Målet med dette prosjektet har vært å utvikle metoder og algoritmer som lar industriroboter interagere med og håndtere objekter som er i bevegelse i sanntid. Prosjektet har hatt to fokusområder: produksjon og helsetjenester. Produksjonscasen fokuserte på utviklingen av en demonstrator for å plukke opp og plassere ulike gjenstander på hengere som beveger seg på skinner i taket ved hjelp av en robotarm. Roboten kompenserer for pendelbevegelsen til hengerne i sanntid og plasserer gjenstandene sømløst på hengerne. Helsecasen tok sikte på å utvikle en demonstrator for autonome ultralydundersøkelser av hovedpulsåren i magen. Her er roboten utstyrt med både en ultralydprobe og et 3D-kamera. Først har vi implementert en strålesporingsalgoritme for å estimere banen som roboten skal følge basert på et anatomisk atlas. Roboten bruker så 3D-kameraet i kombinasjon med en metode for pasient-til-atlas-registrering for å finne utgangsposisjonen på magen til pasienten og starte undersøkelsen. Til slutt blir ultralydbildene automatisk analysert i sanntid for å finne hovedpulsåren, og en adaptiv baneplanlegger tilpasser banen til roboten tilsvarende.
The impact of the results of this project have been extended in both the manufacturing and the health industry.
As far as the manufacturing industry is concerned, the results of the project and espesially the compensation of the swinging motion of the hangers in real time gives the possibility for a more secure and effective production line. Objects in hanging conveyors can be treated more effectively and with greater acurracy being able to predict their future motions. In addition, methods for prediction of motion and collision avoidance, also have a direct effect the manufacturing industry for interaction with moving objects.
As far as the health industry is concerned, the project results have a direct effect in the help of modernization of the health industry in hospitlas giving equal opportunities for health ultrasound examinations, both for all people no matter where they live. Furthermore, physicians can gain advantages on doing ultrasound examinations with the help of robotic arms, since they can have new measurements opportunities and improved image quality of the scanned area.
The next two big steps in IOT towards 2025 is ubiquitous positioning for context specific data and services, and teleoperation - the ability to monitor and control distant objects.
The industrial need for tracking and manipulating moving objects in real-time has been identified through industrial and research projects. In this project, we will develop methods that allow the interaction and manipulation of moving objects in real-time. The motion of the objects is constrained, but their position, orientation and velocity are only partly known in advance and therefore a real time estimation is needed. The objects that we are going to act upon may be rigid or soft, and the force used in the manipulation of the object has to be measured and controlled. Lastly, the robot has to be able to do complex operations on the object whilst the object is still in motion thus requiring a complex path planning and path adjustments in real time.
Painting of objects or welding of small parts on moving objects are examples from the manufacturing industry where being able to manipulate moving objects will be very useful. Inspection and maintenance in the process industry, either tele operated or by an autonomous mobile robot, requires close contact measurement on moving (vibrating) objects and has similar challenges.
In healthcare, assistive robotics require safe interactions with flexible objects (humans). Automation of interventional procedures necessitates research on real-time methods for detection and compensation of motion, in order to safely interact with patients and health personnel.
The project group consists of four young researcher from medical imaging, estimation and robotic control with a good track academic track record and experience in working on industrial applications.