Autonomous Systems Research Theme
This theme has the strategic aim of developing innovative scientific methods that lead to technologies requiring minimal human intervention. Research partnerships have been established with BAE Systems, National Nuclear Laboratory, North West Aerospace Alliance, Rolls-Royce, and Roke Manor Research. Achievements include intelligent agent-based energy-management and autonomous mission re-planning, developed through participation in the Integrated Electrical Power Networks Evaluation Facility (IEPNEF), and autonomous robots being developed for nuclear decommissioning applications through the £20m Dalton Cumbrian Facility, which is now a National Nuclear User Facility.
Autonomous systems are used for a variety of purposes including performing remote tasks in hazardous environments and remote sensing. Autonomous systems need to interact with an unknown, unstructured environment and need to sense and decide upon unexpected events that they encounter while performing their planned tasks. Autonomous Systems typically manifest themselves as mobile robotic platforms which operate in a wide range of domains.
AMORR@Manchester - our integrated, easy and flexible solution for implementing reliable autonomy
Reliable autonomy is our concept for a robot to be autonomous despite several sources of uncertainties and faults. The reliable autonomous solution, developed entirety by the Autonomous Systems Research Theme at The University of Manchester, AMORR@Manchester (Autonomous MObile Reliable Robots) comprises several algorithms for autonomy with fault tolerant and predictive maintenance capabilities. AMORR@Manchester can be used with several types of robotic platforms, such as ground, aerial, and water robots.
Demos at The UoM
We are pioneers in using autonomous robots at The University of Manchester. Autonomous robots equipped with software developed in our labs were successfully tested for the first time in 2014. The autonomous robots were integrally programed in our labs using AMORR@Manchester and are able to perform probabilistic and guaranteed algorithms for: Simultaneous Localisation and Mapping (SLAM), dynamic path-planning, navigation with fault tolerant capabilities, etc. In the following movies, robots are shown moving around in our labs at The University of Manchester.
Flexible Robotic Platform for Teaching
This video is showing a demo with a flexible educational robotic platform. In the demo, the robot is exploring an unknown environment. Several algorithms are needed to perform the required task:
Collaboration with MBDA Systems
Set membership approaches were implemented in AMORR@Manchester as an alternative to more traditional probabilistic methods. In critical applications, such as in military applications, the need for guaranteed solutions is crucial. During the project with MBDA Systems we developed a new guaranteed SLAM technique, interval SLAM (i-SLAM). This approach can work with nonlinear, multi-valued robot motion and observation models. Moreover, interval methods are guaranteed approaches, as they do not use sampling techniques; therefore, all possible solutions are obtained. AMORR@Manchester with i-SLAM was successfully demonstrated for MBDA Systems in 2016.
Collaboration with Sellafield Ltd
The Autonomous Systems research theme has implemented the AMORR@Manchester robotic solution on robots used by Sellafield ltd. This was a great success for our team since the robots used by Sellafield ltd were fully teleoperated before AMORR@Manchester enabled them to operate autonomously for the first time.
The autonomous capabilities were transferred, tested and demonstrated at Sellafield in December 2017. This work with Sellafield has demonstrated the efficiency of our consultancy strategy. Experts from Sellafield informed us of their needs in terms of autonomous robots for inspection, they supplied their teleoperated robots and their specific sensors and we in turn gave them the autonomous integrated solution which solved their problem. This pioneering work in autonomous robots for Sellafield was a great success for The University of Manchester.
Mechanical and electronic parts of the robots were developed by other groups working closely with us at the University of Manchester.
In the following movies the robots developed for Sellafield are tested in labs provided by the company.
Robotics Challenge in Australia
Participation at the International Atomic Energy Agency (IAEA) Robotics Challenge in Australia, November 2017. In collaboration with colleagues across the University of Manchester, the Autonomous Systems Research Theme implemented autonomous capabilities for the robots developed by other research groups (mobile robot and floating robot). The floating robot was selected for a field trial during the next stage of the evaluation process and effectively demonstrates that not only companies, but also other research groups at the university benefit from our consultancy strategy.
Unmanned Aerial Vehicles (UAVs)
Unmanned Aerial Vehicles are basically aircrafts which can operate without a human pilot aboard. In the last decade the applications involving UAVs increased significantly. They are used in civil and military applications, such as: autonomous transportation, surveillance for agricultural crops, surveillance of large forest and jungles to find possible sources of fire or unauthorised hunting, etc.
Unmanned Underwater Vehicles (UUVs)
UUVs are vehicles which operate underwater with minimal or no interaction with a base ship or with the shore. Applications include undersea exploration, the detection of pollution in rivers, exploration of ice-shelves in Antartica and the mapping of large scale, liquid-based industrial processes.
Unmanned Ground Vehicles (UGVs)
UGVs operate on the ground and drive without operator assistance. Such ground vehicles can interact with the environment and manipulate objects as they can exert considerable forces on the environment. Burrowing vehicles, which support the investigation of subsurface phenomena, can also be considered a sub-set of UGVs.
Common issues across all of the above categories include propulsion, control and sensing, supporting hardware and software designs, autonomy algorithms and cognitive processes. Higher level strategies for surveillance, mapping and exploration are also important, as are techniques that combine data from a variety of sensors to enable a detailed description of an environment or phenomenon to be constructed. This research theme is about creating supporting technologies (AMORR@Manchester) for the above types of vehicles.
Researchers in the Autonomous Systems Theme develop autonomous systems for real-world industrial applications. We work with unmanned vehicles that operate in all the above three domains. We explore new sensing technologies, novel vehicle platforms, new control strategies, new cognitive algorithms, power management and optimisation, and methods and tools for perception, abstraction, path-planning and decision making.