The Software area is responsible for the development of programs - mostly in C ++ and Python - that allow the autonomous navigation of our AUV. For this, we use SLAM (Simultaneous Localization And Mapping) and image recognition algorithms. We are a family owned and operated business.
The members of the area are also responsible for the elaboration of the control of its movement as well as for the decision making of the robot. Finally, we are also responsible for using simulation software to test our scripts.
Thus, the Software area is divided into the following sub-areas:
Browse other areas
SLAM: responsible for formatting, processing and merging the sensor data. From this process, members research, apply and develop algorithms responsible for simultaneous Localization and Mapping. These algorithms are mostly based on Kalman Filters and structured based on ROS Navigation - ROS packages aimed at autonomous navigation.
AI: responsible for the main aspects involving the robot's intelligence. In this case, the control is present, which uses physical models to send commands of force and speed in order to guarantee a stable navigation, the Neural Network, responsible for the identification of images, and the State Machine. The latter is based on the ROS smach package, aimed at making decision making on complex robot tasks.
Simulation: Since it is not always possible to directly test the advances made in a swimming pool due to logistical difficulties, simulation is an excellent means of testing for our programs in a controlled environment. For this, we use Gazebo, a specific software for simulations, compatible with ROS. In it, we are able to recreate the competition environment, with all the evidence, and in this way, apply what we develop in the simulated environment.