Objectives: To develop an underwater robotic system that autonomously performs a full-coverage scan of aquaculture infrastructure such as a fin-fish cage or seaweed/shellfish lines. A low-cost, autonomy-capable ROV equipped with multibeam sonar and camera will be augmented with custom-developed smart software and hardware to achieve three sub-objectives: 1. To develop a novel acoustic localization system termed LBL-iUSBL that gives the vehicle knowledge of both position and orientation; 2. To develop an automated full-coverage path planning algorithm to map the infrastructure fully; 3. To develop a SLAM approach that cohesively fuses all sensor data to produce a dense/detailed model of the infrastructure.
Methodology: The system will use a BlueROV2 Heavy, integrate an Oculus M750d multibeam and low-light HD camera, and made autonomous using ROS. LBL-iUSBL localization uses our previous work in OWTT iUSBL with similar hardware/software, using three beacons and a vehicle-mounted array. Path-planning algorithms will be developed on prior work by Hover, Galceran, Palomeras, and will output paths for vehicle following with embedded stand-off and pose set-points for full coverage. The SLAM algorithm uses our prior work in underwater mapping to consider uncertainties of all sensed data and perform cohesive data fusion to produce dense and detailed models of the aquaculture infrastructure.
Rationale: Offshore aquaculture represents a potentially resource-efficient means of producing protein-rich seafood. However, the environment imposes physical, chemical and biological processes leading to structural degradation and biofouling, requiring divers for inspection – a significant ongoing cost for farm operators. Lack of consistent farm monitoring can lead to financial disaster from stock losses, can cause danger to community and environment, and may lead to costly regulations. The remote nature of offshore farms means that safe and effective methods for monitoring represents a significant logistical and technical challenge. Automated inspection can limit risks and costs by limiting diver deployment to intervention and repair.