Control & Trajectory Optimization

← Home  |  All Research Areas

Path-Tracking Controller Benchmark (Pure Pursuit, MPC, MPPI) — ongoing. A full ROS 2 (Jazzy) navigation stack built from the ground up — AMCL localization, A* global planning, and three path-tracking controllers compared head-to-head on a seeded, repeatable goal tour with a custom 7-metric benchmarking harness. Results so far: Pure Pursuit 97% success at 0.069 ms/tick; MPC 98% success, 30% faster time-to-goal at ~330x Pure Pursuit's compute cost; MPPI 98% success at ~6x less compute than MPC with no solve-time spikes. (view the code)

Path-Tracking Controller Benchmark Pure Pursuit MPC MPPI ROS2 Navigation Stack AMCL Localization A*

Control Framework for an Inchworm Hybrid Steel-Bridge Inspection Robot. A dual-mode autonomous control system — mobile and inch-worm transformation, switched via a 3D point-cloud surface-availability algorithm — for autonomous visual inspection of steel bridges. (view the paper — IROS 2020)

Inchworm Robot Hybrid Robot Steel Bridge Inspection Dual-Mode Control 3D Point Cloud Object Detection FSM

Control Framework for Manipulators in Sorting Application. A vision-guided grasping framework integrating object detection, grasp-pose estimation, and trajectory planning for pick-and-place with an AUBO manipulator. (view the paper — IRC 2020)

Control Framework Sorting Application AUBO Manipulator Vision-Guided Grasping Object Detection Grasp-Pose Estimation Trajectory Planning

Distributed Control in Real-Time OS for an Inverted Rotary Pendulum. A state-space controller — one control input stabilizing a two-output pendulum system — modeled, simulated, and tuned in MATLAB/Simulink. (Published — IRJET 2016.)

TTCAN FreeRTOS Inverted Rotary Pendulum State-Space Control MATLAB/Simulink Distributed Control Embedded Systems