Post 19: Progress update

Diana: In Week 9, I worked on the navigation part so that the robot can replay poses saved in the map. While the robot is able to replay these poses, there is still a significant amount of error, so the accuracy of the navigation needs to be improved. I updated the gripper auxiliary part to help the robot avoid obstacles more effectively, but further improvements are needed to ensure more stable performance.

Shey: I worked on the UI implementation for our project this week. The interface features a simple, user-friendly design with two primary buttons: "Record Audio" and "Send Request to Robot". The "Record Audio" button lets users start recording their voice commands for the robot, while the "Send Request to Robot" button initiates the process of parsing the recorded audio and extracting keywords.

Evan: For the past week I worked on the writing a script that connects the perception and manipulation. I reorganized the codebase so that there's a single CLI tool that can send the robot to align to a marker, detect a given object, and then move to grab it. I also worked with Tony to further finetune the adjustments needed for accurately grabbing an object. Currently the given align_to_aruco code doesn't work perfectly. The end state of the robot doesn't quite align with what we need. More specifically, the camera will not be facing the marker or the table after it's aligned. Thus the next step is to modify the code to allow this.
For evidence, we show the "magic numbers" that we found through trial and errors and the codebase that contains all the features I mentioned above:
https://github.com/evanzhaodev/CSE481C-Sp25-Team3