Week 1 & 2 - minaamakary/MENA GitHub Wiki

05-07-2024 - Setting up the workspace

Downloaded all the necessary libraries on Parallels Desktop (Virtual Machine - Ubuntu Linux) and successfully created a workspace on visual studio code.

We are running this project on a Mac M1 2020.

06-07-2024 - Creating Gazebo world

Created a neat workspace on ROS with a bedroom containing a bed, a bench and a TV on top of it. The world seems to run fine, but there are no walls around the room, and the bench, bed and tv are in different locations.

07-07-2024 - Literature review

Downloaded all the necessary research papers, showcasing previous projects that have imnplemented a similar painting robot.

08-07-2024 - Modifying the room (Adding walls)

Modified the room now to have 4 walls and modified the locations of the bench, tv and bed to make the room look more realistic. Also added different colors to each wall. While this is not realistic and might not look very appealing, this would be for testing purposes on one room, to see if the machine learning model suggest templates that appeal for each wall color.

Below is a picture of the world created on Gazebo up to date.

09-07-2024 - Modifying the wall part II - Adding Accessories

Added some accessories in the room but there's an issue with positioning that I am trying to solve.

10-07-2024 - Adding 2D LiDAR

Added a 2D LiDAR to the room. Created a script for reading data from the 2D LiDAR but currently not running.

14-07-2024 - Adding 3D LiDAR

Added a 3D LiDAR to the model replacing the 2D Lidar for an effective vertical scan. There seems to be a logical error with the reading. The code is running, the topic is being echoed and displays data but the script doesn't display all the methods in the class created.

Below is the representation of the world with the 3D LiDAR installed and running.

15-07-2024 - Script Debugging

The code seems to have logical errors. Modified the code instead of a class with methods, removing the class and adding independent functions but getting the same exact result. The code runs, the topic /velodyne_points is subscribed, and when it is echoed in the terminal gives results, but the script doesn't. From debugging it seems to be an issue with the arguments in the function because we are passing (pointcloud2, pointcloud_callback)

16-07-2024 - Literature Review

Worked on the literature review, most of the related works tend to use 3D LiDAR for effective scanning, some works focused on using a vertically moving 2D LiDAR but that could lead to some miscalculations and errors. There are works that combined ultransonic sensors and 2D LiDARs to increase accuracy. Most of the work that was done using 3D LiDARs had limitations would computing power and high cost.

17-07-2024 - Successful 3D LiDAR Readings

Successfully worked out the issue with the code preventing the reading of data from the 3D LiDAR. Since the topic /velodyne_points is publishing messages of type sensor_msgs/LaserScan, we modified the code to handle LaserScan messages. Also modified the code to discard infinite readings from the LiDAR and converted LaserScan to cartesian co-ordinates. The pointcloud arguments passing through the function was replaced with (LaserScan, laserscan_callback) and the readings are now being displayed. Below is a representation of the result.

18-07-2024 - Added read image script

Added a line to the lidar3Dwall_measurement.py to publish the wall dimensions to a node (wall_dimension node). Created a readymade.py script to read an image from the directory and based on the published data on wall dimension node, the script will resize the image to fit the wall. In this testing phase we assume 1 meter = 100 pixels.

19-07-2024 - Testing Script

Testing worked in rescaling an image.