WORK ACCOMPLISHED:
A lot of progress has been made but at a cost of a step backward. So the LiDAR compatibility with the raspberry pi has been a thorn for the last week and a half and a few days ago, I threw in the towel and outsourced. The first group I met with was Robo-club as we were using their 3d printers anyway. They had me follow some guides but I have followed most of the available info online already. When this did not work, I scheduled a meeting with a PhD student in Professor Rowe’s lab. After about two hours of debugging, he came to the conclusion that this LiDAR model is simply not compatible with our RbPi OS. We confirmed this by trying with an older model, a d435. When this worked, he let me borrow a d435 and we proceeded from there, concluding it was not possible to connect a d455. In addition, we have 3d printed an enclosure for our parts to sit underneath the board. We used a slightly modified version of a design found on thingiverse and will compile pieces tomorrow. I also wrote code to back up and move around an object that has been detected. Now, I need to write code with the new LiDAR to obtain a point cloud to be able to recognize objects and their size to move around them.
PROGRESS:
THe progress was monumental this week and we got over the LiDAR hump. We will grind tomorrow to produce a good demo and then in the subsequent week to piece all the parts together.
NEXT WEEK’S DELIVERABLES:
-
- Completing an object detection/avoidance end to end pipeline and to smoothly integrate that with the return feature
- Complete chasis by resizing enclosure, drilling LiDAR holes, drilling enclosure holes, drilling holes in enclosure for charging and power button
- Hopefully obtain some shorter wires
- bring velcro, window sill liner for enclosure