[Paper Review] Real-Time Kinodynamic Motion Planning for Omnidirectional Mobile Robot Soccer using Rapidly-Exploring Random Tree in Dynamic Environment with Moving Obstacles

Fahri Ali Rahman, Igi Ardiyanto+1 more|arXiv (Cornell University)|May 12, 2019

Robotic Path Planning Algorithms14 references4 citations

TL;DR

This paper presents a real-time kinodynamic motion planning system for omnidirectional mobile robots in dynamic soccer environments using kinodynamic-RRT* with double integrator dynamics for translational motion and a minimum-time trajectory generator for rotational motion. The system enables online replanning and trajectory tracking via PI control, achieving collision-free, dynamically feasible trajectories even with moving obstacles in ROS-based simulations.

ABSTRACT

RoboCup Middle Size League (RoboCup MSL) provides a standardized testbed for research on mobile robot navigation, multi-robot cooperation, communication and integration via robot soccer competition in which the environment is highly dynamic and adversarial. One of important research topic in such area is kinodynamic motion planning that plan the trajectory of the robot while avoiding obstacles and obeying its dynamics. Kinodynamic motion planning for omnidirectional robot based on kinodynamic-RRT* method is presented in this work. Trajectory tracking control to execute the planned trajectory is also considered in this work. Robot motion planning in translational and rotational direction are decoupled. Then we implemented kinodynamic-RRT* with double integrator model to plan the translational trajectory. The rotational trajectory is generated using minimum-time trajectory generator satisfying velocity and acceleration constraints. The planned trajectory is then tracked using PI-Control. To address changing environment, we developed concurrent sofware module for motion planning and trajectory tracking. The resulting system were applied and tested using RoboCup simulation system based on Robot Operating System (ROS). The simulation results that the motion planning system are able to generate collision-free trajectory and the trajectory tracking system are able to follow the generated trajectory. It is also shown that in highly dynamic environment the online scheme are able to re-plan the trajectory.

Motivation & Objective

• To address the challenge of real-time motion planning for omnidirectional mobile robots in highly dynamic, adversarial environments typical of RoboCup Middle Size League.
• To integrate robot dynamics and moving obstacle prediction into motion planning for improved safety and feasibility.
• To enable online replanning and trajectory tracking under time-varying conditions with low planning frequency.
• To decouple translational and rotational motion planning for computational efficiency and constraint satisfaction.
• To validate the system in a realistic simulation environment using ROS and Gazebo.

Proposed method

• Decomposed motion planning into translational and rotational components, using double integrator dynamics for translational trajectory planning via kinodynamic-RRT*.
• Employed a minimum-time trajectory generator for rotational motion, enforcing velocity and acceleration limits as hard constraints.
• Applied input weight in the RRT* cost function to softly enforce dynamic constraints on translational motion.
• Implemented a concurrent software architecture for online motion planning and trajectory tracking with separate loop rates (3 Hz for planning, 60 Hz for control).
• Used PI control for trajectory tracking, where control inputs are computed based on current pose error relative to the planned reference.
• Simplified robot and obstacle geometry to points and circles for efficient collision checking in the RRT* algorithm.

Experimental results

Research questions

• RQ1 Can kinodynamic-RRT* with double integrator dynamics generate collision-free, dynamically feasible trajectories for omnidirectional robots in dynamic environments with moving obstacles?
• RQ2 How effectively can online replanning maintain trajectory feasibility when obstacle motions change unpredictably?
• RQ3 To what extent does decoupling translational and rotational motion planning improve computational efficiency and constraint handling?
• RQ4 How well does PI-based trajectory tracking follow the planned reference trajectory under real-time constraints?
• RQ5 Can the system maintain performance with a low planning frequency (3 Hz) while the control loop runs at 60 Hz?

Key findings

• The motion planning system successfully generated collision-free trajectories in dynamic simulations with moving obstacles, even when obstacle paths were not precomputed.
• The trajectory tracking system closely followed the planned pose and velocity profiles, with realized trajectories closely matching planned ones as shown in Figure 8.
• The system demonstrated robustness in online replanning, adapting to changing obstacle positions within 1.4-second intervals as shown in Figure 9.
• The use of input weight in RRT* reduced solution cost and improved convergence, while hard constraints on rotational motion ensured feasibility.
• The decoupled approach enabled efficient computation, allowing real-time performance despite a low planning frequency of 3 Hz.
• The implementation was validated in a ROS-based Gazebo simulation, with video and code publicly available for reproducibility.