What I Personally Accomplished This Week
This week I focused on integrating the complete navigation stack with SLAM and beginning end to end testing of autonomous path planning. The major accomplishment was successfully configuring Cartographer SLAM to handle faster robot motion, which was causing map distortion issues with our previous slam_toolbox implementation.
I created and tuned a custom Cartographer configuration file that increases the scan matching search window and adjusts motion filtering parameters to prevent tracking failures during rapid movements. The configuration specifies appropriate parameters for our RPLidar A1M8 sensor including 12m max range, and 5cm map resolution. This resolved the map distortion problem where the entire map would shift and warp when the robot moved quickly.
In coordination with Paul and Justin, I worked on initial motor setup and constructed a temporary testing car to validate the navigation with actual movement. We assembled a basic wheeled chassis and connected the motors to the Raspberry Pi. This temporary test vehicle allowed us to verify that velocity commands from Nav2 can successfully control physical motors, though we encountered some challenges with motor response calibration that will need further tuning. Having a mobile platform, even in this preliminary form, was essential for testing whether our SLAM and navigation algorithms perform correctly during actual motion rather than just simulation.
I also completed the required ethics module assignment for the course.
Progress Status
My progress remains on track with the project timeline. The Week 6 milestone (March 14) called for closed-loop navigation capability, and while the navigation planning pipeline is now functional and generating valid paths and velocity commands, full closed loop execution awaits motor controller integration from Paul. The remaining work involves connecting these velocity commands to actual motor control and integrating the human detection data from Justin’s vision pipeline to enable the human avoidance behaviors specified in our design.
Deliverables for Next Week
Next week I will begin implementing the behavioral finite state machine that manages transitions between our five operational modes (NORMAL, CROWD_AVOIDING, DENSE_CROWD, EMERGENCY_FALL, EMERGENCY_RUNNING). This FSM will subscribe to human detection data and velocity commands to determine appropriate robot behavior. Once Paul completes the motor controller, I will integrate it with Nav2’s velocity commands to achieve actual autonomous navigation. Finally, I will conduct comprehensive testing of the navigation stack in various hallway scenarios to validate path planning performance and identify any parameter tuning needed for smooth operation.