19. Project development¶
The information about the final project can be found here. Here I am gonna drawn some conclusion about the project developed in this course. I will try to answer the following questions in this text:
- What tasks have been completed, what hasn’t?
- What has worked, what hasn’t?
- What questions need to be resolved?
- What will happen & when?
- What have you learned?
It was possible to design a prototype of the arm that have the necessary movement to catch objects. However, the servo I used was not strong enough, and because of the short time I could not replace it for a strong one. Even though, the arm could responde the signal of the object detection to catch the object. For the next version, improvements will be done to enable the full movement of the arm.
The mobile application was done and allow the user to select which type of object the arm should look for. Improvements could be added in the app, for example validation of content. But these validations does not invalidate the app and it was possible to integrate it in the system. The user need to work consciently and should not perform invalid actions that would trigger errors in the application. This is an improvement for the next phase of the project.
The Object detection worked as expected. It process the images from a camera stream and provide the label and accuracy of the objects detected. The database model used was experimental and for a real application it should be adapted for specific object of interest. However, the project showed that machine learning could add extra functionality to a generic robotic arm.
The arm should be able to move around the space looking for the object of interested. Unfortunatelly, it was not possible to develop the navigation system of this project. Because of this, I could not check if the arm can work autonomously (one of the questions not answered in this project).
Even though this was an important step, it did not take the merit of the project. Because the project itself required more time to be able to develop each component of the system. It was able to finish many steps of the final idea and show a prototype arm working with the application interface and object detection. Next phase of the project might fill the mentioned gaps and provide a better prototype of the entire system.
I tried to develop the final project during the course, using as much as possible the assignments to design part of the final project. It was not easy, because the assignments are most of the time specific and not related to the final project. But I could use the time to develop important parts of the application and finish one prototype working with the mobile application and object detection.
I have learn many things with this final project. How to operate laser machine, milling machine, 3D printer, electronic production, to name a few. But mostly, I learned that having the right tools, we can design a better world. We can use the resources available to create meaningful projects and delivery nice applications that might improve the quality of live.
Extra questions¶
- Did you have/made a plan in the beginning?
Yes. I have a big interesting for the area of artificial intelligence and I wanted to build something that I could use it. I have done some research on the internet to get insperation and when I saw the robotic arm, I got the idea I tried to implement in my project. I have thought about the steps I would need to implement the full idea and executed step by step. Unfortunately, I did not have time to implement all the steps, but I got very close to it!
- How did it changed as you moved along in the course?
I had a good plan since the beginning and it did not change along the time. I did not had problems during the execution that required any change in the plan.
- Did you change/pivot/do something else apart from your plan?
Unfortunatelly, I did not have more time to design anything besides the initial plan of the project. I truly wanted to have more time and improve, for example the mechanic arm, perhaps use stronger servo. However, I could demonstrate that the idea work and it is feasible to design a smart robotic arm that catchs objects using object recognition.
- More information about the project development can be found at Principles and practices assignment.