This week we focused on getting everything in perfect order before integration. First, we assembled the drone’s sensors and camera onto the adjustable mount and adjusted the hardware and tolerances until everything fit properly, the wires could connect safely and the sensors and proper line of sight.
Then, we focused on other miscellaneous tasks in preparation for full integration. We got the camera configured and working with the new RPi4, and started prototyping the circuitry for the passives for button control. In addition, we worked on the calibration of the camera in simulation (and measuring the actual position/pose of the real thing from our CAD design). Using this, we were able to perfect the transformation from 2D pixel coordinates to 3D coordinates in simulation and integrate the state estimator into the simulation. The successful integration of the state estimator into the simulation signaled that our drone tracking pipeline was finally complete.
Finally, we worked to start calibrating the LIDAR and optical flow sensor for drone flight. For next week, we plan to get the drone up in the air and perform rudimentary tracking. In preparation for this, we plan to write the ROS scripts to successfully communicate between the RPi and the TX1, fully calibrate the drone’s sensors and implement the button control to start and stop the main tracking program on the TX1.