Steve Macenski

ROS 2's navigation framework. Flexible behaviors, state of the art algorithms, modular architecture, and support any type of mobile robot.

Templated A*-based planning framework. 2D A*, Hybrid-A*, and State Lattice implementations for Ackermann, Legged, and Differential robots.

2D Lidar SLAM toolbox and default vendor for ROS 2. Map editing, session continuation, synchronous and asynchronous modes.

Model Predictive Path Integral controller. Optimizes trajectories through random perturbations with plugin-based customizable critic functions.

3D voxel layer using OpenVDB for dynamic environments. Novel decay functions and sensor models for multi-sensor integration.

Path tracking controller with velocity regulation for sharp turns and collision avoidance. Processes at >1kHz.

Autonomous charging and static infrastructure docking server for any type of robot and dock. Plugin-based framework with perception, approach, and staging behaviors.

Terrain-aware semantic navigation integrating SAM3 foundation model with Nav2. Text-prompted segmentation for obstacle detection and traversability estimation.

Drivers for Ouster lidar series (OS-0, OS-1, OS-2). 3D pointclouds, IMU, 2D laser scans. Donated to ros-drivers.

Official standard interfaces for radar sensors (radar_msgs) and AI applications (vision_msgs). Donated to ros-perception.