Exploiting Novel Representations for Constrained Multi-Robot Collaborative Loco Manipulation

The principle goal of the project is to address multi-robot loco-manipulation tasks; both from a decentralised planning as well as predictive control perspective. The aim is to realise human-robot teaming for handling rigid and/or actuated object.
Description of the Project: 

The aim of this project is to enable a team of mobile manipulators with the capability of dexterous simultaneous manipulation and locomotion.

Multi-robot system (MRS) promises alternative solutions to motion tasks that are beyond the single robot systems. MRSs capable of collaborative loco-manipulation can be extremely effective in today's real world application, e.g. transportation of large or heavy object, waste retrieval and disposal, or the tasks that are currently being handled by multiple human operators. With optimal and adaptive use of the properties of MRS, parallelism and decentralization, we can aim to generate a group dexterity and resilience that a single robot cannot achieve, irrespective of its sophistication and power.

Multi-robot collaborative loco-manipulation tasks always involve motion planning challenges in high-dimensional constrained configuration space.

This project will address this by decomposing the task space into several different sub-spaces to map on to the configuration’s corresponding subsets.  By doing this, we can use different representations and methods for the corresponding subsets of the configuration - -to make planning more reactive, scalable and complete.

This work will build on several developments in motion planning and control research. Several methods can be adopted for the global locomotion, such as search-based algorithms like A*, RRT, PRM. Moreover, the local locomotion can also draw ideas from formation control for robot coordination. We will specifically address challenges for multi-arm manipulation, where close-chain kinematic constraints can be encoded as a relation between end-effector pose constraints [3]. This idea of so-called decomposition has been explicitly and implicitly demonstrated in several works involving complex motion planning recently [1][2][4].  

Resources required: 
two single-armed mobile manipulators, one dual-armed mobile manipulator [hardware exists or is in the process of being acquired]
Project number: 
First Supervisor: 
University of Edinburgh
Second Supervisor(s): 
First supervisor university: 
University of Edinburgh
Essential skills and knowledge: 
C++, Python, ROS, Motion planning in High Dimensions
Desirable skills and knowledge: 
Physics simulators (Pybullet, Gazebo), hardware experience with mobile robot and manipulator
Funding Available: 

[1] Theodorous Stouraitis, Iordanis Chatzinikolaidis, Michael Gienger and Sethu Vijayakumar, Online Hybrid Motion Planning for Dyadic collaborative Manipulation via Bilevel Optimization, IEEE Transactions on Robotics (T-RO) (2020).

[2] Alonso-Mora, Javier, et al. “Multi-Robot Formation Control and Object Transport in Dynamic Environments via Constrained Optimization.” The International Journal of Robotics Research, vol. 36, no. 9, Aug. 2017, pp. 1000–1021

[3] Berenson, Dmitry, Siddhartha Srinivasa, and James Kuffner. "Task space regions: A framework for pose-constrained manipulation planning." The International Journal of Robotics Research 30.12 (2011): 1435-1460.

[4] McConachie, Dale, Mengyao Ruan, and Dmitry Berenson. "Interleaving planning and control for deformable object manipulation." Robotics Research. Springer, Cham, 2020. 1019-1036.