One of the goals of the self driving rover project has always been to have it controlled by a base “brain” unit that does the actual analysis and motion control. I have a few XBee S1 units I had purchased a long time ago and I wanted to put them to good use for this. XBees are quite versatile but notoriously hard to configure. There are tons of configuration options available and the XCTU software is not the most friendly. ...
I made some OK progress on the driving unit last night. I finally got the T connector for the LiPo battery, so I could hook everything up and power both the Arduino and the motors from a portable source. I was worried about powering them together but it turned out to be simple. I soldered a buck converter (voltage down) between the battery line and the motor shield. Set it up to give out a constant 4.5V, which should be more than enough for the 3V motors I have in there. ...
I had my first try at visualizing incoming data from the LIDAR last night. Good news is that it worked, and I got a successful SLAM map of the room. That’s pretty awesome. Bad news is I definitely have to redesign the pulley mechanism. Stepper motor is ridiculously loud, heavy and slow. I need to design a smaller pulley system that runs on an encoded DC motor. I used a Python library called BreezySLAM for the SLAM implementation. It really simplifies the amount of work you have to do. I had to do some figuring out on Python serial comms, but once I got the data from Arduino to Jupyter, I had the image rendered in a few minutes. Here is the relevant piece of code: ...
Now that I have successfully encoded the stepper motor position for LIDAR, I turned my attention to encoding the DC motors that I am planning to use for running the robot wheels. There are several good examples of this online, but most use quadratic encoders (DC motors with embedded dual encoders in them, usually magnetic). Unfortunately I don’t have any such motors at hand, and I have many, many DC motors so I am hesitant to buy more. ...
I finally ran into an issue that I have been dreading for a while: Encoding the actual position of the LIDAR pulley mechanism. Now that the LIDAR scan works, I need accurate position data alongside the measurements. As reliable as my ridiculously powerful stepper is, it’s not a servo and doesn’t know which position it starts from. There are various ways of solving this, with varying levels of budget and difficulty: ...