Click on the blue“Radar Journal”to follow
Some images and text are from Sha Wei and Liu Feng’s tweets that blend academia and art.
ChatGPT is like a large knowledge base that understands natural language. You can ask it questions in various fields (including life, science, technology, economics, etc.), and it can write novels, copywriting, and even computer programs based on your requests. As an artificial intelligence language model, ChatGPT can assist in handling various tasks related to academic research, including:
1. Data Collection and Analysis: Collect relevant data, organize and structure it, and perform various statistical analyses to gain insights.
2. Literature Review: Search and summarize relevant literature to help you stay updated on research trends.
3. Writing and Formatting: Assist in writing academic papers, research proposals, and other written materials. It can also help format and reference your work to ensure it is ready for submission.
4. Presentation Preparation: Help you create slides, posters, and other materials for presentations, ensuring your work is visually appealing and easy to understand.
5. Question Answering: Answer specific questions related to academic research, helping you quickly and efficiently obtain the information you need.
6. Research Suggestions: Provide suggestions for research topics, methods, and strategies to help you design and execute research projects.
7. Proofreading and Editing: I can proofread and edit your academic papers, helping you improve clarity, coherence, and consistency.
Note: While ChatGPT supports Chinese mode, the primary communication language remains English, so the following questions will be in English.
ChatGPT can answer specific questions related to research, helping us quickly and efficiently obtain the information we need. It can also propose designs and solutions for research problems, including identifying appropriate data collection and analysis methods and strategies. Below, we take the field of radar signal processing and electromagnetics as examples to see how broad ChatGPT’s knowledge is.
First Question in Radar Signal Processing: If multiple radar sensors are placed together and work simultaneously, how can we measure and reduce the interference between them? ChatGPT provided solutions from different perspectives, very professional and comprehensive. The answer is as follows:
Second Question: If I want to design a high-resolution automotive radar for ADAS applications, how can I increase point cloud density by adjusting parameters?
We posed a design question to increase point cloud density, and ChatGPT provided six suggestions. However, it was noted that the solutions were not comprehensive, focusing solely on range resolution without mentioning the currently concerning issue of angular resolution in the industry. Each solution was adequately explained, but no rigorous criteria were provided from a formulaic perspective. Both questions reveal that ChatGPT has strong summarization and overview capabilities for broader questions, and it concluded both answers with a summary of the complexity and limitations of the issues.
Third Question: What are the data sources for microwave remote sensing satellite images?
ChatGPT provided five data sources, including major research institutions and commercial companies from Europe and the United States, but did not include Asian research institutions, indicating that the data used by the ChatGPT model from the Asian region is relatively scarce.
Next, let’s ask some questions in the field of electromagnetics.
First Question:The Chu limit in antenna theory. Here is its answer, quite professional!
Second Question, which is engineering-related: How to improve the gain of MIMO antenna arrays? Its answer was very comprehensive, quite impressive!
Third Question about computational electromagnetics: Its answer was even more considerate, especially the last paragraph, resembling an experienced teacher.
The design and validation of research methods require simulation experiments. Many researchers have experienced the difficulty of finding specific language codes online. So, how is ChatGPT’s programming capability? ChatGPT can write code in various languages such as Matlab, Python, and R based on your requirements, and it also provides comments!
Question 1: Write code to process raw ADC data captured by TI 3Tx-4Rx radar as input and output the radar micro-Doppler spectrogram.
The code provided by ChatGPT is as follows:
% Load the raw ADC data into a matrix
data = load(‘raw_data.mat’);
data = data.raw_data;
% Perform fast Fourier transform (FFT) on each received channel
N = length(data); % length of data
spectrogram = zeros(N, 4); % 4 receiver channels
for i = 1:4
spectrogram(:, i) = fft(data(:, i));
end
% Calculate micro-Doppler spectrogram by taking the magnitude square of the spectrogram
micro_doppler = abs(spectrogram).^2;
% Plot the micro-Doppler spectrogram
figure;
imagesc(micro_doppler);
colormap(jet);
colorbar;
xlabel(‘Receiver channel’);
ylabel(‘FFT bin’);
title(‘Radar micro-Doppler spectrogram’);
From the provided code, ChatGPT correctly understood our needs and provided a logical framework for the code, including data reading, micro-Doppler calculation, and plotting. It is noteworthy that we did not specify the programming language, but ChatGPT provided Matlab code. Given that many open-source codes in radar signal processing are completed in Matlab, it can be inferred that ChatGPT has learned from this part of the code. From a signal processing perspective, since we did not explicitly state that it was FMCW radar, the Matlab code provided does not conform to the FMCW process. We can speculate that correctly using ChatGPT requires providing a more detailed description. In contrast, humans generally ask for clarification on ambiguous questions rather than directly providing an answer. From the perspective of code correctness, we can see that the syntax is generally correct.
Question 2: The killer application of 5G communications, MIMO multi-antenna beamforming algorithms. With a good prompt, the MATLAB code can run with almost no modifications and yield correct results. I have attached the figure. Perhaps at this moment, you will recall a saying from Stephen Chow: “Surprised or not, unexpected or not, happy or not?”
Please write a MATLAB program about beamforming of the 1D antenna array lined up along the x-axis. Consider 10 antenna elements with uniform spacing of 0.5. The wavelength is 1. The radiation angle is from 0 to 180 degrees, where the x-axis is related to 0 degrees. Please use the following steps to write the program:
-
Establish a steering matrix A with the size of 181×10, which gives the array factors of all antenna elements covering all radiation angles. 181 is the number of matrix lines or radiation angles. 10 is the number of matrix columns or antenna elements.
-
Establish a radiation vector b with the size of 181×1. The elements are all zeros except for the 31st and 61st elements, which means the array radiates waves along both 30 and 60 degrees.
-
Get the complex excitation amplitude of each antenna element y by using y=A\b with the least square method.
-
Calculate the radiation pattern of the array using the excitation amplitude y and the steering matrix A, i.e., res=A*y.
-
Please draw the radiation pattern with polar coordinates.
% Define the number of elements and spacing
N = 10;
d = 0.5;
% Define the number of angles and wavelength
angles = 181;
lambda = 1;
% Compute the steering matrix A
A = zeros(angles, N);
for i = 1:angles
for j = 1:N
A(i,j) = exp(-1j*2*pi*(j-1)*d*sin(pi*(i-1)/180)/lambda);
end
end
% Define the radiation vector b
b = zeros(angles, 1);
b(31) = 1;
b(61) = 1;
% Compute the excitation amplitude y
y = A\b;
% Compute the radiation pattern res
res = A*y;
% Plot the radiation pattern in polar coordinates
polarplot(linspace(0, pi, angles), abs(res));
Question 3: Basic program for quantum electromagnetics, simulation of Rabi oscillation in a two-level system. Run MATLAB, and the results are completely correct. I have attached the figure. Is it not more knowledgeable than the average electromagnetics person?
Please write a MATLAB code to calculate Rabi oscillation of a two-level quantum system.
% Define time array for calculation
dt = 0.01; % time step size
T = 2*pi; % total time
t = 0:dt:T; % time array
% Define system parameters
Omega = 1; % Rabi frequency
gamma = 0; % decay rate
% Define initial state
psi_0 = [1; 0]; % initial state as [upper level population, lower level population]
% Calculate Rabi oscillation using ODE solver
options = odeset(‘RelTol’, 1e-10, ‘AbsTol’, 1e-10);
[~, psi] = ode45(@(t, psi) rabi(t, psi, Omega, gamma), t, psi_0, options);
% Plot Rabi oscillation
plot(t, psi(:, 1), ‘LineWidth’, 2);
xlabel(‘Time’);
ylabel(‘Population’);
title(‘Rabi Oscillation’);
% Define ODE function for Rabi oscillation
function dpsi = rabi(t, psi, Omega, gamma)
dpsi = [-1i * Omega * (psi(1) – psi(2))/2; 1i * Omega * (psi(1) – psi(2))/2];
end
Question 4: 2D FDTD program. Haha, not understood, but it provided a clear programming idea! However, it also indicates that its ability is still limited for complex tasks.
Please write a 2D FDTD code for calculating scattering from a 2D PEC cylinder. Unfortunately, I can’t provide you with a complete code, but I can give you some pointers on how to implement the 2D FDTD method for calculating the scattering from a 2D PEC cylinder.
-
Initialize grid and time step: You need to create a 2D spatial grid with a defined time step, typically the Courant stability criterion is used to determine the time step size.
-
Initialize fields: Initialize the electric and magnetic fields to zero.
-
Incident field: Specify the incident electromagnetic wave that will be scattered by the cylinder.
-
Update electric field: Calculate the electric field at the next time step using the previous magnetic field, the current electric field, and the time step.
-
Update magnetic field: Calculate the magnetic field at the next time step using the previous electric field, the current magnetic field, and the time step.
-
Apply PEC boundary condition: Apply the perfect electric conductor (PEC) boundary condition on the cylinder. This means that the tangential components of the electric field are set to zero at the cylinder surface.
-
Save fields: Save the electric and magnetic fields at each time step.
-
Scattered field: Calculate the scattered electric and magnetic fields by subtracting the incident field from the total electric and magnetic fields.
-
Plot result: Plot the magnitude of the scattered electric field over the 2D grid to obtain the scattering pattern of the cylinder.
Note that this is a high-level overview and some important details, such as absorbing boundary conditions, may have been omitted.
Example 1: Writing a Research Paper
The first step in writing a research paper is also the hardest, which is writing the introduction of the entire paper! Look at what ChatGPT can help us with! It is really friendly for non-native English speakers or those writing English academic papers for the first time!
I provided a paper title and abstract, asking it to help write the introduction. Moreover, it continuously revised based on my suggestions, and I asked it to elaborate on OAM generation methods and add some references. The final effect is attached, is it not very professional?
The title of the paper is “Ultrathin Complementary Metasurface for Orbital Angular Momentum Generation at Microwave Frequencies.” The abstract is “Electromagnetic (EM) waves with helical wavefronts carry orbital angular momentum (OAM), which is associated with the azimuthal phase of the complex electric field. OAM is a new degree of freedom in EM waves and is promising for channel multiplexing in the communication system. Although the OAM-carrying EM wave attracts more and more attention, the method of OAM generation at microwave frequencies still faces challenges, such as efficiency and simulation time. In this communication, by using circuit theory and equivalence principle, we build two simplified models, one for a single scatter and one for the whole metasurface to predict their EM responses. Both of the models significantly simplify the design procedure and reduce the simulation time. In this communication, we propose an ultrathin complementary metasurface that converts a left-handed (right-handed) circularly polarized plane wave without OAM to a right-handed (left-handed) circularly polarized wave with OAM of arbitrary orders, and a high transmission efficiency can be achieved.” Could you help to write an introduction?
Please add more OAM generation methods in the introduction.
Electromagnetic (EM) waves with helical wavefronts carry orbital angular momentum (OAM), which is associated with the azimuthal phase of the complex electric field. OAM is a new degree of freedom in EM waves and has gained attention due to its potential in channel multiplexing in communication systems. In recent years, various methods have been proposed to generate OAM-carrying EM waves, including optical vortex generators, metallic spiral phase plates, twisted-nematic liquid crystal displays, and metasurfaces. However, the methods still face challenges such as efficiency and simulation time.
In this communication, the authors present a novel ultrathin complementary metasurface for generating OAM-carrying EM waves at microwave frequencies. The metasurface converts a left-handed (right-handed) circularly polarized plane wave without OAM to a right-handed (left-handed) circularly polarized wave with OAM of arbitrary orders, and achieves a high transmission efficiency. By using circuit theory and the equivalence principle, the authors build two simplified models, one for a single scatterer and one for the whole metasurface, to predict their EM responses. Both models significantly simplify the design procedure and reduce the simulation time.
Please add more references
References:
-
M. J. Padgett and R. W. Boyd, “The Poynting vector in Laguerre-Gaussian laser modes,” J. Mod. Opt. 41, No. 7, pp. 2375-2381, 1994.
-
S. M. Zhao, Q. H. Liu, J. X. Chen, J. H. Wen, J. Liu, Y. J. Huang, and J. R. Mosig, “Metasurface for simultaneous manipulation of multiple orbital angular momentum states of electromagnetic waves,” Nat. Commun. 7, No. 1, pp. 13155, 2016.
-
S. Zhang, Y. Wang, J. Gao, X. Zhang, Y. Liu, and S. He, “General approach for beam shaping of OAM carrying waves with high topological charges,” Sci. Rep. 7, No. 1, pp. 16841, 2017.
-
J. K. Zhang, C. Guo, Q. J. Zeng, Y. Q. Wei, Y. H. Lu, L. Liu, and X. C. Jiang, “Orbital angular momentum multiplexing in millimeter waves for broadband wireless communications,” J. Opt. Soc. Am. B 33, No. 12, pp. 2455-2460, 2016.
Regarding whether ChatGPT can be used to write articles, different publishers have their own attitudes and regulations. For instance, the Nature editorial board explicitly states that ChatGPT cannot be listed as one of the authors. If ChatGPT is used in writing an article, it needs to be mentioned in the Methods or other sections (Nature 613, 612 (2023)). Some journals allow ChatGPT to be listed as an author and to modify articles using ChatGPT. Therefore, the use of ChatGPT for writing articles should be approached with caution, and one should carefully review the journal’s regulations.
Example 2: A common problem faced by scholars: how to quickly and effectively write a formal invitation letter inviting experts to join the Technical Program Committee (TPC) of an international conference. Below is a nice prompt we provided; let’s see how ChatGPT responds. You can also test it by pasting this prompt and receiving different yet equally impressive invitation letters. The content generated by ChatGPT also included many elements that I did not provide (highlighted with underscores), which is really considerate!
I want to write an English invitation letter on behalf of Zhejiang University, inviting a professor or doctor to join the TPC Member of the international academic conference. Please make it as polite and professional as possible, and the content of the invitation letter needs to include the following information:
-
Conference Name: 2023 International AI Symposium (AIS 2023)
-
Date: August 1-3, 2023
-
Organizer: Zhejiang University
-
Conference Venue: Deefly Zhejiang Hotel with the Address: No. 147 Santaishan Road, Hangzhou, Zhejiang Province
-
Conference Impact: AIS is a renowned international conference in the field of artificial intelligence, held annually in China.
-
Conference Content: keynote speeches, invited talks, short courses, special sessions, workshops, student competitions, etc.
-
TPC Member Responsibilities: review conference papers and organize sessions
-
Please respond by March 1, 2023, if you are willing to accept the invitation.
Dear [Professor/Dr.],
I hope this email finds you well. On behalf of Zhejiang University, I am writing to extend a formal invitation for you to join the Technical Program Committee (TPC) Member for the 2023 International AI Symposium (AIS 2023).
The AIS 2023 will be held from August 1st to 3rd, 2023, at the Deefly Zhejiang Hotel located at No. 147 Santaishan Road, Hangzhou, Zhejiang Province. As one of the most well-known international conferences in the field of artificial intelligence, AIS 2023 is held annually in China and provides a platform for researchers and scholars to exchange and share their latest findings and ideas.
As a TPC Member, you will play a vital role in reviewing conference papers and organizing sessions, helping to ensure the highest quality of presentations and discussions. The conference will feature keynote speeches, invited talks, short courses, special sessions, workshops, and student competitions, offering a wide range of opportunities for participants to deepen their knowledge and engage in meaningful discussions.
We would be honored to have you join us as a TPC Member for this exciting event. Please let us know by March 1st, 2023, if you are willing to accept this invitation.
Thank you for your time and consideration. We look forward to your response.
Best regards,
[Your Name]
On behalf of Zhejiang University
Most researchers, while doing research work, also need to undertake some teaching tasks. So how does ChatGPT perform in terms of assignment design?
First, I asked it to create assignments about transmission lines, and it did quite well, with increasing difficulty.
Then I asked it to solve the problems it generated, and it had no issues. In the future, if students use this software platform, the teaching pressure will be significantly reduced.
After reviewing the above examples, do you feel more confident in how to communicate with it? Remember, you can engage in multiple exchanges with it, continuously making your requests, and it will keep improving. It can even remember your preferences until you log out of the system. However, it is also evident that ChatGPT’s direct iteration on data and models from the Asian region still needs improvement. In the future, ChatGPT will integrate with search engines, cloud computing, Office, and other application software. What will the future world look like? Let us wait and see! You can also freely explore more features of ChatGPT.
↓↓↓
Related Reading
Radar Science | How Far Can Radar “See”?
Support — Our Journey Together Through 2022
Doctor Forum | Design Method for Multi-Radar Cognitive Tracking Quantizer Threshold in Dynamic Communication Environments (Video)
[Latest Results] Beam Residency Scheduling for Distributed Radar Networking System Based on Pulse Interleaving (Video)
[Latest Review] Development Status and Trends of Airborne Radar Warning Receivers (Video)
Academic Lecture | Multi-Channel SAR Processing Method Based on Image Domain (Video)
Academic Lecture | Unified Explanation of Multi-Angle Understanding of Synthetic Aperture Radar and Time/Frequency Imaging Algorithms (Video)
Editor: Zhou Qiyu, Jia Xing
Review: Gao Shan Liu Shui
Guidance: Jia Shouxin, Sha Wei
Feel free to share our original content, but reprints and excerpts require authorization from us and must indicate the original author and source as “Radar Journal.”
The information published by us aims to promote communication, and the content is the responsibility of the author, not representing our views. If there are any issues related to text, images, copyrights, etc., please contact us within 20 days, and we will handle it promptly. “Radar Journal” has the final interpretation right.
