Path Planning Algorithms For Autonomous Underwater Vehicles

Imagine a world where robots explore the mysterious depths of our oceans

Navigating narrow channels, dodging obstacles, and mapping uncharted underwater territories—this is the realm of Autonomous Underwater Vehicles (AUVs). In our latest research, we dive into their intelligent path planning strategies.

The full paper is available here (PDF) and published at Zenodo - MARESEC_2023 Conference.

Navigating the Underwater Maze

The ocean covers most of our planet, yet so much of it remains a mystery. AUVs are the unsung heroes of ocean exploration, tasked with gathering essential data for environmental monitoring, research, and even search-and-rescue missions. But here's the catch: underwater environments are incredibly complex. Unlike a simple road map, these environments come with unexpected obstacles—think rock formations, wreckage, or even dynamically changing conditions.

Our research tackles this challenge by investigating clever algorithms that can chart the best possible route for an AUV. The idea is to ensure these underwater explorers can make smart decisions and adjust their paths in complicated underwater landscapes.

The Brain Behind the Route

At the heart of our work lies the fascinating world of path planning algorithms. In simpler terms, these are the "road maps" the AUVs use to get from point A to point B without bumping into obstacles. We explored several algorithms before choosing one called D* (Dynamic A*). This algorithm is a bit like having a GPS that not only gives you the fastest route but can also recalculate if a road is suddenly blocked. Even though D* works off planned (offline) paths rather than instant, live updates, it still offers a lot of flexibility in underwater conditions where the layout is not always fully known beforehand.

But what happens when more than one robot is on the mission? Imagine a team of AUVs working together to survey a large area. In these scenarios, avoiding collisions becomes extra challenging. To solve this, we employed a technique called Conflict-Based Search (CBS). With CBS, each AUV can plan its path independently—and then, a coordinating "brain" steps in to fix any overlapping routes. This ensures that every vehicle can do its job without running into another.

Simulating the Real-World Underwater Adventure

Before sending our AUVs into the real ocean, we needed to test our ideas in a controlled environment. We used a powerful simulation tool called OMNeT++. This simulator allowed us to mimic real underwater conditions and watch how the AUVs moved along the routes calculated by our planning algorithms. By combining simulation with an innovative mobility model (think of it as a digital ocean floor), we could validate our approach and make sure that the routes not only looked good on paper but would work in real-life scenarios.

Real-Life Case Studies

To bring our ideas from theory to practice, we designed two exciting case studies:

Single AUV Mission

One of our missions was all about navigating a single AUV from a narrow river channel into open sea. Using the D* algorithm, the AUV was able to find the optimal route, carefully maneuvering through tight spaces without hitting any obstacles.

Multiple AUV Mission

In a more complex simulation, several AUVs were deployed to inspect underwater structures, such as wind turbine foundations. This mission was a perfect opportunity to test the CBS approach, ensuring that even when multiple vehicles are operating in the same area, they can avoid collisions and work together seamlessly.

Why It Matters

Our work doesn't just stop at theoretical research—it lays the groundwork for safer, more efficient underwater exploration. Whether it's for scientific research, environmental monitoring, or even industrial inspections, these intelligent path planning strategies can vastly improve the way AUVs perform their missions.