Multi-limbed Robots empowered by whole-body Loco-manipulation – MeRLin
The MeRLin project addresses the challenge of replacing or assisting human workers in strenuous and hazardous tasks by developing advanced loco-manipulation skills in multi-limbed robotic platforms. These tasks, such as pushing/pulling heavy loads on uneven surfaces or operating hand-held power tools in confined spaces, pose health and safety risks to humans. To enable robots to perform these tasks effectively, the project focuses on enhancing their loco-manipulation skills by Whole-Body Control (WBC) approaches. Multi-limbed robots face complexities in planning and controlling their loco-manipulation skills, especially when dealing with redundancy in motions and contacts, simultaneous tasks, physical and hardware constraints, and non-linear dynamics. WBC approaches have emerged as a solution, but a comprehensive and robot-agnostic WBC approach for agile and heavy loco-manipulation tasks is lacking. The MeRLin project aims to fill this gap by developing a versatile WBC framework capable of handling loco-manipulation skills for tasks requiring high interaction forces and dynamic movements. The project's objectives include achieving long-term whole-body planning and short-term reactive control. This will be accomplished by integrating state-of-the-art model-based techniques from sample-based planning and optimal control, along with learning methodologies. The synergy between these approaches will be harnessed to maximize their strengths, resulting in a robot-agnostic framework solution. MeRLin seeks to create a seamless and efficient interaction between robots and their environment, ultimately improving safety and efficiency in various labor-intensive tasks.
Project coordination
Enrico Mingo Hoffman (Institut national de la recherche en informatique et automatique)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
Partnership
Institut national de la recherche en informatique et automatique
Help of the ANR 187,842 euros
Beginning and duration of the scientific project:
March 2025
- 42 Months