代做EE5606 Artificial Intelligence for Antennas in Wireless Communication代写数据结构语言程序

日期：2025-11-08 06:41

Mini Project Worksheet: Dipole Antenna Array Optimization

Course: EE5606 Artificial Intelligence for Antennas in Wireless Communication

Report Submission Deadline: Week 12 (at the commencement of the lecture)

Project Background

Design a linear half-wave dipole antenna array that has a main beam direction of x as shown in Fig. 1. Traditional full-wave simulations (e.g., HFSS, CST) are computationally expensive. For simplicity, no full-wave simulations are required in this project. In this project, you will use an array factor formula to generate training data, train a fully-connected neural network (FCNN) to obtain a surrogate model, and then use an optimization algorithm to find an optimal array configuration.

Part I: Data Generation and FCNN Training

Fig. 1

Design Goal:

Generate a dataset using the array factor formula for a linear half-wave dipole antenna array, and train a FCNN to obtain a surrogate model that replaces the array factor formula.

Parameters:

1.   Operating frequency: f = 2.4 GHz

2.   Wavelength: λ = c/f

3.   The radius of each dipole element is b = 0.5 mm

4.  Number of elements: N = 32  for the entire project.

Design Variables:

1.   Element spacing d (in terms of λ)

2.   Progressive phase shift α (in radians) between elements

Electric Field Vector:

The electric field vector generated by the antenna array at a point in the far field.

where k = 2π/λ is the wavenumber in vacuum. For a half-wave dipole antenna, the element pattern he(θ) is given by:

where:

1.   the term accounts for the spherical wave propagation, which can be regarded as a constant in this project.

2.   he (θ) characterizes the radiation pattern of a single half-wave dipole element.

Analytical Formula – Array Factor:

The array factor for a linear array along the z-axis is given by:

where:

1.   an  = 1 (i.e., assumed uniform excitation),

2.  X is the angle from the x-axis, where cos X = sinθ cosφ .

Task 1:

Train a FCNN to predict the value of |E| for a given data input of (d, α , θ, φ), with d ∈ [0.3λ,0.8λ] , α ∈ [-π, π] , θ ∈ [0°, 180°] ,  and φ ∈ [0°, 180°] .  This  FCNN  (surrogate model) is used to replace the array factor formula.

Task 2:

By using the FCNN (surrogate model) and an optimization algorithm (Genetic Algorithm or PSO), find the optimized (d, α) that gives the maximum |E| at (θ = 59.8o, φ = 30.3o).

Submission Requirements

Upload your ZIP file to “Assignments” in Canvas. The ZIP file should contain:

1.  Code: Provide  comprehensive  comments  for  each part  (Data  generation,  FCNN  training, Optimization algorithm).

2. Brief Report: including methodology, results (e.g., training and validation errors of FCNN, convergence curve of optimization algorithm), discussion, and conclusion (no more than 4 pages, single line spacing, New Times Romance, font 12, 1-inch margin from top, bottom, left, and right).

3. Demo video: including the processes ofFCNN training and optimization.

4. Note: In your presentation slides, use diagrams, plots, figures, etc. as far as possible. Avoid long paragraphs and keep the text as concise as possible.

5. Presentation: (10 min + 2 min Q&A)

Suggestion:

(1) Use flowcharts to explain your algorithm workflows.

(2) Include figures to visualize algorithm performance (e.g., convergence curves, prediction error). Grading Criteria

Category	Weighting
Code Quality	25%	Readability, modularity, and documentation
Algorithm Implementation	25%	Correctness of optimization and FCNN
Report	25%	Organization, conciseness, and analysis
Presentation	25%	Clarity and time management