AUV control using geometric constraint-based reasoning

The geometric constraint approach to motion planning with moving obstacles for autonomous underwater vehicles (AUVs) is reviewed. The authors discuss how constraints for optimizing sensor operation, nonuniform obstacle avoidance and predictable nonlinear obstacle motion can be generated by the system. The process of generating and combining the following diverse constraint types is explained: sensor orientation; flow noise; self noise; and droop avoidance; nonuniform obstacle avoidance; and predictable nonlinear obstacle motion. This method allows motion constraint descriptions using any number of separate speed and direction components, and provides a flexible means of controlling their subsequent combination.<<ETX>>