Aditi’s Status Report 4/11
In the first half of this week, I wrote a script that took commands in real time to move the robot, and simultaneously supplied these as odometry measurements to the SLAM, and at the end of a 30 second time frame saved the generated map. The position of the roomba is displayed as a series of points.
Here, the contours of the room in the area that the robot managed to explore, including the bottom left corner and the walls to the bottom and the left have been mapped quite accurately. However, the roomba made a roughly 90 degrees turn which has not been mapped quite as accurately. I hope to be able to fix this by making the calculation in the conversion from rotation speed (in mm/s) to (deg/s) more representative of the roomba.
We decided to address this issue together our work next work session. However, the team ran into a major hurdle during our work session – the roomba stoped responding to serial commands. We tried a number of ways to fix this, including restarting the system several times (raspberry pi + roomba), making sure the roomba and the raspberry pi were charged, sending different series of commands over serial (like reset, safe mode, passive mode), holding down the power button for 30 seconds etc. Unfortunately, even after spending a few hours on this issue, we were not able to fix the problem. I will call iRobot on Monday (tomorrow) to ask for assistance.
This hurdle has definitely hindered our progress. This coming week, I hope to fix it as well as write a more robust and accurate SLAM script.