Subscribe for email alerts

Don’t miss a single Science & Enterprise post. Sign up for our daily email alerts.

Please share Science & Enterprise

Algorithms Plot Optimal Autonomous Underwater Vehicle Routes

Math Formulas (João Trindade/Flickr)Engineers at Massachusetts Institute of Technology have devised mathematical methods to plot the optimal routes for autonomous underwater vehicles (AUVs), increasingly used for industrial, research, and security applications. The team led by mechanical engineering professor Pierre Lermusiaux will discuss its work in May 2012 at the IEEE International Conference on Robotics and Automation.

Setting an underwater course from one point to another does not always follow a straight line. Currents, physical barriers, battery limits, and security factors (e.g., piracy) can force planners to change the proposed route of an AUV. Add in the needs of planning the routes of packs or swarms of AUVs, and the complexity of the problem quickly expands.

Self-propelled and gliding AUVs are used today for mapping and oceanographic research, military reconnaissance, harbor protection, deep-sea oil-well maintenance, and emergency response. Fleets of up to 20 AUVs have been deployed for these kinds of tasks, but Lermusiaux expects the size of AUV fleets to grow as their needs expand.

Earlier attempts to capture these factors in a useful mathematical model foundered on the challenges of ocean currents and complex topography, compounded by the needs of exorbitant computational power that rendered many proposed solutions infeasible for real-time use. “What was missing,” says Lermusiaux, “was the integration of ocean prediction, ocean estimation, control, and optimization” for planning the routes of multiple vehicles in a constantly changing situation.

Among the factors considered in the MIT team’s model was the need for the AUV to ride on ocean currents rather than expend its own power, thus expanding the vehicle’s range. The researchers found that in some cases, it may be quicker to ride on the current, even if deflected from the most direct route, than  trying to fight the current in a straight line. The same idea applies to finding the most energy-efficient route, which require navigating over, under, or around currents and eddies in the ocean rather than sticking to a more direct route.

The algorithm allows, as well, for real-time control and adjustments. This feature enables the planning of routes to track a plume of pollution, such as an oil spill, to its source, or determine how it is spreading. In addition, the model can plan for packs of research AUVs to collect the most useful data in the fastest time.

Lermusiaux has successfully tested the model in simulations of the Philippines, which have many islands and convoluted shorelines, with a virtual fleet of 1,000 AUVs, deployed from one or more ships and seeking different targets. The simulation included areas where sea traffic was forbidden, as well as fixed physical obstacles and dynamic obstacles, such as passing ships.

Lermusiaux says that a similar approach could build computational systems to guide automated vehicles through any kind of obstacles and flows, such as aerial vehicles coping with winds and mountains. Another potential application, he adds, is the navigation of miniature medical robots through the circulatory system.

In the following video, Lermusiaux tells more about his work with planning routes for AUVs.

 

Read more: Institute Developing Autonomous Underwater Robots

Photo: João Trindade/Flickr

*     *     *

2 comments to Algorithms Plot Optimal Autonomous Underwater Vehicle Routes