Den här sidan är utskriven från Högskolan i Halmstads webbplats (www.hh.se). Texten uppdaterades senast den 2016-11-24. Besök webbplatsen om du vill vara säker på att läsa den senaste versionen.
Smart lift trucks are safe trucks, and they can work side by side with humans in, for example, a warehouse. To be smart, the lift trucks need maps with information. Saeed Gholami Shahbandi recently presented his thesis work about how these so called semantic, or “rich”, maps are created.
The overall objective of the AIMS-project, which Saeed Gholami Shahbandis’ licentiate thesis is a part of, is to build a semantic map. With lots of information about the environment, the robot knows how to move around safer and more efficiently. That is a big step towards autonomous trucks or robots.
– In a workspace where trucks and humans operate side by side, it is important that the automation is reliable and therefore the trucks have to be aware of their surroundings. Semantic maps are one means of providing this awareness to the robot. Well-behaved auto guided lift trucks, robots, are essential for warehouse automation, says Saeed Gholami Shahbandi.
Today the technique allows robots to move around by following paths of preinstalled markers or sensors.
– What I have been trying to do is to automate the maping of the environment, says Saeed Gholami Shahbandi who sees a scenario where a truck can be sent into a warehouse and by itself create the map it needs to move around and work.
To make the auto guided lift truck smarter and to be able to ask it to survey the space and create its own map, there has to be means of developing these maps. Saeed Gholami Shahbandi’s research tackles the problem of environment modelling for the lift trucks. In order to do this, he uses information from different sensors such as laser scanners and cameras.
– We want to build this semantic map by fusing maps from different sensors. In my licentiate thesis I show how to have an abstract version of each map independent of sensor type, so that it enables us to fuse maps built from different sources.
For his thesis, Saeed Gholami Shahbandi and his research colleagues have been in warehouses and collected information to test the methods in real situations.
– We installed sensors on auto guided lift trucks and drive it around the environment, collecting data and creating maps from the data.
The robots also need to be able to detect if something in the environment changes, so that the map can change as well. This is also about safety:
– The better the map is, the more aware the lift truck is of its surroundings and the safer is its decision making.
Saeed Gholami Shahbandi is now working on his phd, were he develops this work and stacks the different maps together towards developing an enriched map.
– This is a very advanced field, relying and benefiting from a lot of different improvements in robotics as well as improving them or developing a specific method that can handle very challenging environments such as a warehouse.
Saeed Gholami Shahbandis points out that his research, and the AIMS project, regarding maps are using warehouses as an example.
– But the core technique is not specific to warehouses. It could be used for mobile robots walking in homes or in hospitals, for example. Another research project at the University concerns management and control of big automated trucks in harbours where many containers need to be sorted, says Saeed Gholami Shahbandi.
Text: KRISTINA RÖRSTRÖM
Saeed Gholami Shahbandi defended his licentiate thesis Semantic Mapping in
Warehouses on the 23rd of September. Read the thesis
Automatic Inventory and Mapping of Stock, AIMS
Researchers at Halmstad University have, in collaboration with three industrial partners, developed systems for automated guided vehicles so that they can become more intelligent, safe and flexible – which in turn will lead to a more efficient warehouse management. The research has been done in collaboration with Center for Applied Intelligent Systems Research (CAISR) at Halmstad University, Toyota Material Handling, Kollmorgen and Optronic. The project has led to continued cooperation between the involved parties in order to further develop and improve the automated guided vehicles.