robot_bridge
Connect embedded sensors to ROS 2 reliably.
A custom TCP bridge that receives LiDAR, IMU, and wheel-encoder telemetry from ESP32 and Raspberry Pi hardware and publishes normalized ROS 2 messages.
Robotics-first developer tools
Tools I'm building for robots, not humans.
Contributing to Pixel Engineering
Built an end-to-end ROS 2 SLAM pipeline
Built a complete ROS 2 robotics pipeline from sensor drivers to stable SLAM, using a custom TCP bridge, EKF sensor fusion, and slam_toolbox. Diagnosed encoder scaling, LiDAR timing, and gyroscope bias through TF analysis and controlled calibration—achieving stable loop closure, approximately 1.8% rotational error over 360 degrees, and approximately 1 cm positional error over a 1 m drive. Next: Nav2 autonomous navigation.
- ~1.8%Rotational accuracyerror over a complete 360-degree turn
- ~1 cmPositional accuracyerror over a 1 m straight-line drive
- Nav2Next milestoneautonomous navigation
The Stack
sensor_fusion
Turn noisy measurements into dependable motion estimates.
An EKF pipeline built with robot_localization, calibrated encoder scale, corrected gyroscope bias, and timing aligned with LiDAR updates.
slam_pipeline
Build stable maps that survive turns and close loops.
A slam_toolbox mapping pipeline tuned through TF analysis, timing inspection, and controlled calibration tests to eliminate map tearing.
embedded_telemetry
Stream real-world sensor data from constrained hardware.
Firmware and transport for LiDAR, MPU6500 IMU, and wheel encoders running across ESP32 and Raspberry Pi 3A+ devices.