All classes MIMO-OFDM Channel Estimation Deep Learning Syllabus · Episode 4 of 6

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid

0 views

Tutorial on massive MIMO-OFDM channel estimation with CNN in PyTorch: dilated convolution, residual learning, hybrid LMMSE. Episode 4 of the massive MIMO-OFDM channel estimation deep learning series.

Software: Python (NumPy, SciPy, PyTorch, Matplotlib)
Level: Advanced
Duration: ± 11 min
Series total: 57 min · 6 videos
Published: October 25, 2026

Episode 4 of 6 · view syllabus

Download PDF (6 PDF pages)

0 downloads

Download Project File· 380 KB Unlock All Premium Classes ▶ Watch the video lesson0 plays

▶ Video Lesson

Watch every step above performed live, with English narration.

0 plays

Class book

Page 1 / 6

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Cover: Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation with DNN/CNN on Pilot Grid

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Before you start, steps list, and Step 1

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Step 2: Training with noise augmentation

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Step 3: Training & Comparing NMSE

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Step 4: Loss & Estimated Channel Maps

Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid: Done: Final Result and Further Practice

Cover: Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation with DNN/CNN on Pilot Grid

Click a page or use ◀ ▶ to turn · swipe on mobile · to enlarge

Want to go deeper?

Join one of our courses or ask us directly on WhatsApp.

Download PDF· 0 downloads View the full 6-episode syllabus View MIMO-OFDM Channel Estimation Courses Ask on WhatsApp

Recommended classes for you

Picked from the software, topic, and level of “Class MIMO-OFDM Deep Learning Channel Estimation #4: Channel Estimation using DNN/CNN on Pilot Grid”.

Deep Learning CNN-LSTM Pneumonia Detection

6 episodesFree preview

Python · Intermediate–Advanced

Deep Learning CNN-LSTM Pneumonia Detection

6 video lessons with PDF guides, 57 min in total. Starts with: Chest X-ray Dataset: Loading, Exploration & Data Splitting.

57 min 22 0

View class

18 episodesFree preview

Python · Beginner–Advanced

Machine Learning Python

18 video lessons with PDF guides, 3 h 41 min in total. Starts with: Introduction to Machine Learning & First Model.

3 h 41 min 61 0

View class

EKG Arrhythmia Classification Wavelet Deep Learning

6 episodesFree preview

Python · Intermediate–Advanced

EKG Arrhythmia Classification Wavelet Deep Learning

6 video lessons with PDF guides, 56 min in total. Starts with: EKG & EEG Signals: Characteristics and Public Datasets (MIT-BIH).

56 min 2 0

View class

50 episodesFree preview

Agentic AI · Beginner–Intermediate

Agentic AI Without Coding

50 video lessons with PDF guides, 14 h 29 min in total. Starts with: Chatbots, Automation, and AI Agents: What's the Difference?.

14 h 29 min 66 0

View class

100 episodesFree preview

Claude · Beginner–Advanced

Claude from Zero to Expert

100 video lessons with PDF guides, 15 h 1 min in total. Starts with: What Claude Is and How It Differs from ChatGPT and Gemini.

15 h 1 min 62 0

View class

6 episodesFree preview

Python · Intermediate–Advanced

Transfer Learning for Leaf Disease

6 video lessons with PDF guides, 1 h 5 min in total. Starts with: PlantVillage Dataset: Folder Structure, Class Distribution & Data Split.

1 h 5 min 9 0

View class

6 episodesFree preview

Python · Intermediate–Advanced

5G Deep RL Resource Allocation

6 video lessons with PDF guides, 1 h 1 min in total. Starts with: Concept of 5G/6G Resource Allocation: Resource Block, Power, User.

1 h 1 min 0 0

View class

6 episodesFree preview

Python · Intermediate–Advanced

Brain Tumor Segmentation with U-Net

6 video lessons with PDF guides, 1 h 10 min in total. Starts with: Brain MRI Dataset & Tumor Mask: Loading, Visualization & Patient-wise Splitting.

1 h 10 min 28 0

View class

Need more than a video?

Other ways we can help with Python.

Learn with a mentor

Scheduled AI & Machine Learning courses with live guidance and a certificate.

Course

Multi-Objective Optimization Course: NSGA-II, MOPSO & Hybrid Algorithms (Python)

Course

ECG Arrhythmia Classification Course with Wavelet & Deep Learning (PyTorch)

Course

ANFIS + PSO/GA Course for Non-Linear Time Series Forecasting (Python)

See all in AI & Machine Learning →

Have it built for you

No time to learn Python? Our team can do the project for you.

Service

Python Machine Learning — Prediction Models, Customer Segmentation, Fraud Detection & AI Apps

Service

AI Image Creation with Stable Diffusion — Unlimited Custom Visual Production

★ 4.9

Service

AI Image & Video Creation with Leonardo AI — All-in-One Visuals for Business

★ 4.9

See all in AI & Machine Learning →

Get notified about new Python classes

At most one email a week. Unfollow anytime with one click.

Transcript

The full narration of this video. The current sentence is highlighted while the video plays; click any sentence to jump there.

Show transcript

Introduction

This is the final result, a CNN that looks at the entire 32 antenna x 64 subcarrier grid as an image, then refines the LMMSI estimation. At 0dB SNR, it's 2.5dB better than LMMSI. Let's get started. Welcome to episode four of the massive MIMO FDM channel estimation series with Deep Neural Networks. Our LMMSI only uses correlation between subcarriers at each antenna separately. Massive MIMO channels are also highly correlated between antennas because there are only a few paths. Today we are making a CNN that uses both.

Step 1: Input & CNN Architecture

Create a file named model_underscore_cnn.py. The prepare_input function converts the LMMSI estimation into a 3-channel tensor, like a color image, with real, imaginary, and SNR fields, so the CNN knows how much noise to clean. The CNN channel class uses 3x3 convolutions with dilation 1, 2, 4, and 8, so the field of view covers almost all antennas and subcarriers with back normalization. The output is residual, insert LMMSI plus correction, so the CNN only needs to learn what LMMSI missed. The train function generates new pilot noise in each epoch as data augmentation, so the model doesn't memorize specific noise.

Step 2: Training with Noise Augmentation

Optimizer Adam with learning rate decreasing following a cosine curve, loss mean squared error against the actual channel, and back-64. The estimate function runs the model in chunks of 500 samples and returns the complex channel. Create a file named train_underscore_cnn.py and run it.

Step 3: Training & Comparing NMSE

This model only has about 48,000 parameters and is trained for 15 epochs on GPU, about 1 minute. Then we compare it on the test data. At -10dB, CNN is -1.6dB compared to LMMSI -0.8dB. At 0dB, CNN is -7.07dB compared to LMMSI -4.58dB, which is 2.5dB better. At 10dB, CNN is still 1.2dB better, and at 20dB both are equal, because when noise is small LMMSI is already almost perfect.

Step 4: Loss & Estimated Channel Map

Let's look at the graph. The training loss decreases steadily thanks to the cosine schedule. The next three images show the actual channel. LMMSI estimation at 0dB SNR still has noisy dots, and the CNN estimation is much cleaner and clearly shows the diagonal line pattern of the channel. CNN successfully used the correlation between antennas to clean the noise.

Done!

Done. We have designed a dilated residual CNN that refines LMMSI. Training with noise augmentation on GPU, and proving its superiority at low and medium SNR. In the next episode, we will investigate the effect of the number of pilots and the training SNR range. Currently working on a thesis, dissertation, or research in the field of telecommunications and artificial intelligence? Cerdas Inovasi Digital helps implement communication system research, from OFDM and multipath channels, massive MIMO simulations, least squares estimators, and LMMSI until channel estimation based on deep learning with PyTorch, as well as opening a Python programming course for electrical engineering and telecommunications. Want to master the entire series, from basic OFDM to NMSI and BER evaluation? All 6 episodes, complete with a step-by-step PDF guide and source code, Available for premium tutorial subscribers. Subscribe starting at Rp69.000 per month, save 28% with 12-month package, or one-time payment of Rp999.000 for lifetime access. Thank you for watching, see you in the next episode.

Ask the class material

Type a question or keyword; we find the part of this episode that explains it and jump the video there.

My notes

Notes are saved with the current video time and are only visible to you.

Questions & answers

Stuck on a step? Ask here; the instructor and other learners can answer.

No questions yet.

Your instructor

Cerdas Inovasi Digital Team

Software developers, BIM & 3D engineers, and AI practitioners

53 video classes1,296 episodes

We build software, BIM models, games and AI automations for clients every day, and turn that hands-on work into step-by-step classes: every step is shown on screen and comes with a PDF guide you can keep.

See our clients →

Memuat

▶ Video Lesson

Recommended classes for you

Need more than a video?

Transcript

0:03Introduction

0:46Step 1: Input & CNN Architecture

3:36Step 2: Training with Noise Augmentation

5:40Step 3: Training & Comparing NMSE

7:57Step 4: Loss & Estimated Channel Map

9:23Done!

Ask the class material

My notes

Questions & answers

Your instructor

▶ Video Lesson

Recommended classes for you

Need more than a video?

Transcript

0:03Introduction

0:46Step 1: Input & CNN Architecture

3:36Step 2: Training with Noise Augmentation

5:40Step 3: Training & Comparing NMSE

7:57Step 4: Loss & Estimated Channel Map

9:23Done!

Ask the class material

My notes

Questions & answers

Your instructor

Introduction

Step 1: Input & CNN Architecture

Step 2: Training with Noise Augmentation

Step 3: Training & Comparing NMSE

Step 4: Loss & Estimated Channel Map

Done!

Introduction

Step 1: Input & CNN Architecture

Step 2: Training with Noise Augmentation

Step 3: Training & Comparing NMSE

Step 4: Loss & Estimated Channel Map

Done!