Currently, I am working on the design and development of algorithms for autonomous systems in virtual reality environments, such as Gazebo, Matlab Virtual Reality (VR) Toolbox, by using machine learning techniques and properly operating system, i.e., ROS (Robotic Operating System) for the analysis and the control of small Unmanned Aerial Vehicles (UAVs) systems (aka MAV, Micro Aerial Vehicle).
The control of small UAVs, such as Parrot AR.Drone 2.0 and Bebop 2.0 (micro-quadcopter) and Bitcraze Crazyflie 2.0 (nano-quadcopter), has been extensively investigated worldwide for functionalities augmentation and costs reduction with respect to a single larger vehicle. Machine learning techniques (e.g., reinforcement and deep learning) are often used to reduce the time-consuming process of programming the UAV desired behavior, an activity that requires a large amount of time by an interdisciplinary team. In this way, programmers can specify what action the robot should perform without ever detailing how it should perform the action. This abstraction is especially beneficial when controlling a robot with multiple actuators, such as aerial robots, moving through a dynamic environment.
The aim of this research is to propose a co-design approach based on simulation-in-the-loop and software-in-the-loop methodologies for performance evaluation, designing and developing algorithms for autonomous systems. The simulation scheme consists of specific middleware, namely ROS, where one node or computer simulates the control unit (MATLAB/Simulink equipped with Robotics System Toolbox), and another one the vehicles in the scenario (sensor models and obstacles are simulated, too), through virtual reality environments (e.g., Gazebo, V-REP), with the aim of implementing the overall control system together with the flight dynamics and the environment.
If you'd like to know more about me, please stay tuned on this web page, or feel free to contact me. Also, check out my Publications and Software pages for my latest publications and repositories, respectively.