Week 4: Change of Operating System and Problems

Change system and problems

In the lab section of this week, the research group had a difficult time. In detail, it is difficult to have the video of the camera been transmitted to the computer. The data flows to the raspberry in the form of coordinate, length and height of the marked blocks. In this way, the processor of the pi could be relaxed. However, in the first place，the basic function of the car is that the user could see what the car sees with the use of camera. If the control system is not visual, it is meaningless to have such a complex system.

To make it possible to visualize the streaming video, the research group checked the driver of the camera. It turns out that it is not designed for Raspbian, and this made its function been limited to this system.

To solve this problem, the research group had to change the system of the smart car and recode the programs that have been tested. In specific, the driver is designed for a pure Linux environment.

Having done a lot of research, the research group learnt that ubuntu could be adapted to a raspberry pi 4b. Then, we began to change the system. The first step was to clean the SD and install a new one. Int his part, we chose ubuntu 19.10.1 as the new operating system. We copied the XZ file from the official web and flashed it into the SD. Secondly, we need to install a desktop to the system. However, even though we have tried to do this under the guidance of the instruction, it responded in a failure. It says we needed to refill the content. After checking this for a second time, we discovered that we needed to put the pi onto the internet to have the pi download the rest part of the system and come into function.

As is said from the beginning, the work this week was difficult. Having succeeded installed the system, we found that it is very slow to enter the system and the programs could hardly be run smoothly on the pi. Then we noticed that our pi was a low version which only has a memory of 1 GB, and it is insufficient for analyzing video while controlling different parts of the car. Which means we had to divide the job of processing the video to the camera itself.

In this way, we had to change our method that we use the ubuntu as our backup plan and use Raspbian as our first choice. We bought another SD and reinstalled the Raspbian to reappear our previse functions while trying to use ubuntu to achieve further functions.

Use of object identification 

In this part, the research group have succeeded identifying three objects of different colours. The detailed video is shown below.