Interplanetary Missions

Interplanetary Multiple Gravity-Assist missions are among the most challenging missions to design for aerospace engineers. The European Space Agency initiated the Global Trajectory Optimization Competition (GTOC) in 2005 to allow a comparison among the different global optimization models and tools used by different research groups. At present, the GTOC is an event taking place every one-two years over roughly one month during which the best aerospace engineers and mathematicians worldwide challenge themselves to solve a “nearly-impossible” problem of interplanetary trajectory design. Our laboratory participated, within a joint team with Polytechnic of Turin, in the 10th edition of the GTOC, organized by NASA JPL, ranking 7th out of 73 registered teams.

The same team won the GTOC 6th edition in 2012.

Stochastic Evolutionary Optimization Algorithms

Space trajectory optimization problems often present several features that make them hard to solve with local or deterministic optimization methods. For this reason, a variety of stochastic meta-heuristic techniques have been applied to space-based problem over the last decades. A meta-heuristic is a general purpose procedure designed to find a good-quality solution to an optimization problem in a limited amount of time.  Prominent examples are methods inspired by natural systems, as Simulated Annealing and Evolutionary Algorithms. Since 2015, our laboratory is developing an in-house    solver named EOS, Evolutionary Optimization at Sapienza, a self-adaptive, multi-population Differential Evolution (DE) algorithm for constrained global optimization.

Rocket Ascent Trajectory Optimization

Ascent trajectory optimization is a relevant problem in the space industry. 

Chemical propulsion is the only reliable way to inject a payload in orbit and only a small fraction of the total launch vehicle mass can be delivered. Ascent trajectory optimization consists in searching for the rocket trajectory that maximizes the payload delivered in orbit while ensuring the respect of several mission constraints, among which (i) small injection errors, with respect to a desired final orbit, (ii) path constraints to limit aerodynamic and thermal loads, and (iii) safety constraints on the impact points of the spent stages.

Convex Optimization

Convex optimization is a class of mathematical programming for which convergence toward the global optimum is guaranteed in a limited, short, time thanks to the availability of state-of-the-art highly-efficient numerical algorithms. Since most aerospace problems cannot be readily solved as convex optimization problems, our research focuses on converting a given nonconvex problem into a convex one, through a process referred to as convexification. We successfully applied convexification techniques on several aerospace problems, among which the ascent of a multistage launcher and a cooperative rendezvous mission. 

Cooperative UAV swarm control

Cooperative UAV swarm control focuses on enabling a group of unmanned aerial vehicles to coordinate their actions autonomously to accomplish shared mission objectives, such as load transportation. 

Unlike single-agent systems, swarm control requires the development of distributed algorithms that ensure real-time cooperation, collision avoidance, and adaptability to dynamic environments.
Our research focuses on the development of scalable and robust control strategies that rely on limited inter-agent communication and decentralized decision-making.