1st Edition
Trajectory Planning Using Dynamics and Power Models A Heuristics Based Approach
This book shows how to plan trajectories (i.e. time-dependent paths) for autonomous robots using a dynamic model within the A* framework.
Drawing from optimal control's model predictive control framework, the book develops a paradigm called Sampling Based Model Predictive Optimization (SBMPO), which generates graph trees through input sampling of a dynamic model, enabling A*-type algorithms to find optimal trajectories. The book covers various robotic platforms and tasks, including manipulators lifting heavy loads, mobile robots navigating steep hills, energy-efficient skid-steered movements, thermally informed space exploration planning, and climbing robots in obstacle-rich environments. It also explores methods for updating dynamic models for robust operation and provides sample code for applying SBMPO to additional problems.
This resource is aimed at researchers, engineers, and advanced students in motion planning and control for robotic and autonomous systems.
Chapter 1. Introduction
1.1 Emergence of Sampling Based Path Planning
1.2 Kinodynamic Motion Planning
1.3 Sampling Based Model Predictive Optimization
1.4 Book Organization
Chapter 2. Review of SBMPO
2.1 Propagation Model
2.2 Input-Sampling
2.3 Tree Graph
2.4 Sampling Based Model Predictive Control
2.5 SBMPO Algorithm
2.6 Algorithm Completeness
2.7 Illustration of SBMPO
2.8 Application of SBMPO to High Dimensional Systems with Dynamic Constraints
2.9 Effect of Grid Resolution and Sample Period
2.10 Baseline Comparison of SBMPO against RRT∗
Chapter 3. Development of Heuristics for Direct Generation of Trajectories
3.1 Heuristic for Time Optimality
3.2 Naive Heuristic for Distance Optimality
3.3 Velocity-Aware Heuristic for Distance Optimality
3.4 Heuristic for Energy Optimality
Chapter 4. Illustration of SBMPO Motion Planning Using Dynamic Models
4.1 Momentum Based Planning
4.2 Minimum Time Planning for Autonomous Spacecraft
4.3 Energy Efficient Planning for Skid-Steered Vehicles
4.4 Thermally Informed Motion Planning
Chapter 5. Learning Models and Heuristics: Current and Future Work
5.1 Learning Models
5.2 Learning Heuristics
Chapter 6. Contributions and Future Work
Appendix A: Code Repository
A.1 Unicycle Steering
A.2 Ackermann Steering
A.3 Double Integrator
Biography
Camilo Ordonez received a B.S. in Electronics Engineering from Pontificia Bolivariana University in 2003. He obtained his M.S and Ph.D. degrees in Mechanical Engineering from Florida State University in 2006 and 2010. Currently, he is a faculty member in the department of mechanical engineering at the FAMU-FSU College of Engineering. He is part of the Center for Intelligent Systems, Controls, and Robotics (CISCOR) and the Energy and Sustainability Center. His research interests include dynamic modeling of legged and wheeled vehicles, terrain identification, and motion planning.
Mario Harper is a professor of Computer Science at Utah State University and the director of the Decision-making, Intelligence, Robotics, Electrification, and Transportation (DIRECT) Lab. With expertise spanning Artificial Intelligence, Machine Learning, Robotics, and Finance, Dr. Harper has contributed to many projects involving satellites, Mars rovers, military systems, and electrified transportation. His research integrates AI with electrification, space robotics, and intelligent systems, with a focus on practical applications in extreme environments. He received a B.S. in Physics and Economics from Utah State University, as well as an M.S. in Finance and Computational Science and a Ph.D. in Computer Science, both from Florida State University.
Jonathan Tyler Boylan earned a Bachelor’s degree in Mechanical Engineering with a Minor in Computer Science from Florida State University in 2023. He is currently pursuing a Master’s degree in Robotics at Florida State University, where he conducts research in the Scansorial and Terrestrial Robotics and Integrated Design (STRIDe) Lab at the FAMU-FSU College of Engineering. His work focuses on advancing decision-making algorithms for autonomous robotic systems, including autonomous ground vehicles (AGVs), quadrupedal robots, and other platforms. His research interests span dynamic modeling, motion planning, computer vision, and robotic control, aiming to bridge theoretical insights with practical innovations in autonomous robotics.
Emmanuel Collins currently serves as Dean of the J.B. Speed School of Engineering at the University of Louisville. He has had a long career as a researcher in the fields of controls and robotics. Upon graduating with his Ph.D. in Aeronautics and Astronautics from Purdue University, he was employed at Harris Corporation where he worked in the emerging field of flexible space structure control. He made major contributions to the development and demonstration of effective robust vibration control algorithms, culminating in an Honorary Superior Accomplishment Award from NASA. As a professor, he has made contributions to a variety of areas, including robust control, robust fault detection, proprioceptive terrain classification for robots, intelligence for both mobile robot planning and manipulator motion planning, and nonlinear adaptive control. He has always focused on developing or interpreting state-of-the-art optimization algorithms and applying them to real-world problems. This remains one of his passions.
