Researching existing systems
Hi!
I will be researching existing robotic systems that have been published online by identifying the type of robot, components and algorithms that the creator(s) have used to aid my project development. This is also to avoid completely copying everything that they have implemented within their projects and to allow myself to have a personal contribution to the project topic of having a robot play air hockey. For a more detailed analysis, please click here.
Most projects documented on paper have used a Cartesian Robot controlled by 1-2 independent motors. Basic systems only uses 1 motor that moves along the y-axis of the robot, with the main goal of only saving the puck from going inside the robot's goal. Whereas, more sophisticated systems can save and hit the puck to the opponents side by having 2 motors that move in both x- and y-axis of the robot.
Listed below are some links that used a Cartesian Robot:
I will be researching existing robotic systems that have been published online by identifying the type of robot, components and algorithms that the creator(s) have used to aid my project development. This is also to avoid completely copying everything that they have implemented within their projects and to allow myself to have a personal contribution to the project topic of having a robot play air hockey. For a more detailed analysis, please click here.
Most projects documented on paper have used a Cartesian Robot controlled by 1-2 independent motors. Basic systems only uses 1 motor that moves along the y-axis of the robot, with the main goal of only saving the puck from going inside the robot's goal. Whereas, more sophisticated systems can save and hit the puck to the opponents side by having 2 motors that move in both x- and y-axis of the robot.
Listed below are some links that used a Cartesian Robot:
- PDF: Autonomous Air Hockey - Chalmers University of Technology
- YouTube: New Air Hockey Robot: Smartphone vs Human
Other documentations have used a SCARA which moves at x-, y- and z-axis of the robot. Although, I personally do not see the advantage of using a SCARA instead of a Cartesian Robot in this case, as the z-axis is bounded on the table's surface. Reading the documentations have not suggested any reasons for the need of moving at z-axis for their project.
Listed below are some links that used a SCARA:
- PDF: Air-Hockey Robot that autonomously changes strategies
- YouTube: Autonomous Robotic Air Hockey Test
Apart from these two types of robots, I have only read few papers that document using other types of robots. The most notable robot type I have seen is through YouTube videos which uses an articulated manipulator, although there are no papers that document how the whole robotic system works.
Listed below are some video links that used an articulated robot:
Listed below are some video links that used an articulated robot:
As a result, I plan on using an articulated robot. The university have multiple robots of that type, therefore, I will run some tests on the ABB IRB 120 or Dobot Magician and decide on what robot I will use as the robot player.
_______________________________________________________________________________________________________________________
_______________________________________________________________________________________________________________________
Almost all of the documentations uses computer vision to detect the location of the puck within the table by using a camera which is mounted on top of the table to detect the colour of the puck. This is the most efficient way for localising the position of the puck as it only uses one sensor. Therefore, there is not much I can personally contribute to the problem of detecting the puck. However, if necessary I can implement more hardware such as having multiple laser sensors laid around the edges of the table to aid with localisation.
_______________________________________________________________________________________________________________________
_______________________________________________________________________________________________________________________
To predict the trajectory of the puck, the documentations have used many different methods from comparing the location of the puck between frames, comparing the location of the puck in relative to the opponent's paddle, and applying AI methods. The methods used are quite varied and not one solution is used for all cases of projects. Therefore, there is a possibility that I can personally contribute to the topic of predicting the puck's trajectory by implementing a new solution, or even integrating existing solutions together.
_______________________________________________________________________________________________________________________
_______________________________________________________________________________________________________________________
Reflecting my idea with previous existing robotic systems, I have concluded that there are many existing projects of having a robot play air hockey. Though, the main objectives of the task i.e puck detection and trajectory prediction have already been solved, there are still some functionalities that I can contribute to the topic.
For example:
- Having ABB IRB 120 or Dobot Magician as a player
- Using a raspberry Pi for processing data, existing projects documents using Arduino
- Using ROS (Robot Operating System) as a middle ware between different hardware that may speak at different languages.
- Creating an autonomous scoring system
By this week, I hope to decide on what manipulator I will use: either Dobot Magician or ABB IRB120.
Immanuel
Comments
Post a Comment