ROS 2 Nav2 Expert Needed for Industrial Mecanum AGV Navigation Stack

## Project Overview We're developing a large industrial warehouse AGV (1.23m x 1.0m) with mecanum drive running ROS 2 Jazzy (Ubuntu 24.04). The robot has dual RPLidar sensors, high-quality Phidget encoders, IMU, and a custom software stack. We've completed odometry validation and IMU integration but need expert assistance with Nav2 tuning and navigation stack validation. ## Current Status **Completed:** - Motor control and encoder odometry validated (~2-5% accuracy on mecanum wheel seyt.) - Wheel offset calibration for mecanum rollers - Dual LiDAR merger working needs calidation - Basic Nav2 parameter overhaul done (untested on hardware) **Needs Completion:** - TF tree and merged LiDAR scan alignment verification - AMCL localization testing and tuning - Nav2 configuration validation on real robot - Mission manager review and debugging (~650 lines Python) - IMU heading sign corrected - Mecanum wheels encoders. Odometry PODS? ## Scope of Work Following a structured consultant plan, complete points 3-6: 1. **TF & Scan Validation** - Verify transform tree, scan timestamps, coordinate alignment 2. **AMCL Testing** - Test localization with corrected odometry baseline, tune motion model 3. **Nav2 Configuration** - Validate costmap settings, DWB controller for mecanum, goal tolerances on real hardware 4. **Mission Manager** - Review state machine, action flow, error handling, multi-threading issues ## Required Skills - **Essential:** Strong ROS 2 (Jazzy/Humble) and Nav2 experience - **Essential:** Proven experience tuning autonomous navigation for real robots - **Essential:** Understanding of mecanum/omnidirectional kinematics - **Preferred:** Experience with large industrial AGVs (not just TurtleBots) - **Preferred:** Familiarity with AMCL, DWB controller, costmap configuration - **Nice to have:** robot_localization package for sensor fusion ## Technical Environment - **ROS 2 Distro:** Jazzy (Kilted) on Ubuntu 24.04 - **Compute:** Intel NUC - **Drive:** Mecanum (4x steppers via Arduino Mega) - **Sensors:** 2x RPLidar C1M1-R2, Phidget 4000 CPR encoders, Phidget IMU - **Languages:** Python (primary), C++ (existing nodes) ## Deliverables 1. TF tree and scan alignment validation report 2. AMCL configuration tuned for the platform 3. Nav2 parameter file validated on real robot 4. Mission manager code review with fixes/recommendations 5. Testing documentation and recommendations for production use ## Project Duration Estimated 15-25 hours needed asap ## Communication - Remote work (robot is in Durban, South Africa - can provide remote access or recorded test runs) - Video calls for real-time testing sessions preferred or even onsite testing