Modeling irregular small bodies gravity field via extreme learning machines and Bayesian optimization

Close proximity operations around small bodies are extremely challenging due to their uncertain dynamical environment. Autonomous guidance and navigation around small bodies require fast and accurate modeling of the gravitational field for potential on-board computation. In this paper, we investigate a model-based, data-driven approach to compute and predict the gravitational acceleration around irregular small bodies. More specifically, we employ Extreme Learning Machine (ELM) theories to design, train and validate Single-Layer Feedforward Networks (SLFN) capable of learning the relationship between the spacecraft position and the gravitational acceleration. ELM-base neural networks are trained without iterative tuning therefore dramatically reducing the training time. Analysis of performance in constant density models for asteroid 25143 Itokawa and comet 67/P Churyumov-Gerasimenko show that ELM-based SLFN are able learn the desired functional relationship both globally and in selected localized areas near the surface. The latter results in a robust neural algorithm for on-board, real-time calculation of the gravity field needed for guidance and control in close-proximity operations near the asteroid surface.     

Modeling and identification of an electro-mechanical system: The LISA grabbing positioning and release mechanism case

In the frame of space missions, mechanisms often constitute critical systems whose functionality and performance need to be tested on ground before the mission launch. The LISA scientific space mission will detect gravitational waves by measuring the relative displacement of pairs of free-floating test masses set into geodesic motion onboard of three spacecrafts. Inside each satellite, the injection of the test masses from the caged configuration into the geodesic trajectory will be performed by the grabbing positioning and release mechanism. To provide a successful injection, the test masses must be dynamically released with a minimal residual velocity against adhesion with the holding device. A parameter that determines the test mass residual velocity is the quickness of the retraction of the holding device. The need arises then to characterize the dynamic response of the release mechanism in order to predict its behaviour in the in-flight conditions. Once a validated model of the mechanism is available, the compliance of the system to the tight requirement on the maximum allowed residual velocity of the test mass may be verified. Starting from an electro-mechanical model of the mechanism dynamics, this paper presents the results of the experimental identification of its relevant parameters.     

Image based control of the “PINOCCHIO” experimental free flying platform

A free floating platform is realized with a pneumatic suspension system which enables a two-dimensional test of complex space operations, such as rendezvous and docking. The platform is equipped with a IMU and actuated via cold gas thrusters. In addition, an on-board camera is used to acquire a target and its image is processed for evaluating the control actions needed to reach it. A technique for determining the relative position and velocity with respect to target using the same visual device is proposed and realized. The novel algorithms and relevant experimental results are presented. The tested accuracy of the relative navigation system is not very high, but the guidance algorithm, which is image based and has just a weak dependence on the position information, is robust enough to perform a successful maneuver. The error between the final acquired target image and the desired final target image is of the order of one pixel, notwithstanding all the testbed disturbances.     

A small mission for in situ exploration of a primitive binary near-Earth asteroid

We present a concept for a challenging in situ science mission to a primitive, binary near-Earth asteroid. A sub-400-kg spacecraft would use solar electric propulsion to rendezvous with the C-class binary asteroid (175706) 1996 FG3. A campaign of remote observations of both worlds would be followed by landing on the ∼1 km diameter primary to perform in situ measurements. The total available payload mass would be around 34 kg, allowing a wide range of measurement objectives to be addressed. This mission arose during 2004 from the activities of the ad-hoc Small Bodies Group of the DLR-led Planetary Lander Initiative. Although the particular mission scenario proposed here was not studied further per se, the experience was carried over to subsequent European asteroid mission studies, including first LEONARD and now the Marco Polo near-Earth asteroid sample return proposal for ESA’s Cosmic Vision programme. This paper may thus be of interest as much for insight into the life cycle of mission proposals as for the concept itself.     

Measurements of the upper troposphere and lower stratosphere during tropical cyclones using the GPS radio occultation technique

Water vapour transport to the upper troposphere and lower stratosphere by deep convective storms affects the radiation balance of the atmosphere and has been proposed as an important component of climate change. The aim of the work presented here is to understand if the GPS radio occultation technique is useful for characterization of this process. Our assessment addresses the question if severe storms leave a significant signature in radio occultation profiles in the upper troposphere/lower stratosphere. Radio occultation data from the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) were analyzed, focusing on two particular tropical cyclones with completely different characteristics, the hurricane Bertha, which formed in the Atlantic Basin during July 2008 and reached a maximum intensity of Category 3, and the typhoon Hondo, which formed in the south Indian Ocean during 2008 reaching a maximum intensity of Category 4. The result is positive, suggesting that the bending angle of a GPS radio occultation signal contains interesting information on the atmosphere around the tropopause, but not any information regarding the water vapour. The maximum percentage anomaly of bending angle between 14 and 18 km of altitude during tropical cyclones is typically larger than the annual mean by 5–15% and it can reach 20% for extreme cases. The results are discussed in connection to the GPS radio occultation receiver which will be part of the Atomic Clock Ensemble in Space (ACES) payload on the International Space Station.     

Tomographic reconstruction of wet and total refractivity fields from GNSS receiver networks

One of the most attractive scientific issues in the use of GNSS (Global Navigation Satellite System) signals, from a meteorological point of view, is the retrieval of high resolution tropospheric water vapour maps. The real-time (or quasi real-time) knowledge of such distributions could be very useful for several applications, from operative meteorology to atmospheric modelling. Moreover, the exploitation of wet refractivity field reconstruction techniques can be used for atmospheric delay compensation purposes and, as a very promising activity, it could be applied for example to calibrate SAR or Interferometric-SAR (In-SAR) observations for land remote sensing. This is in fact one of the objectives of the European Space Agency project METAWAVE (Mitigation of Electromagnetic Transmission errors induced by Atmospheric Water vapour Effects), in which several techniques are investigated and results were compared to identify a strategy to remove the contribution of water vapour induced propagation delays in In-SAR products. Within this project, the tomographic reconstruction of three dimensional wet refractivity fields from tropospheric delays observed by a local GNSS network (9 dual frequency GPS receivers) deployed over Como area (Italy), during 12–18 October, 2008, was performed. Despite limitations due to the network design, internal consistency tests prove the efficiency of the adopted tomographic approach: the rms of the difference between reconstructed and GNSS observed Zenith Wet Delays (ZWD) are in the order of 4 mm. A good agreement is also observed between our ZWDs and corresponding delays obtained by vertically integrating independent wet refractivity fields, taken by co-located meteorological analysis. Finally, during the observing period, reconstructed vertical wet refractivity profiles evolution reveals water vapour variations induced by simple cloud covering. Even if our main goal was to demonstrate the effectiveness in adopting tomographic reconstruction procedures for the evaluation of propagation delays inside water vapour fields, the actual water vapour vertical variability and its evolution with time is well reproduced, demonstrating also the effectiveness of the inferred 3D wet refractivity fields.