n is defined as
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
He noises at different antennas is independent so
The model is now extended to operate over discrete time instants, indexed by k:
The noise covariance matrix is given as
This noise process, which is uncorrelated across different antennas and time instants is termed
as spatio temporally uncorrelated noise.
2. Explain MIMO zero forcing Receiver. (10 marks)
• The Zero-Forcing (ZF) receiver is a linear detection method used in MIMO systems to
recover the transmitted signal vector x from the received signal vector y.
• Its core principle is to completely eliminate the inter-symbol interference (ISI) caused by
the channel mixing the data streams. It achieves this by applying the pseudo-inverse of
the channel matrix H to the received signal.
• The derivation is based on minimizing the Least-Squares error between the received
signal and the signal predicted by the channel model (MMSE).
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Least-Squares Estimation
For the over-determined case (r>t), the ZF receiver finds the estimate 𝑥̂ that minimizes the
squared error between the actual received signal and the signal that would be received if 𝑥̂
were transmitted.
The error function (or cost function) is defined as:
This is known as the Least-Squares error function. The vector 𝑥̂ that minimizes this function
is the Least-Squares estimator.
Step 1: Expand the Error Function
Step 2: Apply Vector Calculus to Find the Minimum
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
̂
Step 3: Solve for the Optimal Estimate 𝒙
Step 4: Generalize to Complex-Valued Signals
Step 5: The Final Zero-Forcing Receiver
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
3. Explain MIMO MMSE Receiver. (10 marks)
The Minimum Mean-Squared Error (MMSE) receiver is a linear detection scheme for MIMO
systems that aims to find an optimal balance between two opposing goals:
a) cancelling inter-stream interference and
b) suppressing noise amplification.
The Zero-Forcing (ZF) receiver, which only focuses on cancelling inter symbol
interference.
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
The MMSE receiver minimizes the average total error between the transmitted and estimated
signals, factoring in both the channel characteristics and the noise power.
This makes it more robust and performant than the ZF receiver, especially at low and
moderate Signal-to-Noise Ratios (SNR).
The fundamental goal of the MMSE receiver is to find the estimate 𝑋̂𝑀𝑀𝑆𝐸 that minimizes the
Mean-Squared Error (MSE):
Bayesian Approach: The MMSE receiver treats the transmitted signal vector xx as a random
variable with known statistical properties (e.g., power), unlike the deterministic view
sometimes taken for ZF.
Linear Estimator: It restricts itself to a linear estimator of the form 𝑥̂ = 𝐶 𝐻 𝑦 where C is a
receiver matrix to be optimized.
a) Let us consider SISO system
estimate a single transmitted symbol x from a vector of received signals
y=h x + n
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
b) Generalize the MIMO: The solution from the SIMO case is then extended to the full
MIMO case where a vector 𝑥̂ must be estimated.
The Final MMSE Receiver and its Alternative Form
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
4. Explain singular value decomposition of MIMO channel. (10 marks)
The Singular Value Decomposition (SVD) is a powerful mathematical tool that decomposes
the MIMO channel matrix HH into three simpler, structured matrices: U, Σ, and 𝑉 𝐻 .
This decomposition reveals that the MIMO channel can be interpreted as a set of parallel,
independent, non-interfering sub-channels.
The number of such usable sub-channels is equal to the rank of the channel, and their quality
is determined by their singular values.
1. The SVD Equation and Matrix Dimensions
2. Key Properties of the Matrices U and V
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
3. Key properties of the singular values
The SVD provides a profound physical interpretation of MIMO operation:
1. Precoding (V): At the transmitter, the data vector ss is multiplied by the matrix V. This
"rotates" the transmitted signal to align with the channel's eigenmodes.
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
2. Parallel Channels (Σ): The channel's effect is now simplified to a set of independent,
parallel sub-channels. Each sub-channel has a gain equal to a singular value σi. The first
sub-channel (gain σ1σ1) is the strongest, and subsequent ones are weaker.
3. Combining (𝑉 𝐻 ): At the receiver, the received signal is multiplied by 𝑉 𝐻 . This "rotates"
the signal back, separating the independent data streams perfectly.
This process transforms the complex, coupled MIMO channel into a collection of independent
Single-Input Single-Output (SISO) channels, dramatically simplifying the receiver design (no
need for complex ZF or MMSE to separate streams from each other).
6. Note on a) SVD and MIMO capacity
c) Alamouti and Space-Time Codes
d) Nonlinear MIMO receiver: V-Blast, MIMO Beamforming.
i) SVD and MIMO capacity
The Singular Value Decomposition (SVD) of the MIMO channel matrix H is a powerful
mathematical tool that reveals the channel's inherent parallel spatial channels or
"eigenmodes." By using the SVD for specific linear processing at both the transmitter
(precoding) and receiver (combining), the original coupled MIMO system can be transformed
into a set of independent, parallel Single-Input Single-Output (SISO) channels.
This process, known as spatial multiplexing, allows for the simultaneous transmission of
multiple data streams and provides a direct formula for calculating the fundamental channel
capacity of the MIMO system.
1. The SVD-Based MIMO System Transformation
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
a) Receive Combining: The receiver processes the received vector y by multiplying it
with the matrix 𝑈 𝐻 .
𝑦̃ = 𝑈 𝐻 𝑦
This aligns the receiver with the channel's spatial directions to separate the data
streams.
b) Transmit Precoding: The transmitter pre-processes the data vector 𝑥̂ by multiplying
it with the matrix V.
This aligns the data streams with the strong spatial directions (eigenmodes) of the
channel.
We have x=V𝒙
̂
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
2. The Equivalent Parallel Channel Model
The covariance of the modified 𝒏
̂ can be derived as
The ith MIMO channel is given by
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
3. MIMO Capacity and the "Water-Filling" Principle
ii) Optimal MIMO capacity
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
iii) Asymptotic MIMO capacity
b) MIMO nonlinear Receiver V-blast
V-BLAST is a nonlinear MIMO reception strategy
The core idea is to decode the transmitted data streams one by one, and after each
successful decoding, to cancel its contribution from the received signal before decoding
the next stream. This process reduces interference for the remaining streams.
The standard MIMO system model:
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
This means that the receive beamformer qi is perfectly orthogonal to all channel
columns hj except for hi. When applied to the received signal y, it will nullify (zero-force)
the interference from all streams except the ii-th one.
c) MIMO beamforming
(i) System Model and SVD of the channel
MIMO Beamforming is antenna technology that uses multiple antennas that create a
focused data signal directed at a specific direction, resulting in high SNR.
Consider a MIMO system.
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
(ii) Transmit Beamforming
The transmitted signal vector is therefore:
(iii) Signal Propagation through the Channel
Now, let's see what happens when this beamformed signal xx passes through the channel
HH.
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
(iv) Receive Beamforming (Maximal Ratio Combining)
At the receiver, the multiple antennas receive the signal y. To optimally combine these
signals, we perform Maximal Ratio Combining (MRC). The optimal combiner is the
matched filter to the effective channel, which is u1.
(v) Signal-to-Noise Ratio (SNR) Analysis
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]
Module-4 BEC703 Wireless communication systems
Prof. Shobha S Dept. of ECE SCE, Bengaluru-57
Studied smart, not hard — thanks to [Link]