LSSU Team AMORE - ASV (RoboBoat)

Our team needs your support

Support Team AMORE

ASV:

The Sardaine

Autonomous Surface Vehicle (ASV)

Team AMORE's 2025 RoboBoat Vessel

ASV History (2023-2025)

RoboBoat 2023- LSSU

The first iteration of the ASV featured a children's kayak with a wooden deck. Extra floatation was added to improve the stability of the vessel. The propulsion system consisted of two ~90N (20lb) thrust Hawk Hobby thrusters. These thrusters were mounted at the middle of the hull. The control system of the vehicle used the Guidance, Navigation, and Control (GNC) system for our WAM-V from the 2022 RobotX competition. The vision box and GNC battery box from the WAM-V were also integrated on the ASV. The only box dedicated to the kayak was the electrical box. The electrical box consisted of the motor controllers for the thrusters, the batteries for the thrusters, and the safety system (E-stop).

RoboBoat 2024- LSSU/FAU

The second iteration of the ASV was a collaboration between Florida Atlantic University (FAU) and Lake Superior State University (LSSU). The hull and electrical system were developed by FAU, while the GNC and vision system were developed by LSSU. This is the first iteration of the ASV to feature a dedicated GNC system rather than re-using the GNC system on the WAM-V. The hull consisted of two pontoons made of foam with a fiberglass outer layer coated in marine epoxy. The vision system for the ASV used the same vision system meant for the WAM-V. The vision system consisted of an interface box and the perception mast. The perception mast consisted of a Velodyne LiDAR and a Zed2i stereo camera. This vessel helped FAU and LSSU place 5th in the autonomy challenge portion of the competition.

RoboBoat 2025- LSSU

The most recent iteration of the ASV features characterisitics from both previous iterations. The current ASV features the same hull, thrusters, and electrical system as the first iteration and the GNC system from the second iteration. This year's design features an improved vision system, with a more powerful processor and a dedicated vision box. The GNC system features an upgraded processor. This iteration of the ASV has been utilized for ecological research such as Zooplankton monitoring. The ASV will feature a ball and water delivery system during the 2025 RoboBoat competition.

ASV System Overiew

Overview

The ASV is a 6-foot kayak with a raised, stable platform and a payload of 130 pounds. The kayak is made of durable high-density polyethylene, which is lightweight and has a high strength-to-density ratio. This platform is designed for children’s play, but can handle many of the surface conditions found on the Great Lakes, which is where AMORE conducts testing. This led to the ASV’s use being expanded beyond the competition to also include biological research.

On the base hull, AMORE has added a variety of systems for unmanned operation. The first being the GNC system, which serves as the brain of the vessel. The perception system, also known as the vision system, enables the ASV to make decisions based on its surroundings. The communication system allows monitoring of the ASV from shore. There is an integrated safety system consisting of indicators and safety switches. Finally, there is a modular racquetball launcher developed specially for Roboboat. A model of the design of the ASV for RoboBoat 2025 is shown in the figure on the left.

Propulsion System

Main Components:

2x HAWK HOBBY Brushless Underwater Thruster
- 24 V, 1080 W
- 90 N (20 lbs) of thrust

Overview:

The ASV's propulsion system, first developed for RoboBoat 2023, utilizes two bidirectional thrusters mounted on opposite sides of the hull. The thrusters were placed towards the center of the ASV to simplify the control software and to enable the ASV to rotate efficiently. With the current propulsion system configuration, it was determined that the ASV would utilize differential thrust, where each thruster uses varying thrust outputs to reach a desired state. The thrusters used for the ASV were Hawk Hobby 24V bidirectional thrusters; these thrusters are capable of 90 N of force each, providing sufficient thrust for the ASV. The figure below shows the thruster used for the ASV.

In order for the ASV to operate autonomously, a low-level controller is required to dictate the behavior of the vessel while operating in autonomous mode. There are two separate Proportional Integral Derivative (PID) controllers that fill this role. During waypoint navigation, a Heading and Speed controller utilizes the thrusters with differential thrust, providing sufficient speed and maneuverability. For Station keeping, the thrusters operate similar to the Heading and Speed controller, using differential thrust to maintain position and heading. The free-body diagram of the forces from the thrusters is shown in the figure below.

Sensors

Main Components:

Velodyne VLP-16 LiDAR (Donated by Robonation)
ZED2i Stereo Camera
HolyBro M9N Global Positioning System (GPS)
IST8310 - 3D Magnetic Sensor (Compass)
ICM-42688-P (6-axis IMU)

Overview:

The ASV requires an array of sensors to enable autonomous navigation, consisting of both navigation and perception sensors. The navigation sensors provide information on the ASV's global position, velocity, orientation, and direction of travel. These sensors are the GPS, compass, and IMU. For the ASV, the navigation sensors are implemented using a Pixhawk 6C flight controller. The flight controller houses every navigation sensor (GPS, IMU, Compass) required for autonomous navigation in a compact device. The team's design philosophy is space optimization, so a low-profile device that is reliable is advantageous for the ASV. The Pixhawk 6C flight controller is shown in the figure below.

The perception sensor array consists of two devices, the Velodyne LiDAR and the ZED2i stereo camera. In basic terms, the LiDAR is used to tell where objects are and the stereo camera tells you what the object is. The LiDAR is used to detect and localize data within an area in front of the vessel within a ± 45 degree angle of the vessel coordinate frame and a certain distance. An occupancy grid of the course is generated to map the course. The basic idea of the occupancy grid is to represent a map of the environment as an evenly spaced field of binary random variables each representing the presence of an obstacle at that location in the environment. This map is an input to the path planner, used to calculate the necessary trajectories and avoid obstacles.

Once the objects in front of the ASV are mapped, then the ZED2i stereo camera classifies them. Using a combination of OpenCV and convolutional neural networks the system can determine the size, shape, and even pattern of each detected object providing task data directly to our control system. Here the path planner gains information about navigating buoys based on rules defined based on particular missions, such as following general buoy rules or identifying buoy pairs for competition tasks. The perception sensor array is shown in the figure below.

Perception System

Main Components:

Jetson Orin AGX (Vision Processor)
Velodyne LiDAR interface box
Ethernet Hub

Overview:

The ASV's perception system processes the data received from the LiDAR and stereo camera and turns it into data that can be used by the autonomy processor in the GNC system. The perception system uses a Jetson Orin AGX as its main vision processor, which replaces the Jetson TX2 that was previously used. The Jetson Orin AGX runs several algorithms such as object detection and the trajectory generator. The object detection algorithm needs to be trained through the collection of images of specific objects and identifying them to develop a baseline model. The trajectory generator uses the LiDAR to determine where objects, such as buoys, are and uses the stereo camera to tell what each detected object is. With both the LiDAR and stereo camera data, the trajectory generator is able to generate waypoints based on a set of rules determined by the team. If the ASV is navigating a channel with red and green buoys, the trajectory generator will output waypoints bisecting the closest set of buoys and while continue to do so until the ASV has navigated the entire buoy channel. The perception system allows the vehicle to navigate a course autonomously without the need for teaching waypoints. The perception box, also known as the vision box, is shown in the figure below.

Perception System Video Overview:

Guidance, Navigation, and Control (GNC) System

Main Components:

Jetson Orin Nano
Pixhawk 6C Flight Controller
R12DS RC Receiver
Ethernet Hub
Battleswitch Relay (Remote E-stop)
Teensy 4.1 Microcontroller

Overview:

The Guidance, Navigation, and Control (GNC) box is located aboard the deck and is responsible for all communication aboard the ASV. The GNC box houses the Jetson Orin Nano computer, a Teensy 4.1 microcontroller, a Pixhawk 6C flight controller, and the integrated safety system. The main function of the GNC box is to act as a management device for all sensors and components aboard the ASV. This is achieved using Robotic Operating System (ROS) as middleware to communicate between various sensors and components in software. Using ROS, the low-level controllers aboard the ASV have access to vital peripheral data, such as the current GPS location, that may be used to dynamically adjust the propulsion system’s output to reach a desired state and respond to changing environmental conditions. This setup allows the ASV to achieve accurate and reliable movement, essential for tasks such as navigating complex waterways, performing scientific data collection, and conducting autonomous operations. A diagram of the GNC box configuration is shown in the figure below.