Minimum-Time Trajectory Planning for a Differential Drive Mobile Robot Considering Non-slipping Constraints
- PDF / 1,080,707 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 84 Downloads / 182 Views
Minimum-Time Trajectory Planning for a Differential Drive Mobile Robot Considering Non-slipping Constraints I. F. Okuyama1
· Marcos R. O. A. Maximo1
· Rubens J. M. Afonso2,3
Received: 14 February 2020 / Revised: 24 August 2020 / Accepted: 8 October 2020 © Brazilian Society for Automatics–SBA 2020
Abstract We propose a real-time minimum-time trajectory planning strategy with obstacle avoidance for a differential-drive mobile robot in the context of robot soccer. The method considers constraints important to maximize the system’s performance, such as the actuator limits and non-slipping conditions. We also present a novel friction model that regards the imbalance of normal forces on the wheels due to the acceleration of the robot. Theoretical guarantees on how to obtain a minimum-time velocity profile on a predetermined parametrized curve considering the modeled constraints are also presented. Then, we introduce a nonlinear, non-convex, local optimization using a version of the Resilient Propagation algorithm that minimizes the time of the curve while avoiding obstacles and respecting system constraints. Finally, employing a new proposed benchmark, we verified that the presented strategy allows the robot to traverse a cluttered field (with dimensions of 1.5 m × 1.3 m) in 2.8 s in 95% of the cases, while the optimization success rate was 85%. We also demonstrated the possibility of running the optimization in real-time, since it takes less than 13.8 ms in 95% of the cases. Keywords Trajectory planning · Optimization · Differential-drive mobile robot · Robotics
1 Introduction Differential-drive mobile robot (DDMR) is a very pervasive class of robots. They are car-like and have only two wheels driven by electric motors, making them cheap and easy to manufacture. However, this hardware simplicity comes with nonlinear, nonholonomic, and underactuated dynamics (LaValle 2006); hence trajectory planning and control are hard, especially in competitive scenarios, such as military
B
I. F. Okuyama [email protected] Marcos R. O. A. Maximo [email protected] Rubens J. M. Afonso [email protected]
1
Autonomous Computational Systems Lab (LAB-SCA), Computer Science Division, Aeronautics Institute of Technology, São José dos Campos, São Paulo, Brazil
2
Department of Aerospace and Geodesy, Institute of Flight System Dynamics, Technical University of Munich, Munich, Bavaria, Germany
3
Electronic Engineering Division, Aeronautics Institute of Technology, São José dos Campos, São Paulo, Brazil
applications and robot soccer, where high performance is needed to beat an opponent. The Very Small Size Competition League (VSS) is a robot soccer competition where two teams compete using three autonomous mobile robots each one (IEEE 2008). Each robot is constrained to fit within a cube of side 7.5 cm. The poses of the robots and the ball are determined by a vision system based on an overhead camera. Moreover, computationally intensive algorithms, including computer vision, decisionmaking, and motion planning, run in a central computer.
Data Loading...
