Automation of Aquaculture with Underwater Drones (ATAQUD)

Purpose and goal:

The challenges facing this project were to develop methodology for autonomous underwater vehicles (so-called AUVs) to be able to perform more robust operations in real environments. Within the project, data from realistic environments has given us valuable information to train on. We have developed and presented a system and methods for the operation of an AUV in an algae farm, including localization and navigation based on a priori knowledge of the algae farm as well as dead-reckoning and sensor-based relative navigation. The results are demonstrated in a simulated algae farm.

Expected results and effects:

The results generated are algorithms for navigating with an AUV in an algae farm based on sonar data from the side-scan sensor with realistic data from a real algae farm and disturbances to the sensor data from both the vehicle itself and from the surroundings as well as expected variations in rigging and buoys for the farm. The results generated are expected to have effects on how future robots can be developed to be effective tools in growing the blue economy in Sweden (and globally) within algae farming, but also with wider areas of application within socially critical functions.

Approach and implementation:

The project has been carried out by senior researchers, PhD students and master students at KTH. The challenges in a project as short as this (apart from the concrete research questions), lie in staffing as there is no space to recruit for such a short project. Implementation therefore becomes completely dependent on the resources available within the organization right from the start.