Iara

Mechanical Subsystem

Electrical Subsystem

Software Subsystem

Iara represents not only the beginning of a journey, but also the overcoming of challenges. With a tight budget and still affected by strikes in federal universities, we are proud to develop a reliable AUV capable of performing RoboSub tasks. With a Brazilian DNA, Iara combines robustness and maneuverability.

Our AUV is the result of initial learning, conceived at the team's inception in December 2023. There is still much to improve, and we believe RoboSub 2024 will be a great opportunity to acquire knowledge and experiences. We are determined to bring innovations to our project and contribute to the development of others.

Structure

With extruded aluminum profiles and acrylic cylinders, Iara has a rigid and lightweight structure. With better pressure distribution, our AUV is hydrodynamic and safe.

Power Management

Since each component requires different voltages and currents, we need a Power Management Board. With Vin+,12v and 5V outputs, our board also contains an integrated kill switch for emergency situations.

Torpedo

The torpedo is launched by a solenoid valve, activated by a N-Channel MOSFET. Since the transistor functions as a normally open switch, the torpedo is launched when tha MOSFET's gate receives a HIGH signal (3.3V) from the Pixhawk.

Hydrophones

We built our own hydrophones using piezoelectric transducers and resin for waterproofing. To synchronize with the RoboSub pinger, we built amplifiers and band-pass filters in the hydrophones circuit.

Maneuverability

Iara is equipped with 8 thrusters strategically positioned to provide 6 degrees of freedom, enhancing their torque. This design ensures precise and agile maneuverability.

Watertightness

Using resin-sealed penetrators and waterproof connectors, we ensure the AUV's watertight integrity. We also conduct internal pressure tests to detect potential leaks.

SLAM Method

Iara executes SLAM with its stereo cameras for mapping. With the support of sensors, our AUV can create a map of the environment and localize itself within it.

Computer Vision

A neural network trained and implemented on the Jetson Nano assists in the AUV's decision-making. We applied Visual Servoing techniques, using images from the cameras to guide Iara's operations.

Simulation Environment

Based on Unity and ROS, we built a simulation environment including a rendered version of the pool and RoboSub tasks. We achieved a high-resolution visualization of the competition environment, allowing the computer vision algorithms tests.