Motion & Path Planning

 

How a robot gets from A to B safely and efficiently in an alien environment is one of the defining problems of space robotics. Our group has developed a family of planners built on continuous mathematical frameworks—principally the Fast Marching Method (FMM) and the Ordered Upwind Method (OUM)—that treat the terrain as a cost field and compute globally optimal paths in a single wave propagation pass. Because these methods naturally encode anisotropy (slope direction matters differently going up vs. down), they outperform classical grid-search heuristics on realistic planetary terrain models. We extend these foundations to multi-layered planners for reconfigurable rovers that select the best locomotion mode at each waypoint, kinodynamic local planners that respect joint-level actuator constraints, and coupled rover–manipulator planners that coordinate mobility and arm motion for sample-fetch tasks. More recently, work under the ESA RAPID & FASTNAV programs has pushed toward high-speed, semi-autonomous traversal enabling decimeter-per-second rover speeds—a step change from the centimeter-per-second speeds of past missions.

KEY CAPABILITIES
Global path planning (FMM/OUM) Single-pass wave propagation over cost maps derived from DEMs; optimal w.r.t. energy, safety, or combined metrics.	Multi-locomotion-mode planning Two-layer global/local planner selects drive, wheel-walking, or crabbing mode at each grid cell to minimize power.
Kinodynamic local planner Warm-started LQR cascade generates feasible, constraint-satisfying joint trajectories for over-actuated platforms.	Rover–manipulator coupling Coupled path and motion planner maximizes arm workspace to safely reach science targets in restricted configurations.

 

IN THE LAB & IN THE FIELD

 

SELECTED PUBLICATIONS
01	Coupled Path and Motion Planning for a Rover–Manipulator System Pérez-del-Pulgar C.J., Romeo-Manrique P., Paz-Delgado G.J., Sánchez-Ibáñez J.R., Azkarate M. Space Robotics: State of the Art and Future Trends, Springer Nature (2023) [also ASTRA 2019]
02	Multi-stage Warm Started Optimal Motion Planning for Over-actuated Mobile Platforms Paz-Delgado G.J., Pérez-del-Pulgar C.J., Azkarate M., Kirchner F., García-Cerezo A. Intelligent Service Robotics, 16, 247–263, Springer (2023)
03	Combined path and motion planning for workspace restricted mobile manipulators in planetary exploration Paz-Delgado G.J., Sánchez-Ibáñez, R., Domínguez, R., Pérez-del-Pulgar C.J., Kirchner F., García-Cerezo A. IEEE Access, 11, 78152-78169 (2023)
04	Optimal path planning using a continuous anisotropic model for navigation on irregular terrains Sánchez-Ibáñez J.R., Pérez-del-Pulgar C.J., Serón, J., García-Cerezo A. Intelligent Service Robotics 16 (1), 19-32 (2023)
05	Path Planning for Autonomous Mobile Robots: A Review Sánchez-Ibáñez J.R., Pérez-del-Pulgar C.J., García-Cerezo A. Sensors, 21(23), 7898 (2021)
06	Dynamic Path Planning for Reconfigurable Rovers Using a Multi-Layered Grid Sánchez-Ibáñez J.R., Pérez-del-Pulgar C.J., Azkarate M., García-Cerezo A. Engineering Applications of Artificial Intelligence, 86, 32-42 (2019)
07	Multi-scale path planning for a planetary exploration vehicle with multiple locomotion modes JR Sánchez, M Azkarate, CJ Perez-del-Pulgar-Mancebo International Symposium on Artificial Intelligence, Robotics and Automation in Space (i-SAIRAS) (2018)

OPEN-SOURCE SOFTWARE & DATASETS
	Multi-staged Warm started Motion Planning (MWMP) C++ library Motion planning library that uses Sequential Linear Quadratic regulator (SLQ) in a Multi-staged Warm-Started manner to plan the movements of a mobile platform github.com/spaceuma/MWMP-Cpp  >
	Multi-staged Warm started Motion Planning (MWMP) MatLab library Motion planning library that uses Sequential Linear Quadratic regulator (SLQ) in a Multi-staged Warm-Started manner to plan the movements of a mobile platform github.com/spaceuma/MWMP-MatLab >
	CAMIS - Continuous Anisotropic Model for Inclined Surfaces  PYTHON package including CAMIS and biOUM path planner github.com/spaceuma/CAMIS_python >

 

SEE ALSO