Precision landing on large and small planetary bodies is a technology of utmost importance for future human and robotic exploration of the solar system. In this context, the Zero-Effort-Miss/Zero-Effort-Velocity (ZEM/ZEV) feedback guidance algorithm has been studied extensively and is still a field of active research. The algorithm, although powerful in terms of accuracy and ease of implementation, has some limitations. Therefore with this paper we present an adaptive guidance algorithm based on classical ZEM/ZEV in which machine learning is used to overcome its limitations and create a closed loop guidance algorithm that is sufficiently lightweight to be implemented on board spacecraft and flexible enough to be able to adapt to the given constraint scenario. The adopted methodology is an actor-critic reinforcement learning algorithm that learns the parameters of the above-mentioned guidance architecture according to the given problem constraints.
Impedance and Admittance Control are two wellknown controllers to accomplish the same goal: the regulation of the mechanical impedance of manipulators interacting dynamically with the environment. However, they both are affected by a strong limitation deriving from their fixed causality, which causes their inability to provide good performance over a large spectrum of environment stiffnesses. In this paper an adaptive hybrid system framework is proposed to unify Impedance and Admittance formulations and consequently overcome this limit. Indeed, the hybrid framework interpolates the opposite performance and stability characteristics of the above-mentioned impedance-based control strategies leading to a family of controllers with intermediate properties, and thus suitable for several conditions. Moreover, the adaptivity allows the hybrid system to operate properly in an environment characterized by unknown and even time-varying stiffness. Especially, the work focuses on the development of this latter aspect and an adaptive solution based on a feedforward Neural Network is presented. The effectiveness of the novel control strategy is demonstrated by means of numerical simulations
This paper analyzes the real-time relative pose estimation and attitude prediction of a tumbling target spacecraft through a high-order numerical extended Kalman filter based on differential algebra. Indeed, in the differential algebra framework, the Taylor expansion of the phase flow is automatically available once the spacecraft dynamics is integrated and thus the need to write and integrate high-order variational equations is completely avoided making the presented solution easier to implement. To validate the technique, the ESA's e.deorbit mission, involving the Envisat satellite, is used as reference test case. The developed algorithms are implemented on a BeagleBone Black platform, as representative of the limited computational capability available on onboard processors. The performance is assessed by varying the measurement acquisition frequency and processor clock frequency, and considering various levels of uncertainties. A comparison among the different orders of the filter is carried out.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.