PhD Studentship: Developing a Multi-Robot Task Planner for Robotic Disassembly of Electrical-Vehicle Battery Disassembly in Industrial Scale

This project aims to develop a hybrid (logical and geometric) task coordination planner for multi-robot where each robot has an specific task; and has to collaborate jointly with other robots. The developed method focuses on disassembly of electrical vehicle Lithium-ion battery (LIB) battery packs.

To efficiently generate feasible and optimized task execution plans for a group of robots to disassemble a lithium-ion battery, a hierarchical multi-robot temporal task planning framework is required, in which a central server allocates the collaborative tasks to the robots, and then individual robots can independently synthesize their task execution plans in a decentralized manner. The robots solely or jointly, autonomously or semi-autonomously, will carry out a variety of manipulation action primitives (e.g. unscrewing, cutting, grasping, unplugging, separating etc.) for both robotic disassembly and robotic testing of EV battery.

The work exploits the Behaviour Trees model for task execution and monitoring, which links different robot capabilities such as object tracking and motion planning in a modular fashion. A behaviour tree is a graphically represented mathematical model of plan execution that describes switching between a finite set of tasks in a modular fashion. Action primitives should be capable of being applied, via perception-based online planning, to a wide variety of object categories and object examples, which are present in arbitrary positions, with respect to arbitrary structures of obstacles and clutter. During action executions, robots will continuously update their status to complete the scheduled tasks without collision. Further complexities and extensions may involve bi-manual tasks (e.g. one arm holds an object while other unscrews or cuts it), or scheduling and synchronising simultaneous actions of multiple arms. Each such action primitive will exploit state-of-the-art methods in computer vision and sensory guided autonomous motion planning and control. Robots must adapt these actions to uncertainty and diversity in the objects being handled. The PhD research problem is how to automatically plan a sequence of such actions to disassemble a complex object.

Student will be working in a team of mutually supportive researchers, who together are creating a portfolio of robotic disassembly capabilities of various types, with research spanning: autonomous motion planning; dynamics and control of forceful actions; advanced human-robot teleoperative interfaces; machine learning, computer vision and multi-sensor fusion.

Entry requirements

• The successful applicants should have a first-class degree in engineering and preferably an MSc in robotics or a closely related subject.
• a solid background in linear algebra, model parameterisation and optimisation, kinematics and dynamics.
• excellent programming skills (C++, Python, Matlab),
• good communication skills (oral and written).

Environment and support

Lab Tour: https://my.matterport.com/show/?m=du6jbWMzRk4

The University of Birmingham Extreme Robotics Lab, is one of the largest and best equipped robotics labs in UK. Facilities include a 1,000m2 lab on the main university campus, coupled with a full scale heavy duty industrial robot test-bed off campus at Birmingham Energy Innovation Centre.

How to apply

1- Put your CV, 1-page cover letter on a safe cloud (e.g. Onedrive, Dropbox etc.) and get the link.
2- Send an email and share the link with Dr Alireza Rastegarpaah (Email: a.rastegarpanah@bham.ac.uk)
** Please do not attach your files to your email; the inbox is almost full**

Informal enquiries should be addressed to Dr Alireza Rastegarpanah via email: a.rastegarpanah@bham.ac.uk

Advert information