Developing optimal controllers for aggressive high-speed quadcopter flight poses significant challenges in robotics. Recent trends in the field involve utilizing neural network controllers trained through supervised or reinforcement learning. However, the sim-to-real transfer introduces a reality gap, requiring the use of robust inner loop controllers during real flights, which limits the network's control authority and flight performance. In this paper, we investigate for the first time, an end-to-end neural network controller, addressing the reality gap issue without being restricted by an inner-loop controller. The networks, referred to as G&CNets, are trained to learn an energy-optimal policy mapping the quadcopter's state to rpm commands using an optimal trajectory dataset. In hover-to-hover flights, we identified the unmodeled moments as a significant contributor to the reality gap. To mitigate this, we propose an adaptive control strategy that works by learning from optimal trajectories of a system affected by constant external pitch, roll and yaw moments. In real test flights, this model mismatch is estimated onboard and fed to the network to obtain the optimal rpm command. We demonstrate the effectiveness of our method by performing energy-optimal hover-to-hover flights with and without moment feedback. Finally, we compare the adaptive controller to a state-of-the-art differential-flatness-based controller in a consecutive waypoint flight and demonstrate the advantages of our method in terms of energy optimality and robustness.

Aggressive time-optimal control of quadcopters poses a significant challenge in the field of robotics. The state-of-the-art approach leverages reinforcement learning (RL) to train optimal neural policies. However, a critical hurdle is the sim-to-real gap, often addressed by employing a robust inner loop controller-an abstraction that, in theory, constrains the optimality of the trained controller, necessitating margins to counter potential disturbances. In contrast, our novel approach introduces high-speed quadcopter control using end-to-end RL (E2E) that gives direct motor commands. To bridge the reality gap, we incorporate a learned residual model and an adaptive method that can compensate for modeling errors in thrust and moments. We compare our E2E approach against a state-of-the-art network that commands thrust and body rates to an INDI inner loop controller, both in simulated and real-world flight. E2E showcases a significant 1.39-second advantage in simulation and a 0.17-second edge in real-world testing, highlighting end-to-end reinforcement learning's potential. The performance drop observed from simulation to reality shows potential for further improvement, including refining strategies to address the reality gap or exploring offline reinforcement learning with real flight data.

This Review discusses the main results obtained in training end-to-end neural architectures for guidance and control of interplanetary transfers, planetary landings, and close-proximity operations, highlighting the successful learning of optimality principles by the underlying neural models. Spacecraft and drones aimed at exploring our solar system are designed to operate in conditions where the smart use of onboard resources is vital to the success or failure of the mission. Sensorimotor actions are thus often derived from high-level, quantifiable, optimality principles assigned to each task, using consolidated tools in optimal control theory. The planned actions are derived on the ground and transferred on board, where controllers have the task of tracking the uploaded guidance profile. Here, we review recent trends based on the use of end-to-end networks, called guidance and control networks (G&CNets), which allow spacecraft to depart from such an architecture and to embrace the onboard computation of optimal actions. In this way, the sensor information is transformed in real time into optimal plans, thus increasing mission autonomy and robustness. We then analyze drone racing as an ideal gym environment to test these architectures on real robotic platforms and thus increase confidence in their use in future space exploration missions. Drone racing not only shares with spacecraft missions both limited onboard computational capabilities and similar control structures induced from the optimality principle sought but also entails different levels of uncertainties and unmodeled effects and a very different dynamical timescale.

While novel artificial intelligence and machine learning techniques are evolving and disrupting established terrestrial technologies at an unprecedented speed, their adaptation onboard satellites is seemingly lagging. A major hindrance in this regard is the need for high-quality annotated data for training such systems, which makes the development process of machine learning solutions costly, time-consuming, and inefficient. This paper presents “the OPS-SAT case”, a novel data-centric competition that seeks to address these challenges. The powerful computational capabilities of the European Space Agency’s OPS-SAT satellite are utilized to showcase the design of machine learning systems for space by using only the small amount of available labeled data, relying on the widely adopted and freely available open-source software. The generation of a suitable dataset, design and evaluation of a public data-centric competition, and results of an onboard experimental campaign by using the competition winners’ machine learning model directly on OPS-SAT are detailed. The results indicate that adoption of open standards and deployment of advanced data augmentation techniques can retrieve meaningful onboard results comparatively quickly, simplifying and expediting an otherwise prolonged development period.

In this paper, we combine the concepts of hyper-volume, ant colony optimization and nondominated sorting to develop a novel multi-objective ant colony optimizer for global space trajectory optimization. In particular, this algorithm is first tested on three space trajectory bi-objective test problems: an Earth-Mars transfer, an Earth-Venus transfer and a bi-objective version of the Jupiter Icy Moons Explorer mission (the first large-class mission of the European Space Agency's Cosmic Vision 2015-2025 programme). Finally, the algorithm is applied to a four-objectives low-thrust problem that describes the journey of a solar sail towards a polar orbit around the Sun. The results on both the test cases and the more complex problem are reported by comparing the novel algorithm performances with those of two popular multi-objective optimizers (i.e., a nondominated sorting genetic algorithm and a multi-objective evolutionary algorithm with decomposition) in terms of hypervolume metric. The numerical results of this study show that the multi-objective hypervolume-based ant colony optimization algorithm is not only competitive with the standard multi-objective algorithms when applied to the space trajectory test cases, but it can also provide better Pareto fronts in terms of hypervolume values when applied to the complex solar sailing mission.

Optimal control holds great potential to improve a variety of robotic applications. The application of optimal control on-board limited platforms has been severely hindered by the large computational requirements of current state of the art implementations. In this work, we make use of a deep neural network to directly map the robot states to control actions. The network is trained offline to imitate the optimal control computed by a time consuming direct nonlinear method. A mixture of time optimality and power optimality is considered with a continuation parameter used to select the predominance of each objective. We apply our networks (termed GCNets) to aggressive quadrotor control, first in simulation and then in the real world. We give insight into the factors that influence the 'reality gap' between the quadrotor model used by the offline optimal control method and the real quadrotor. Furthermore, we explain how we set up the model and the control structure on-board of the real quadrotor to successfully close this gap and perform time-optimal maneuvers in the real world. Finally, GCNet's performance is compared to state-of-the-art differential-flatness-based optimal control methods. We show, in the experiments, that GCNets lead to significantly faster trajectory execution due to, in part, the less restrictive nature of the allowed state-to-input mappings.

Self-supervised learning is a reliable learning mechanism in which a robot uses an original, trusted sensor cue for training to recognize an additional, complementary sensor cue. We study for the first time in self-supervised learning how a robot’s learning behavior should be organized, so that the robot can keep performing its task in the case that the original cue becomes unavailable. We study this persistent form of self-supervised learning in the context of a flying robot that has to avoid obstacles based on distance estimates from the visual cue of stereo vision. Over time it will learn to also estimate distances based on monocular appearance cues. A strategy is introduced that has the robot switch from flight based on stereo to flight based on monocular vision, with stereo vision purely used as “training wheels” to avoid imminent collisions. This strategy is shown to be an effective approach to the “feedback-induced data bias” problem as also experienced in learning from demonstration. Both simulations and real-world experiments with a stereo vision equipped ARDrone2 show the feasibility of this approach, with the robot successfully using monocular vision to avoid obstacles in a 5 × 5 m room. The experiments show the potential of persistent self-supervised learning as a robust learning approach to enhance the capabilities of robots. Moreover, the abundant training data coming from the own sensors allow to gather large data sets necessary for deep learning approaches.