یادگیری تقویتی و رباتیک 67813 /

شرح پروژه

#کد_پروژه: 67813

موضوع: پروژه یادگیری تقویتی و رباتیک /

من دانشجو هستم و دو تا پروژه دارم میخواهم برام اوکی کنین سلام هر دو پروژه رو با جزیات کامل براتون میفرستم Homework: SLAM Problem Solving with Reinforcement Learning

Aim:

In this assignment, students will apply reinforcement learning algorithms to solve the Simultaneous Localization and Mapping (SLAM) problem of a mobile robot using ROS. Students will develop a SLAM solution using a variety of reinforcement learning techniques to map the robot's environment and determine its own position.

Requirements:

A computer with ROS (Robot Operating System) installed.

Gazebo or similar simulation environment.

Python programming skills for reinforcement learning.

Steps:

Installation of ROS Environment:

Install ROS on your computer. Examine the basic concepts and operating principles in ROS.

Preparation of the Simulation Environment:

Create a mobile robot model in Gazebo or a similar simulation environment. The robot model should include sensors (e.g., lidar) and motors.

Here, the robot model must match the robot model given by Doctor Barış!

Understanding SLAM Algorithms:

Research SLAM algorithms and choose a suitable one and apply it.

Application of Reinforcement Learning Algorithms:

Apply reinforcement learning algorithms to map the robot's environment and determine its location. For example, algorithms such as Deep Q-Networks (DQN) or PPO (Proximal Policy Optimization) can be used.

Let's try the algorithm I gave at this stage, if it doesn't work, we can try other algorithms.

Training and Trial:

The robot needs to be trained in a simulation environment and then perform the SLAM task in real time.

Analysis of Results:

Evaluate the results obtained. Analyze the robot's mapping and location capabilities. Discuss the performance of the algorithms, success criteria, and potential improvements.

Reporting:

Report all your findings using visuals and graphs. نرم افزار پروژه ubuntu ve Rose  هست که البته متن رو کامل بخونین متوجه میشین چون استاد همه چی رو راجع پروژه کامل توضیح داده MTCS Homework

The objective of this homework is to gain hands-on experience with Monte Carlo Tree Search (MCTS) and apply it to a specific environment or problem of your choice. MCTS is a powerful algorithm widely used in artificial intelligence for decision-making in various domains, including games, robotics, and optimization problems.

Instructions:

Choose an Environment:

Select a specific environment or problem where you can apply MCTS. This could be a board game, a puzzle, a simulation, or any other scenario that involves decision-making.

Understand the Environment:

Provide a brief overview of the chosen environment. Explain the rules, goals, and any relevant information necessary for understanding the context in which MCTS will be applied.

Implement MCTS:

Implement the Monte Carlo Tree Search algorithm for your chosen environment. You can use any programming language of your choice. If you're not comfortable with programming, consider using existing libraries or tools that implement MCTS.

Experiments:

Conduct experiments to test the performance of your MCTS implementation. This may involve running simulations, playing games, or solving instances of the problem in your chosen environment.

Results and Analysis:

Present the results of your experiments and analyze the performance of MCTS in the given environment. Discuss any insights gained, challenges faced, and improvements that could be made.

Comparison:

Optionally, compare the performance of your MCTS implementation with other algorithms or strategies applicable to the chosen environment. Discuss the strengths and weaknesses of MCTS in comparison to alternative approaches.

Documentation:

Prepare a comprehensive document that includes the following:

Introduction to the chosen environment.

Description of the MCTS algorithm and its implementation.

Details of the experiments conducted.

Results and analysis.

Comparison with other approaches (if applicable).

Conclusion and future work.

Submission:

Submit your documentation along with any relevant code and data files. هردو پروژه در رابطه با یادگیری تقویتی و رباتیک هست استاد های که در این حوزه کار میکنن رو نشون بدین خودشون متوجه میشن زمان هم یک هفته است رشته من فوق لیسانس مهندسی کامپیوتر هستم هروقت اوکی کردین بهم خبر بدین  

Assignment: Classification of Oxford Flowers Dataset

Aim:

In this assignment, students will apply different ResNet models using the Oxford Flowers dataset and compare the results. They will also try to improve the performance of the models using finetuning.

Requirements:

PyTorch

The Oxford Flowers dataset (https://www.robots.ox.ac.uk/~vgg/data/flowers/17/index.html) should be downloaded.

At least three different ResNet models should be selected (for example, ResNet18, ResNet50, ResNet101).

Basic knowledge of fine-tuning.

Steps:

Preparation of the Dataset:

Students should download the Oxford Flowers dataset and prepare it for classification. It should apply preliminary processes such as standardization or augmentation by separating the data into training, validation and test sets.

Selecting ResNet Models:

At least three different ResNet models should be selected and these models should be implemented using deep learning frameworks.

Evaluation of Training and Results:

Train selected ResNet models and monitor their performance on the validation set during and after training. Also, evaluate the performance of the models on the test set.

Fine-Tuning Application:

Train ResNet models using fine-tuning. Monitor performance during and after training and evaluate the effect of fine-tuning.

Comparison of Results:

Compare results obtained using different ResNet models and fine-tuning. Evaluate the performance of models with metrics such as accuracy, precision, recall.

Reporting:

Present the experiments in a detailed report. The report should include a description of the models used, training processes, results and fine-tuning experiences. من دو تا پروژه فرستادم این شماره واسه جفت شون هست؟ این رو هم برسی کنید با محیط pytorch  پایتورج کار میشه و برای طبقه بندی از سه تا مدل استفاده میشه که البته همه توضیحات کامل ذکر شده هر سا تو شون رو نهایی کنین بعد قیمت هر سه تا رو بهم بگین خیلی ممنون 

مهارت ها و تخصص های مورد نیاز

درباره کارفرما

khaneprozhe

عضویت یک سال پیش

7359 پروژه ثبت شده ،

38 پروژه در حال انجام ،

205 پروژه آماده دریافت پیشنهاد ،

نرخ پذیرش پیشنهاد 14%