Semi-automatic harvester crane navigation-interface

Completed: Winter 2015  Duration: 8 weeks  With: Trieuvy Luu and Ezgi Sabir  Collaboration: Skogstekniska klustret


Background

 

The history of forest industry goes long back in time in Sweden, as it is a cornerstone of the Swedish economy. In the past, there used to be an extreme physical workload as most of the tasks were done by hand. Today these tasks are being replaced with heavy forestry machinery that requires a high mental workload instead. 
 

Operating these complex forestry machines brings new complications with itself. It takes many years to prepare new operators to be ready for such complex machines. To operate the forestry machines for a longer period of time will intensely exhaust the mind of the operator.

 

 

Final concept


Research

The world is unpredictable

 
         Artwork by Michael Murdoch

        Artwork by Michael Murdoch

 

 

FROM FACTORIES TO FORESTS

As robots are moving away from factory floors into increasingly unstructured environments, the ability to cope with uncertainty is critical for building successful robots

 

ENVIRONMENTS - Physical worlds are unpredictable

SENSORS - Sensors are limited in what they can perceive

ROBOT - Actuation involves motors that can be unpredictable

COMPUTER MODELS - Models are inherently inaccurate 

 
 

 

"Before making self-sustaining smart forestry machines, we must design semi-autonomous systems where humans and technology work together."

- Daniel Ortiz Morales, PhD Umea University

 

Automation through collaboration

 

COLLABORATIVE CONTROL - WHEN ARE HUMANS NEEDED?

In particular, probabilistic approaches are typically more robust in the face of sensor limitations, sensor noise and environment dynamics. In semi-automation, humans can be seen as very advanced sensors that help to increase the overall robustness.

Bidirectional communication in the form of human–robot dialog can be used to exchange information of different types, such as commands, queries and responses

 

 

 

Research process video