We have multiple robotic arms and a range of sensors and cameras, including a state-of-the-art motion capture system. We aim to address challenges within robotics such as real-time motion planning, including collision avoidance and interaction with moving objects, as well as object tracking and pose estimation.
While robotic arms are widely used today for automation of simple and repetitive tasks, primarily in the manufacturing industry, there are many sectors that have significant innovation potential by applying robotics to automate tasks. The Robot Manipulator Lab, located in Trondheim, is a space where perception and motion planning methods needed for increased autonomy in various domains are researched and tested.
The use of robotic cranes and arms within the aquaculture and fish farming sector, as well as the port-and-shipping industry is investigated for performing autonomous and remote operations, for instance cargo handling, maintenance and inspection tasks. We are also investigating research questions within the manufacturing industry related to tracking and manipulation of moving objects. Within the health sector there is also large potential for increased adaptation of robotics and machine learning, for instance in robotized non-invasive procedures and examinations.
These domains and use cases have many common challenges, such as how to grasp or manipulate objects in motion, how to understand and interact autonomously with an environment that is changing over time and how to track objects in motion. At the SINTEF Digital Robot Manipulator Lab we are trying to answer these research challenges.
Photo: The SINTEF Digital Robot Manipulator Lab investigates motion-planning and perception algorithms for enabling autonomous operations in the aquaculture, manufacturing and health sectors.
One of the major research objectives of the lab is development of novel real-time motion planning methods in order to interact with objects on motion. We also have projects related to human motion analysis and replicating human motion using robotic arms in order to test products for our industry partners. Augmented reality and visual twins are also researched in the lab. These objectives also motivate building tools for automatic calibration of multi-sensor robotic systems.
In order to interact with the world around us, it is also necessary to perceive and understand it. In the Robot Manipulator Lab we are researching how to track and predict the motion of objects, be it hanging crane loads, machined parts or human tissue. Powerful filtering and fusion methods are needed to make sense of and utilize all the available data from multi-sensor robotic systems. We also hold competency in machine learning methods, and are looking at how robots may learn the dynamics of itself and the objects of interest in the environment around it to facilitate better motion planning in a dynamic world.
In the lab we have multiple 6- and 7-DOF robotic arms and a wide range of sensors, such as 2D and 3D cameras and lidars. We also have a motion capture system (HiPPo lab) used for validation and quick development of technology demonstrators.
We have the following equipment in our lab:
- UR10e and UR5 6-DOF robotic arms (with grippers)
- Franka Emika Panda 7-DOF robotic arm
- Qualisys motion capture system (HiPPo lab.)
RGB and RGB-D cameras
- 3D printers (for rapid testing of robot end effector designs and test setups)
Autonomous offshore operations
One of the research goals of the lab is to increase the use of autonomous and robotic operations in the aquaculture sector. While fish farming started in more sheltered coastal environments, there is an increasing push for establishing new fish farms in locations that are more exposed to the elements, because of the radical growth in the industry. This leads to increased health and safety risks for workers, which motivates the development of novel technology and knowledge for autonomous operations at exposed fish farms.
At the Robot Manipulator Lab, we are investigating how to use vessel-mounted robotic arms to perform autonomous operations at fish cage sites. We are testing motion compensation and motion planning methods for doing accurate work tasks, even in rough seas.
Photo: Testing vessel-mounted robotic arms on a motion hexapod platform which emulates realistic vessel motions. The robot autonomously compensates for the wave motions and draws on a whiteboard.