GeoPlanet: Earth and Planetary Sciences

Jerzy Sąsiadek Editor

Aerospace Robotics II

GeoPlanet: Earth and Planetary Sciences

Editor-in-chief

Paweł Rowiński

Series editors

Marek Banaszkiewicz, Warsaw, PolandJanusz Pempkowiak, Sopot, PolandMarek Lewandowski, Warsaw, PolandMarek Sarna, Warsaw, Poland

More information about this series at http://www.springer.com/series/8821

Jerzy SąsiadekEditor

Aerospace Robotics II

123

EditorJerzy SąsiadekMechanical and Aerospace EngineeringCarleton UniversityOttawa, ONCanada

The GeoPlanet: Earth and Planetary Sciences Book Series is in part a continuation ofMonographic Volumes of Publications of the Institute of Geophysics, Polish Academy ofSciences, the journal published since 1962 (http://pub.igf.edu.pl/index.php).

ISSN 2190-5193 ISSN 2190-5207 (electronic)GeoPlanet: Earth and Planetary SciencesISBN 978-3-319-13852-7 ISBN 978-3-319-13853-4 (eBook)DOI 10.1007/978-3-319-13853-4

Library of Congress Control Number: 2013932857

Springer Cham Heidelberg New York Dordrecht London© Springer International Publishing Switzerland 2015This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or partof the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmissionor information storage and retrieval, electronic adaptation, computer software, or by similar ordissimilar methodology now known or hereafter developed.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exemptfrom the relevant protective laws and regulations and therefore free for general use.The publisher, the authors and the editors are safe to assume that the advice and information in thisbook are believed to be true and accurate at the date of publication. Neither the publisher nor theauthors or the editors give a warranty, express or implied, with respect to the material containedherein or for any errors or omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media(www.springer.com)

Series Editors

Geophysics Paweł RowińskiEditor-in-ChiefInstitute of GeophysicsPolish Academy of Sciencesul. Ks. Janusza 6401-452 Warsaw, Polandp.rowinski@igf.edu.pl

Space Sciences Marek BanaszkiewiczSpace Research CentrePolish Academy of Sciencesul. Bartycka 18A00-716 Warsaw, Poland

Oceanology Janusz PempkowiakInstitute of OceanologyPolish Academy of SciencesPowstańców Warszawy 5581-712 Sopot, Poland

Geology Marek LewandowskiInstitute of Geological SciencesPolish Academy of Sciencesul. Twarda 51/5500-818 Warsaw, Poland

Astronomy Marek SarnaNicolaus Copernicus Astronomical CentrePolish Academy of Sciencesul. Bartycka 1800-716 Warsaw, Polandsarna@camk.edu.pl

Managing Editor

Anna DziembowskaInstitute of Geophysics, Polish Academy of Sciences

Advisory Board

Robert AnczkiewiczResearch Centre in KrakówInstitute of Geological SciencesKraków, Poland

Aleksander BrzezińskiSpace Research CentrePolish Academy of SciencesWarsaw, Poland

Javier CuadrosDepartment of MineralogyNatural History MuseumLondon, UK

Jerzy DeraInstitute of OceanologyPolish Academy of SciencesSopot, Poland

Evgeni FedorovichSchool of MeteorologyUniversity of OklahomaNorman, USA

Wolfgang FrankeGeologisch-Paläntologisches InstitutJohann Wolfgang Goethe-UniversitätFrankfurt/Main, Germany

Bertrand FritzEcole et Observatoire desSciences de la Terre,Laboratoire d’Hydrologieet de Géochimie de StrasbourgUniversité de Strasbourg et CNRSStrasbourg, France

Truls JohannessenGeophysical InstituteUniversity of BergenBergen, Norway

Michael A. KaminskiDepartment of Earth SciencesUniversity College LondonLondon, UK

Andrzej KijkoAon BenfieldNatural Hazards Research CentreUniversity of PretoriaPretoria, South Africa

Francois LeblancLaboratoire Atmospheres, MilieuxObservations Spatiales, CNRS/IPSLParis, France

Kon-Kee LiuInstitute of Hydrologicaland Oceanic SciencesNational Central University JhongliJhongli, Taiwan

Teresa MadeyskaResearch Centre in WarsawInstitute of Geological SciencesWarsaw, Poland

Stanisław MasselInstitute of OceanologyPolish Academy of SciencesSopot, Polska

Antonio MeloniInstituto Nazionale di GeofisicaRome, Italy

Evangelos PapathanassiouHellenic Centre for Marine ResearchAnavissos, Greece

Kaja PietschAGH University of Scienceand TechnologyKraków, Poland

Dušan PlašienkaPrírodovedecká fakulta, UKUniverzita KomenskéhoBratislava, Slovakia

Barbara PopielawskaSpace Research CentrePolish Academy of SciencesWarsaw, Poland

Tilman SpohnDeutsches Zentrum für LuftundRaumfahrt in der HelmholtzGemeinschaftInstitut für PlanetenforschungBerlin, Germany

Krzysztof StasiewiczSwedish Institute of Space PhysicsUppsala, Sweden

Roman TeisseyreEarth’s Interior Dynamics LabInstitute of GeophysicsPolish Academy of SciencesWarsaw, Poland

Jacek TronczynskiLaboratory of Biogeochemistryof Organic ContaminantsIFREMER DCN_BENantes, France

Steve WallisSchool of the Built EnvironmentHeriot-Watt UniversityRiccarton, EdinburghScotland, UK

Wacław M. ZuberekDepartment of Applied GeologyUniversity of SilesiaSosnowiec, Poland

Preface I

Dear Colleagues and Friends,Space robotics is a fascinating field that was developed for space exploration andspace missions. Space is a hostile environment for humans and could not beexplored earlier because of technological factors. Simply speaking, our technologyin the past was not advanced enough to allow us to venture safely into space. Also,in many cases, traveling in space takes more than a human’s life span. This isexactly the reason why we have to use robots of a variety of sorts and types. Robotsallow humans to extend their abilities to perform missions that otherwise would bedangerous or simply impossible. Robots make long-duration missions not onlypossible but also present opportunity for exploration of the solar system safely andefficiently. The space automation and autonomous missions are also a viablesolution to deep space travel and exploration problems.

One of the issues that needs urgent solution is the problem of space debrisremoval from Earth high orbits. One of the viable possibilities is the application ofrobotics including spacecraft base with two-arm manipulator.

It is with great pleasure that we would like to show once more to the interna-tional space robotics community, selected papers presented at the II Conference onAerospace Robotics held in Warsaw, Poland during 1–2 July 2013.

We have selected 18 papers from the conference that reflect the interests andtopics of all papers presented at this conference. This book includes papers rangingfrom space manipulators control and applications to mobile robots and sampletaking instrumentation.

We hope that our readers will find these papers not only interesting but alsohelpful and inspiring in their professional activities.

Warsaw, June 2014 Marek BanaszkiewiczJerzy Sąsiadek

ix

Preface II

Dear Friends,It is with great pleasure that we invite the international robotics community to the IIConference on Aerospace Robotics to be held in Warsaw, Poland during 1–2 July2013.

This is a time that is particularly important for the Polish aerospace roboticscommunity. In July 2012, the European Space Agency Council unanimouslyapproved the accession of Poland to the ESA Convention. In November 2012,Polish President Bronisław Komorowski signed the final ratification documents andPoland became the 20th member of the Agency. The timing was significant becausesoon after, an important meeting was held, where the ESA space policy for the nextdecade was discussed and shaped.

Without any doubt, it is also a singular moment in terms of global spaceexploration. In August 2012, as a part of NASA Mars Science Laboratory,Curiosity rover landed on the surface of Mars. On the other hand, the EuropeanSpace Agency is planning to launch two missions to Mars—ExoMars (possiblelaunch in 2016), as a part of the Aurora Program, and Mars Sample Return Mission,in cooperation with NASA, planned for 2020–2022.

However, Space Agencies are not the only entities interested in shaping thepriorities for the development of space activities. The European Union Commissionitself is an important player in designing and shaping space sector activities andpolicies. Article 189 of the Lisbon Treaty gives the EU an explicit role in designinga policy for the exploration and exploitation of space; the European Union spacepolicy is a key element of the Europe 2020 strategy and an integral partof the European industrial policy. Flagship space projects of the EU include thedevelopment of the European navigation satellite program Galileo, and theimplementation of the Copernicus (previously named European Global Monitoringfor Environment and Security—GMES) program.

The Polish space community activities are synchronized with broader Europeanactivities and for a number of years Polish institutes, universities, and privatecompanies have subscribed to the priorities defined by ESA or EU and cooperatedwith European partners. This includes Polish participation in GMES, EGNOS,

xi

Space Situational Awareness, collaboration with ESA within the framework ofPECS (Plan for European Cooperating State) Charter, and many more.

We hope that our II Conference on Aerospace Robotics will help in promotingdevelopment and cooperation within the robotics community.

We would like to thank the Space Research Centre (CBK PAN) in Warsaw fortheir organizational and financial support of this conference. Also, we would like tothank all CBK PAN employees for their time and efforts devoted to this conference.In particular, we thank Dr. Karol Seweryn for his valuable contributions andimmeasurable devotion.

Warsaw, June 2013 Marek BanaszkiewiczJerzy Sąsiadek

xii Preface II

Contents

Robots for Space Exploration—Barriers, Perspectivesand Implementations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1Teresa Zielinska

On the Simple Adaptive Control of Flexible-Joint SpaceManipulators with Uncertainties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Steve Ulrich and Jurek Z. Sasiadek

Hybrid Method of Motion Planning for Driftless Systems . . . . . . . . . . 25Ignacy Duleba and Michal Opalka

The Formation Flying Navigation System for Proba 3 . . . . . . . . . . . . . 37João Branco, Valentín Barrena, Diego Escorial Olmos,Lorenzo Tarabini Castellani and Alexander Cropp

Innovative Resistojet Propulsion System—Use in RoboticSpace Platforms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Łukasz Mężyk, Łukasz Boruc, Arkadiusz Kobiera, Jan Kindracki,Karol Seweryn and Tomasz Rybus

Analyses of a Free-Floating Manipulator Control Scheme Basedon the Fixed-Base Jacobian with Spacecraft Velocity Feedback . . . . . . 59Tomasz Rybus, Tomasz Barciński, Jakub Lisowski and Karol Seweryn

TwinCube—Preliminary Study of a Tether Experimentfor CubeSat Mission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71Tomasz Szewczyk, Tomasz Barciński, Tomasz Rybus,Łukasz Wiśniewski, Agata Białek, Jerzy Grygorczuk,Marcin Krzewski, Tomasz Kuciński, Jakub Lisowski,Marek Morawski, Rafał Przybyła, Hanna Rothkaehl,Marta Tokarz and Roman Wawrzaszek

xiii

Novel Type of Inertial Actuator for Satellite Attitude ControlSystem Basis on Concept of Reaction Sphere—ELSA Project . . . . . . . 85R. Wawrzaszek, M. Sidz, M. Strumik, M. Banaszkiewicz,K. Seweryn, Ł. Wiśniewski, L. Rossini and E. Onillon

Deployable Manipulator Technology with Application for UAVs . . . . . 93Tomasz Kuciński, Tomasz Rybus, Karol Seweryn,Marek Banaszkiewicz, Tomasz Buratowski, Grzegorz Chmaj,Jerzy Grygorczuk and Tadeusz Uhl

MERTIS/BEPI COLOMBO Pointing Unit Mechanism—PointingAccuracy Test Procedure, Setup and Results . . . . . . . . . . . . . . . . . . . 105Mirosław Rataj, Robert Pietrzak and Piotr Wawer

Design of the New Pi of the Sky Robotic Telescope Controlledvia Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117Jerzy Grygorczuk, Marcin Dobrowolski, Grzegorz Juchnikowski,Marek Morawski, Lech Mankiewicz, Aleksander Filip Żarnecki,Mikołaj Ćwiok, Marcin Zaremba, Tadeusz Batschand Robert Wilczyński

Satellite Guided Navigation Control for Environment Monitoring . . . . 129Marek Zaremba, Fadi Halal, Thomas Hirose and Pablo Pedrocca

Analog Mars Rover Service as a Robotic Hardwareand Team Building Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Sebastian Meszyński and Mateusz Józefowicz

Manipulators of the Phobos-Grunt Project and Lunar Projects . . . . . . 163O.E. Kozlov and T.O. Kozlova

On a Hybrid Genetic Algorithm Solving a Global Path Planningfor a Ground Mobile Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175Piotr Bigaj and Jakub Bartoszek

Outline of an Autonomy Framework for Space Mobile Robots . . . . . . 187A. Medina, A. Bidaux-Sokolowski, G. Binet, M. Avilés, J. Ocón,A. Ceballos and P. Colmenarejo

xiv Contents

Robots for Space Exploration—Barriers,Perspectives and Implementations

Teresa Zielinska

Abstract The article focuses walking machines dedicated to the complex explora-tion tasks. The problem of autonomy increase is discussed taking into account thedesign and control issues. The factors limiting the autonomous robots developmentare summarized. It is emphasised that the biological inspiration suggests the effectivesolutions useful for autonomous robots with increasedmobility. Selected examples ofdesign and control for machines devoted to autonomous exploration are presented.

Keywords Walking machines � Space robots design � Legged robots

1 Introduction

The exploration of the other planets requires the autonomous robots which are able tomove on difficult terrain. The surface conditions are there not so well identified andnon-uniform surface properties with undulating shape must be expected. Thereforethe robots must be able to act autonomously adapting its motion features to theexternal conditions. The level of autonomy in actions (intelligence) is determined byproperties of mechanical structure and abilities of the control system. The animalworld gives a good example of efficient adaptation to the environmental conditions.Analysing the body build and neural structures it can be observed that more complexbodies applies more advanced control. That means that in animals the complexity ofnervous system is proportional to the complexity of the body. But transferring it intotechnical world it does not mean that the increase of robot autonomy should beobtained only by increase of complexity of mechanics and control. Another propertyof animals world it that the body structure and neural system matches the livingconditions. The simple animals, with primitive body and control centres can survive

T. Zielinska (&)Faculty of Power and Aerospace Engineering, Warsaw University of Technology, Ul.Nowowiejska 24, 000-665 Warsaw, Polande-mail: teresaz@meil.pwe.edu.pl

© Springer International Publishing Switzerland 2015J. Sąsiadek (ed.), Aerospace Robotics II, GeoPlanet: Earth and Planetary Sciences,DOI 10.1007/978-3-319-13853-4_1

1

well due to the proper spontaneous reactions (arising from an impulse, not pre-meditated). This suggests that the robot mechanical structure and actuation systemmust be properly chosen to the assumed working conditions and the task. Someactions produced usually by sophisticated control can be obtained much simpler bythe mechanics, and it not means that mechanics must be complex.

2 Technical Limitations

Comparing biological world and the technical abilities we can notice that the fol-lowing factors make the most significant barriers limiting the robots abilities:

Motors Motors have limited efficiency, majority of motors produce revolute motion,they lack muscles flexibility

Sensors Existing sensors are not as effective as those I animals, they types arelimited, they are space and energy consuming

Power sources Power sources are still too big, too heavy and ineffective comparing to theliving world

Mechanicalsystem

Mechanical structures are made of stiff materials, in general design solutionslack of novelty, however recently developed bio-robots represent non-conventional tendencies

Controlsystems

Control systems are space and energy consuming and much less efficientthan the biological ones.

More information about above listed properties of robotics components can befound in (Siciliano and Khatib 2008), the characteristics of animals sensing andcontrol systems from technical point of view can be found in (Doncieux et al. 2010).

3 Design Solutions

3.1 Design Problems

The advantage of legged robots for space exploration is the ability to move onuneven terrain and to adapt the gait features to the surface. But they have manydegrees of freedom requiring actuators what increases the body weight and energyconsumption. Those are main disadvantages comparing such robots with the otherpossible e.g. with wheeled robots. But the engineering technologies are constantlyimproved, in longer perspective we expect more efficient and light actuators andimproved sources of the energy. Therefore the research on walking robots for spaceexploration is continued (Menon et al. 2007). Walking robots for space explorationmust possess greater mobility than those used in Earth.

2 T. Zielinska

The inverted pendulum template consisting of point mass located on top of a stiffrod with only one degree of freedom is used for modelling the periodic fast motion ofarthropods and vertebrates. For modelling the bouncing and running it is used thespring loaded inverted pendulummodel, this allows better representation of biologicalmovement properties. Those models are successfully used for designing of mechan-ical structure and synthesising the control algorithms for jumping and running robots.

Animals leg shows flexibility which is important for dynamical movement(spring effect in leg-end take-off and shock absorption during landing). Theresearch on compliant legs started just recently and it concerns mainly humanoids(Zielinska and Chmielniak 2011). Compliant leg naturally stabilizes the postureminimising the need for additional adjustments, and allowing to obtain smoothmotion trajectories. Significant progress is seen in the development of humanoidalrobots, in that also humanoids substituting man in dangerous space missions. Themost advanced is Robonaut 2 developed by NASA (2012).

3.2 Examples

Multi-legged walking robots often applies identical legs and their contribution intobody displacement is the same, what is not like in animals. The range of motion forrobotic legs is often very limited and large leg workspace allows not only walkingbut climbing or crawling. Moreover the legs can bend for transportation saving thespace what is very important for space transportation. The large leg work space canalso offer manipulation ability. Moreover the turn-over is important when robotcarries special equipment on same part of his body. In complex terrain the robot canoverturn, and coming back to proper position is needed. Analyzing current state ofthe art not many robots with such properties were identified. Besides of our workswe are able to list hexapod IOAN developed by the end of XXc. in Belgium(Premount et al. 1997) for space exploration—Fig. 1.

Fig. 1 Design conceptapplied in IOAN machine.This design employs twoservo motors attached back toback and each motor isresponsible for either the liftor swing axes. This design ifemployed carefully allows thedesign to be invertible

Robots for Space Exploration—Barriers … 3

Our leg concept applies modular design (Zielinska and Heng 2002)—each legtogether with its actuators makes separate module which can be attached to differenttrunk with different number of legs. Applied differential mechanism brings twodegrees of freedom in one joint—Fig. 2. The motors actuating this joint (hip) areattached to the trunk what means that only the knee motor must be transferred whenthe leg is in transfer phase, This saves the energy.

Combined work of two motors in the hip increases the delivered power what isvery important because this joint is the most loaded. The design was successfullyused in family of walking machines, starting from the small size and mass (15 cmheight and 2 kg mass to 60 cm height and 40 kg mass with payload 40 kg). The legscan use different postures during walking, they can work as manipulators and themachine can actively turn-over—Fig. 3.

Fig. 2 Modular leg design assuring large work-space—our works

4 T. Zielinska

4 Control Systems

4.1 Control Concepts

Besides of themechanical structure another barrier in the development of autonomouswalking robots results from traditional approach to the synthesis of control systems.Current robots still miss at least such autonomy of actions as it is observed in the insectworld, but their performances are improving very fast, especially in the recent years.

In first robots controlled by computers the motion patterns were fixed. Theprototypes were not equipped with external sensors (sensors delivering the infor-mation about the environment, or about the relation between environment andmachine i.e. distance to the obstacles) and the motion was not influenced by sur-rounding. In the farther works on increase the motion autonomy model basedapproach was dominating. Motion actions were evaluated using models of envi-ronment and assuming that the environment response matches the patterns. Thosetraditional methods of the so called intelligent control of technical devices werederived from the old Greek tradition of understanding of the source of intelligence.In this approach the role of hierarchic system of reasoning, utilizing knowledge,memory, etc. was underscored. This understanding of source of intelligent behav-iour for long years dominated the works of psychologists and was reflected intechnical sciences by classical methods of artificial intelligence. After the SecondWorld War the development cybernetics (the study of human control functions andof mechanical and electronic systems designed to replace them) was tightly relatedwith the development of control theory and brain research. The role of brain as the

Fig. 3 Quadruped LAVA illustrates the legs motion ranges with turnover ability, the legs can bebend for transportation

Robots for Space Exploration—Barriers … 5

central control unit and feedback from receptors to brain was underscored. Thebrain was considered as the centre of all decisions. The researchers agreed that brainuses models of environment to perceive correct control. Arbib (1972) formulatedthe metaphor “human is a mechanism” what means that it is possible to describehuman functioning using the machine analogy (e.g. analogy to sequential machine).He concluded that it is possible to compare the brain research to the research ofunknown calculating system. In many works from period of cybernetics and cen-tralized control development the role of knowledge, reasoning and modelling ofintelligent behaviours of humans and robots was underscored. In robotics this pointof view is reflected by hierarchic structure of control system, where the decisions ofthe higher levels are obligatory for the lower levels.

The source of new ideas and progress in autonomous motion synthesis were theworks of zoologists. By investigation of the behaviours of primitive animals(usually insects) they became convinced that even complex behaviours are attainedby simple, distributed neural structures quickly transforming information fromreceptors into adequate behaviours (activities) or reflex actions. The generation ofinsect behaviours is described by parallel structures of information flow. Theinformation from receptors is processed simultaneously by the groups of neuronslocated in the different level of the nervous system (Beer 1990), signals from somereceptors are received only by specific neurons. Some neural signals produced byhigher neural levels are sending to lover levels which also receive the additionalinformation from selected sensors. The environment model is not created by neu-rons and the knowledge of the laws governing the environment features is notinvoked, the neural system produces output signals as the responses to the inputsreceived by receptors. This scheme of direct receptors-effectors relation found thereflection in robotics.

The works on robots control by (Brooks and Stein 1994) which have beenappearing since 1986, are usually quoted in this context. He elaborated the idea ofadaptive behaviour underscoring that adaptability is obtained by proper reactions tothe sensory readings and the control system shall be elaborated step by step startingfrom simple behaviours as it is observed in the personal development of livingorganism.He proposed subsumption architecture of control systemswhere the actionsare evaluated due to the deliberation between layers of control systems responsible forseparate behaviours. More complex behaviours are built using simpler ones.

Distribution of the control tasks (control functions) over the software bringsfunctional architecture. It is difficult to give the clear, unique classification ofexisting functional architectures of robot control systems, but classification ofapproaches used for motion generation enables to get systematic view.

In general, the four listed below approaches are used in motion generation:

1. Model based methods employing hierarchic division of motion tasks.Those approaches always dictate traditional controller structure known from

industrial robotics. Information flow is hierarchic. Planning is done by begin-ning with generic notions and ending with details. The decisions of higher levelsare obligatory to lover levels of functional control structure.

6 T. Zielinska

2. Methods focusing on receptor information processing into adequate actions, thestructures of controllers, in this case, may differ a lot. Reactive control,behavioural controls are here examples.

The functional structure of control system is dictated by different type ofutilized information and schemes of information flow which influences thedevice motion and behaviours. Here is used the descriptive concept of intelli-gence but without imitating the biological structures.

3. Methods in which the motion criteria determines the control actions.Those methods are mainly developed for walking robots. In early works

these criteria were building using hypotheses concerning the quality of humanwalk. For example the criterions of the minimization if joint torques, or jointforces, minimization of body centre of gravity accelerations, minimization ofcentre of gravity vibrations or zero moment point (ZMP) criterion were applied.In his recent works on modelling of human motion Khatib pays attention to theinteraction between the posture and the task, the motion must fulfil the criteriawhich results from internal (robot) constraints, environment conditions andposture—i.e. expected trajectory of the centre of robot mass during change ofbody posture (Khatib 2002).

4. Methods reflecting the motion patterns or reflecting the neural structuresobserved in living world (biomimetic control).

As the example we can give the motion generation methods in which the gaitimage or some gait features are imitated utilizing the results of observation ofliving creature locomotion, for example reproducing during machine motion thepattern of distribution of reaction forces observed during insect motion or fol-lowing the fixed sequence of gait pattern delaying the motion of appropriate legson the ground with pebbles. The investigations of insects neural gait patterngenerators encompass construction of such generator models in six leggedwalking machines control systems (Collins and Richmond 1994).

4.2 Sensing and Sensors

Biologically inspired understanding of the mechanism of intelligence emphasisesthat the adaptability is obtained by proper reactions to the sensory readings.

The fundamental questions determining the choice of sensors for newly devel-oped robot are following:

• What is the expected level of device autonomy?• To what extend environment is unstructured?• What type of environment information is required (e.g. acting external forces or

torques, distances, object shapes, object coordinates, surface profile)?• How reliably the information must be captured?• How should the sensors be mounted on the mechanical structure?

Robots for Space Exploration—Barriers … 7