The research area is related to Unmanned Aerial Vehicles, Ground Robots and Space systems. The main research topics are real-time guidance and control algorithms for advanced autonomous vehicles and robots, Guidance, Navigation and Control (GNC) algorithms for space proximity operations, design of robust and variable structure controllers for autonomous systems. Robotic systems are the mainly platforms considered in the research activities, focusing on distributed navigation and control algorithms. Implementation issues and real tests are main features of the research.
- Robust On-board Algorithms for distributed and cooperative robotic systems
Complex robotic systems and missions in unstructured and unknown environments lead to challenging nonlinear robust and generalized motion planning, which remains a critical point at the state of the art. Moreover, the decentralization of the mission is crucial to achieve scalable strategies for controlling any number of robotic systems.
The key features of this research are the combination of distributed/decentralized algorithms and the cooperation between robots. A multi-platform approach is considered, to enable robots, both aerial (UAV) and ground-based (UGV), to efficiently plan their path (path planning and trajectory planning), to flexibly manage their activities (task management) and to optimally control their dynamics (optimal control).
The project requires coordinating the actions of multiple autonomous agents that must operate in dynamic and uncertain environments and the focus shall be on:
- Design of control techniques, based on robust approaches, which consider system uncertainties, unmodelled dynamics and disturbances from the external environment.
- A distributed framework allowing to achieve a global agreement among the members of the swarm exploiting only local communication.
- Increase of flexibility and efficiency of the robotic systems, which must operate in complex and uncertain scenarios.
- Adaptive Sensor Management and Distributed Data Fusion Algorithms
Despite the significant results achieved in terms of accuracy in real-time implementation of on-board algorithms for small autonomous systems, assessed solutions are not available and complex technical challenges still need to be addressed. For example, in adverse situations (i.e close to structures) the effects of the environment could compromise the stability of the autonomous systems and therefore its mission. Navigation and control algorithms should be able to handle these aspects. The problem of disturbance attenuation and rejection with unknown nonlinear and missing measurements as active disturbance rejection control was studied with the design of linear/linearize control systems, with an approximation of the disturbance rejection. Some limitations on these studies are mainly related to (i) the linearization of the dynamic systems and single input case and (ii) approximation of unmodeled dynamics and measurement noises.
This proposal aims to advance the field of sensor data fusion by developing robust nonlinear algorithms that improve perception and decision-making in critical environments/applications.
Sensor data fusion is essential to enhance accuracy and reliability by integrating information from diverse sensors effectively. To enhance the accuracy and reliability of perception and decision-making in these domains, there is a need for robust nonlinear algorithms that can effectively integrate data from different sensors. This proposal aims to develop and advance robust nonlinear algorithms for sensor data fusion, for enhancing mission performance. Online adaptive management and configuration of constrained sensor resources to optimize the result of various tasks in accordance with multiple, possibly conflicting, performance measures.
The primary objectives of this project are as follows:
- Research and development of novel nonlinear sensor data fusion algorithms that can handle complex sensor data distributions.
- Implementation and optimization of the proposed algorithms to achieve real-time performance suitable for practical applications.
- Evaluation and validation of the algorithms using simulated and real-world datasets to demonstrate their effectiveness compared to existing fusion techniques.
- Integration of the developed algorithms into specific use-case scenarios, to assess their practical applicability and performance.
- Nonlinear and Adaptive Variable Structure Controllers for Spacecraft Proximity Operations
A key aspect of a successful space mission is the proper orientation of the spacecraft during each phase of the mission itself. All the satellite’s on-board systems contribute directly or indirectly to this, but the management of the spacecraft attitude is mostly the responsibility of the Guidance, Navigation and Control (GNC), actuation, and attitude sensors systems. Moreover, the key challenge of this research is the definition of a framework for a modular spacecraft, in which the sensors are distributed on variable-swarm systems. The main advantages of formation flight are improved flexibility and redundancy of the space mission, allowing adaptation to mission, target, and number of elements. The research area is dedicated on the development, implementation and validation of advanced Guidance Navigation and Control algorithms for spacecraft proximity formation flight.
The objective of the project is increasing the autonomy of the onboard systems and improving the robustness of the mission. Guidance and Control are combined to deal with different formation maneuvers (keeping, reconfiguration, deploying) in a compact implementation.
The main focus are:
- Modeling a modular spacecraft, including flexibility and reconfiguration of the system
- The design of control system for time-varying and uncertain systems
- Definition of non-hierarchical formation approaches.