THE BENEFIT OF MULTIMODAL TELEPRESENCE FOR IN-SPACE ROBOTIC ASSEMBLY

Transcription

1 THE BENEFIT OF MULTIMODAL TELEPRESENCE FOR IN-SPACE ROBOTIC ASSEMBLY Enrico Stoll, Daniel Kwon Space Systems Laboratory, Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, Massachusetts, USA ABSTRACT Multimodal telepresence systems are often used for terrestrial applications but they not yet are practically applied in space. The deployment of robots with visual as well as haptic feedback for servicing operations in space is a valuable addition to the existing autonomous systems since it will provide flexibility and robustness in mission operations. The operator on Earth will no longer be a pure observer but will have the capability of real-time interaction with the space environment. For demonstrating the feasibility and showing the benefits of multimodal space applications a test environment is being developed, which is based on the SPHERES nano-satellite testbed at the MIT Space Systems Laboratory. KEY WORDS multimodal, telepresence, in-space robotic assembly, onorbit servicing, haptics, free flyer 1 Introduction Telepresence systems enable a human operator to actively manipulate and intervene in a remote environment. In this connection the term telepresence refers to a state, in which a human operator can no longer differentiate between an interaction with a real environment and a technical mediated one. There is a great variety of terrestrial applications for telepresence technology. The technology has been driven by applications in nuclear power plants [1]. The handling of hazardous material is less dangerous for the human operator if a telerobotic system enables a spatial separation of the human operator and the dangerous material. In general, terrestrial applications can be differentiated, depending on the type of the barrier between operator and the teleoperator, located in the remote environment. The handling of dangerous material as already mentioned demands matter as the barrier. If scale is the barrier, as it is for the case in minimal invasive surgery [2] or micro-assembly systems [3], the telepresence system enables the human operator to carry out macro-sized procedures on the operator side, by scaling them down to micro-sized movements on the teleoperator side. Well understood terrestrial applications, in which distance separates the operator from the teleoperator include search and rescue operations [4] and the exploration of deep underwater environments using so-called remotely operated vehicles (ROV) [5]. One of the key elements in modern telepresence systems is the implementation of multimodal feedback, which is the stimulation of more than just the visual sense of the human operator. The total sensorial immersion of the human operator into the remote environment with all sensory abilities is envisaged, which results in a high degree of so-called system transparency and consequently in a better task performance of the operator [6]. For that purpose haptic devices enjoy a great popularity since they allow feeding forces back to the operator. Commonly the positions (or forces) of the human operator are measured by a haptic input device or haptic display as shown in Figure 1. After being communicated via the communication channel, which bridges the addressed barrier in the system, the values are used as set-points for the teleoperator position (or force). Figure 1: General structure of a multimodal telepresence system

2 That way motions and manipulations are commanded from the operator site to the remote site. The resulting forces (or positions) with the remote environment are measured via sensors and fed back via the communication channel to the multimodal man-machine interface. Finally, diverse visual-acoustic displays provide feedback to the respective human sense. This paper emphasizes the benefits of using multimodal telepresence systems in space with a particular focus on in-space robotic assembly. Section 2 gives a short overview on teleoperated space systems with focus on on-orbit servicing. Next, section 3 clarifies how a multimodal feedback channel from space applications can enhance the mission performance. Section 4 depicts the development of a representative test environment, which will demonstrate the stated benefits. 2 On-orbit servicing systems The development of all existing free flyer systems as well as space systems with robotic manipulators in Earth orbit was mainly motivated by on-orbit servicing (OOS) purposes. OOS is basically a ground controlled modification of spacecraft in space. There is a great variety of application areas for OOS, which can be grouped into five main operations [7] as follows. (1) Assembly operations comprise the construction as well as the upgrade of a spacecraft in orbit. (2) Orbit transfer operations involve not only orbit correction procedures, but also the retrieval from an orbit to a graveyard orbit and the controlled deorbiting of a spacecraft. Further, (3) resupplying spacecraft with consumables, or components may enhance the life time of the target spacecraft. Typical applications for (4) maintenance and repair span the spectrum from inspecting the target spacecraft to major modifications to the system, and repair. (5) Special operations include emergency operations or the permanent use of the servicer satellite for attitude control of the target spacecraft. The following subsections briefly reviews missions and experiments that demonstrated various aspects of OOS. It differentiates in this connection between free flyers, which are space systems without the ability to modify a target spacecraft and systems with robotic manipulators (mounted on a free-flying base), which are used to interact with the target spacecraft. 2.1 Free Flyer The role of the free flyer in OOS operations is mostly concerned with proximity operations such as rendezvous and docking with the target satellite. It can also be utilized for inspecting the target satellite or monitoring OOS maneuvers of another servicer satellite with the target satellite. The Autonomous Extra-vehicular Robotic Camera (AERCam) [8] was designed to provide astronauts and ground control visual feedback of the International Space Station and the Space Shuttle. Intended to be used for remote inspections the prototype (AERCam Sprint) was first deployed during a Space Shuttle mission (STS 87) in Teleoperated by an astronaut inside the space shuttle it flew freely in the forward cargo bay. The Experimental Satellite System 10 (XSS-10) and 11 (XSS-11) were developed by the US Air Force [9], [10]. The mission was launched in 2003 and 2005, respectively and was intended to demonstrate key concepts and technologies relating autonomous satellite inspection operations. The spacecraft were attached to second stage of a Delta II rocket and a Minotaur upper stage, respectively. After ejecting, the stages were used to simulate the spacecraft to be inspected. The mission objectives included autonomous navigation and proximity operations. The 2005 Demonstration of Autonomous Rendezvous Technology (DART) [11] mission, developed by NASA, was launched to verify hardware and software for rendezvous and proximity operations. The main objectives were the demonstration of station keeping and collision avoidance maneuvers. However, when DART approached target satellite MUBLCOM, it overshot an important waypoint, and collided with it. A premature retirement of DART was the consequence. The Micro-Satellite Technology Experiment (MiTEx) is a US military project, which was successfully delivered into geostationary orbit in 2006 [12]. The two were developed to execute a variety of experiments in autonomous operations, maneuvering and station-keeping. In 2009 they conducted the first deep space inspection of a malfunctioning spacecraft through the Defense Support Program (DSP 23) missile warning satellite. The lessons learned from OOS missions that involve free flyers depict that they are mostly applied in operations where the target (satellite) is known in detail. That way most of the missions were very successful by using an autonomous approach which had been monitored from ground. 2.2 Robotic manipulators After the proximity operations are successfully fulfilled and the servicer satellite has docked with the target satellite, the objectives of robotic manipulators are to interact with the target and execute manipulations executed from ground. This can either be executed autonomously with the human operator only observing or teleoperated, with the human operator having an active role (cp. Figure 2) as the following OOS experiments show. The first robot in space, which had been controlled from ground was the Robot Technology Experiment (ROTEX) [13] aboard the space shuttle Columbia in The German experiment consisted of a six degrees of freedom (DOF) robot, featuring a multisensory gripper. ROTEX performed several tasks, such as assembling a truss structure and catching a free-floating object, in different levels of robot autonomy. Amongst others a teleoperation by a human operator from ground, using predictive computer graphics, was performed which enabled the operator on ground to successfully perform

3 sophisticated tasks despite a round trip delay of six seconds. The Japanese Engineering Test Satellite ETS-VII [14] consisted of a pair of satellites (servicer satellite, target satellite) it was launched in The servicer satellite featured a 6 DOF robotic arm, which was used to execute various OOS experiments with the target satellite in three different operational modes, that is telemanipulation mode, pre-programmed execution mode, and the real-time execution mode. Besides rendezvous and docking procedures, capturing, inspection and a series of manipulation operations were demonstrated. The experiments showed that a bilateral teleoperation is stably executable in space and the force feedback is helpful for the operator, even under six to seven seconds of round trip delay. Installed in 2005 outside the International Space Station (ISS) at the Russian service module [15], the Robotic Component Verification aboard the ISS (Rokviss) is a German experiment, featuring a two joint robotic manipulator, controlled by a human operator via a direct radio link from a ground station. In contrast to Rotex and ETS-VII, which have been controlled via geostationary satellites, typical round trip delays between operator action and Rokviss haptic-visual feedback are in the vicinity of 20 ms [16]. Orbital Express aimed at validating software for autonomous mission planning, rendezvous, proximity operations, and docking. Robotic OOS scenarios included fuel and electronics transfer, deployment of, and operations with a micro-satellite. Developed by Boeing and Ball Aerospace and launched in 2007, it was the first time a spacecraft was robotically transferring propellant and a battery to a target satellite [17]. Figure 2: Telepresence experiments via ARTEMIS Recent ARTEMIS experiments [18], involving the Institute of Astronautics (LRT) at Technical University of Munich, the German Aerospace Center (DLR), and the European Space Agency (ESA), have shown that it is possible for a human operator to steer virtual as well as robotic applications with multimodal feedback via a geostationary satellite as shown in Figure 2. In the framework of this telepresence experiments two ground stations (in Germany and Belgium, respectively) were connected via the ESA relay satellite ARTEMIS [19] in a way that robotic manipulations could be executed with real-time feedback for the human operator. The use of a geostationary satellite for data relay will drastically enhance the acquisition time of an OOS mission in low Earth orbit. That way it was proven that a telepresent OOS operations in Earth orbit are controllable by an operator on ground [20]. In summary, most of the missions involving robotic manipulators utilized the benefits of a human in the loop. It was repeatedly shown that the human operator is capable of executing OOS maneuvers from ground. 3 Benefits of multimodal telepresence Most of the mentioned OOS operations (cp. section 2) have either already been demonstrated in space or it is evident that they will be demonstrated in the foreseeable future. Here the concept of autonomy plays and will play a very important role, since a majority of procedures are predictable and thus programmable on ground. Nonetheless, telepresence techniques have to be considered as a valuable addition to OOS missions. A few missions have already applied the concept of telepresence control to space. The next subsections outline potential benefits and application areas of multimodal telepresence systems in space. 3.1 Areas of Application In the framework of the development of OOS missions machine pattern recognition, object tracking, and acquisition algorithms have been developed and enhanced in recent years. While the research is still in early stages and the algorithms have to be realized in very complex systems, the human eye-brain combination is already very evolved and trainable. Autonomous systems for guidance, navigation, and control are currently utilizing imaging technologies, which demand the availability of special surface marks on the target satellite or stored geometrical data, and perspectives of the target spacecraft. This is a precondition for calculating the servicer position and attitude with respect to the target satellite [21]. As a result only spacecraft that have been explicitly designed to be approachable and serviceable by another spacecraft can be supported with OOS operations. Further, since all autonomy procedures have to be preprogrammed on ground, it is only applicable to satellite failures, which are in the developed data bank and can exactly be predicted on ground 1. A further prerequisite is the knowledge of satellite geometry and dynamics 2, which 1 The causes of satellite failures cannot always be discerned from satellite telemetry. 2 A software update on-orbit is possible but requires additional attitude and orbit control procedures of the servicer satellite in the time where the software updates are developed and

4 can be very complicated in the presence of flexible appendages or fuel consumption and sloshing. The concept of telepresence utilizes the human operator s trained sensory abilities for recognition and decision-making. Procedures can be executed by the trained user from the ground. Unforeseen incidents can be solved with greater flexibility and robustness. Arbitrary spacecraft could be approached, i.e. spacecraft which were not explicitly designed for rendezvous and docking maneuvers. Further, the capability of a telepresent intervention at the remote environment can be helpful for emergency operations, e.g. in the case of malfunctioning autonomous collision avoidance maneuvers, for example, the DART mission. If the operator on Earth is capable of using an input device, such as a joystick, for sending real-time commands to the servicer satellite and to maneuver instantaneously away from a potential threat, a second (i.e. a manual) level of collision avoidance is always available. After maneuvering the satellite to a safe position and executing several predetermined checkout procedures, autonomous procedures could take again control of the satellite. Analogously, inspections and fly-arounds can be controlled by the human operator. This approach might be beneficial if the cause or the position of a satellite defect is not known beforehand. Based on the acquired information the human operator on the ground can decide how to proceed and which servicing measures to take. Another element in the decision queue is the path planning approach for the target satellite to the capture object. Mapping or remapping of the target satellite is another area of interest for a telepresently controlled inspector satellite. The geometry of the target can be efficiently acquired and used for proximity operations. Besides the stated benefits for the free flyer itself, multimodal telepresence can support robotic manipulations at the target satellite. In case autonomous operations cause the work area to exhibit an unknown and unforeseen state (e.g. when robotically exchanging or upgrading instruments) the human operator on ground can support the operations by either finishing the procedure or returning the system into a state which can be processed by autonomous procedures. The advantage of multimodal telepresence in this connection is the fact that the operator will not only see the remote side, but also feel it due to haptic displays. Although this paper concentrates on feedback from the haptic-visual workspace, it is worth mentioning other modalities such as auditory feedback. This can be artificially generated at the operator site if the robotic manipulator contacts the target satellite at the remote site. The applicability of the telepresence approach, with a human operator located in a ground station, controlling a spacecraft, is limited to the Earth orbit. This is because the round trip delay 3 increases with increasing distance from operator to the teleoperator. A decrease of the telepresence feeling is the consequence, which has a large impact on the task performance. 3.2 Haptic feedback for free flyers Unlike robotic manipulators, where haptic feedback plays an important role for telepresence control, free flyers are commonly only steered using visual feedback. That means that even though free flying experiments are commonly steered with hand controllers, as for example Scamp [22] or the Mini AERCam [23], usually no haptic information is fed back to the human operator. Figure 3: Example application of free flyers with haptic feedback The implementation of haptic feedback into the control of free flyers enriches the telepresence feeling of the operator and helps the operator on ground to navigate. It paves the way for new concepts of telepresent spacecraft control. Collision avoidance maneuvers for example can be made visible and perceptible for the human operator, by placing virtual walls around other spacecraft as shown in Figure 3. Equipping these virtual walls with sufficient high stiffness means that the operator is not able to penetrate them by means of the haptic device, since it exerts to the operator a high resistance force. Areas of fuel optimal paths can be displayed to the operator by implementing an ambient damping force, as depicted in Figure 3, featuring a magnitude which is proportional to the deviation of the actual path from the fuel optimal trajectory and area, respectively. Docking maneuvers can be supported by virtual boundaries as a haptic guiding cone and damping forces which are increasing with decreasing distance to the target. Further, since orbital maneuvers featuring a change of orbit extensively tested. This must be considered in the fuel and power budget of the servicer satellite. 3 The round trip delay is the time period between a telecommand sent and feedback received by an operator.

5 inclination are very fuel intensive, haptic boundaries could be set for fly-arounds around the target satellite which prevents the inspector from leaving the orbital plane. These are just a few examples for enriching the operator perception with haptic feedback and enhancing the task performance of telepresent OOS. Summarizing the benefits it can be seen that the application of telepresence control will extend the amount of serviceable spacecraft and spacecraft failures by involving a well trained human operator. In this connection it is proposed that the task performance of the operator can be enhanced by feeding back high-fidelity information from the remote work environment. Here the haptic feedback plays besides the visual feedback a very important role in human perception. Therefore, for an overall and significant evaluation of the benefits of multimodal telepresence a representative test environment is being developed at the MIT Space Systems Laboratory and is depicted in the next section. SPHERES is a risk-tolerant interactive testbed for telepresence formation flight and docking algorithms. Moreover, SPHERES allows for testing under offnominal conditions which can give vital validation information otherwise unachievable by on-orbit demonstrations without mission loss risk. 4.1 In-space Robotic Assembly Testbed Before applying the telepresence control to space it will extensively tested on ground. Thus, for initial ground tests, the SPHERES terrestrial 5-meter flat floor facility, a two-dimensional version of the same hardware used on the ISS, allows for initial testing and validation of the processes before attempting their implementation in the microgravity environment. This approach inherently reduces risks and allows for assessing repeatability while improving the reliability of the implemented process. 4 Development of a representative test environment The baseline of the test environment is the SPHERES (Synchronized Position Hold Engage Reorient Experimental Satellites) hardware which was developed at MIT Space System Laboratory and consists of a testbed on Earth and a test facility aboard the International Space Station (ISS) [24]. SPHERES, shown in Figure 4 comprises nano-satellites which are 20 cm in diameter. There are currently three 6 DOF free-flyer, self-contained SPHERES on the ISS. The navigation system consists of a custom pseudo-gps ultrasonic-based beacons and receivers located on the SPHERES and walls of the ISS nodes. Each SPHERE has a docking face with a beacon for direct range and bearing measurement, which can be used for relative state measurement. Figure 5: Two SPHERES docked using UDP A Universal Docking Port (UDP) has been implemented on the SPHERES to allow for satellite to satellite (Figure 5) and satellite to payload docking. Use of the UDP and SPHERES allows for an openarchitecture for testing the methods for robotic assembly of large space systems and structures. Figure 6: SPHERES with a docked flexible beam Figure 4: SPHERES inside ISS The benefit of using SPHERES to mature space telepresence technology is that the testbed presents the possibility of easy abort improve repeat approaches since astronauts can assist with tests. In this way, Current work has focused on assembly of a complex space structure with flexible dynamics. The three main technical demonstrations have been (1) control of a flexible beam structure, (2) docking control at both ends of a flexible structure, and (3) reconfiguration and control after docking and undocking. Figure 6 shows a SPHERES on a aircariage system on the flat floor with a docked flexible beam. Using an adaptive controller based on Shahravi the system demonstrated noncollocated docking

6 [25]. The basic steps for beam docking were first an approach maneuver, where the tug translates and rotates the flexible element into position at a safe distance from the target. Next is berthing, where the tug maneuvers close to the target, dampens out vibrations, and executes fine alignment. The final step is capture, where a terminal thrust command is issued to establish a closing velocity between the tug and target at the docking port. Current algorithms use autonomous navigation and control as a baseline method and to demonstrate initial functionality of the hardware-in-the-loop. The next step to be undertaken is to modify the testbed so it can be controlled via a haptic device in order to show that autonomous missions can benefit from telepresence operations. The intended modifications are depicted in the next section. states of the haptic device as well as feeds calculated forces back to it. The received positions will be communicated to a SIMULINK thruster model of the SPHERES spacecraft. The thruster model will calculate forces and torques 4 in accordance with the satellite intrinsic properties. These forces and torques will be evaluated by an estimator using knowledge on the satellite dynamics and returning estimations of the satellite states to the remote environment. This remote workspace will be created using SIMULINK s Virtual Reality (VR) toolbox, allowing for satellite states and environmental properties to be displayed. In addition to the Matlab environment (cp. Figure 7) algorithms in C will be used to create interfaces to the actual SPHERES hardware, described in the previous section (cp. Figure 6). This will be used to transmit Figure 7: Block diagram of the test environment 4.2 Implementing a haptic device and virtual reality into the ISRA test bed The key element, which has to be implemented into the test environment is the Novint Falcon [26], which is a 3 DOF force feedback joystick (as can be seen in Figure 7). All degrees of freedom are of translational nature and servo motors are used to feed forces in three degrees of freedom back to the user. This system has high utility for space applications since it allows the human operator to control the application in three dimensional space. The Falcon will be implemented in a Matlab/SIMULINK environment via the HaptikLibrary [27], which is a component based architecture for uniform access to haptic devices. It is currently used as the interface to Matlab and reads the positions and button position commands via the existing communications system to the SPHERES internal model of its thrusters. Again torques and forces are calculated and directly commanded to the thrusters, which will cause a motion of the SPHERE in two dimensions with three degrees of freedom. Distance information is available due to ultrasound beacons which border the test area. This will be combined with information from the onboard gyroscopes to metrology data, which is necessary for the position and attitude determination system to compute the 4 The degrees of freedom of the Novint Falcon are only translational but it is intended to implement rotational commands by using the button states of the haptic device in connection with certain translational motion.

7 SPHERES state vector via the use of extended Kalman filters [28]. By transmitting the actual states to the VR, the operator obtains information of the estimated and the actual motion of the system, which should be identical if the communication channel is not delayed. In the presence of time delay the predictions should give the user a feeling for the behaviour of the system (e.g. similar to ETS-VII) and enhance the task performance. This is an important point if a human operator on ground will steer an application in space. That way the interactions between the autonomous in-space robotic assembly and a telepresently controlled free flyer can be tested. 4.3 Envisaged experiments After the interfaces between the haptic device and the testbed are optimized, various experiments are planned. Inspection The first and least complex experiment is to use the telepresently steered free flyer to inspect an autonomous assembly procedure. The collision avoidance approach and the fuel optimal paths, described in section 3 can for example be tested in that way. Assembly The following step will be to use the Novint Falcon to test a human assisted assembly procedure. Proximity and docking maneuver can be implemented in that connection. Time delayed operations If a human operator controls an application in space, the system is subject to time delay. Thus, a control architecture has to be developed which works under this delay.tthere exists several approaches found in literature which can be leveraged [29]. For this purpose time delay has to be introduced in the system and the response has to evaluated in experiments. Transparency evaluation However, the addressed controllers cannot preserve system transparency and system stability to the same degree in the presence of time delay [30]. Thus, tradeoffs between stability and transparency have to be sought and evaluated. There is no all-embracing concept for evaluating the telepresence capability or transparency of a teleoperated system in space. One possibility is to conduct a series of tests with human participants and log their task performance with respect to the properties of the system. Another approach is approximating the haptic interface as a device which generates mechanical impedance [31]. It represents the dynamic relationship between displacement (or velocity) and force. An ideal haptic interface would be capable of generating any impedance, which is required to represent the remote environment realistically to the human operator. Accordingly, transparency can be considered as the accuracy in rendering the remote environment to the human operator [30]. For evaluating the performance of a teleoperated system, the Z-width concept can be used. It is defined as the dynamic range of achievable impedance, which the system can stably represent to the human operator and is accordingly limited by the maximum and minimum achievable impedance [32]. Path planning A further evaluation, that has to be conducted, is the influence of the human in the loop on existing path planning algorithms. If a human operator is supporting autonomous OOS operations, the system has to take into account that the motions of the ground controlled spacecraft are not predictable in the same way as autonomous are. Sudden changes in spacecraft direction can occur due to the human operator, which are not common for an efficient autonomous path planner. After finishing the addressed initial evaluation the test environment will be used for specific tests in the framework of in-space robotic assembly, such as inspection or the actual assembly procedures. 5 Conclusions and future work The research on autonomous servicing operations in space is still in the early stages. Up to date it is not possible to service an arbitrary spacecraft in orbit. There is utility for human control of the servicer satellite from the ground and should be considerationed for OOS operations. Thus, the benefits of such telepresent on-orbit servicing operations in addition to autonomy have been emphasized in this paper. This approach allows an instantaneous intervention from the ground with the remote environment in space and provides a large flexibility to mission operations. In particular the advantages of multimodal telepresence feedback for free flyers have been depicted, which is a new research area in astronautics. Based on the potential benefits, a test environment for multimodal telepresent in-space robotic assembly is proposed in this paper, which is based on the existing SPHERES testbed. This testbed has not only been extensively tested on ground but also in space on the ISS. While the C environment of the proposed testbed already exists (due to the SPHERES test campaign), the Matlab/SIMULINK environment and interfaces between both are currently being developed and optimized for the use with the Novint Falcon haptic device. Using the SPHERES in 1-g limits the tests to only feature three degrees freedom, that is two translational and one rotational. Nonetheless, the test environment will give significant information on the feasibility and the advantages of multimodal telepresence in space. Acknowledgements This work was supported in part by post-doctoral fellowship programs of the German Academic Exchange

8 Service (DAAD) and the German Research Foundation (DFG). The authors would like to thank Prof. David W. Miller, Dr. Alvar Saenz-Otero, and Jacob Katz for collaboration on this research. References [1] T. B. Sheridian, Telerobotics, automation, and human supervisory control ( Cambridge, USA: MIT Press, 1992) [2] T. Ortmaier et al., Robot Assisted Force Feedback Surgery, in Advances in Telerobotics (Heidelberg, Germany: Springer, 2007) [3] M. Zaeh, S. Clarke, B. Petzold, J. Schilp, Achieving flexible micro-assembly systems through telepresence, proc. of the International Conference in Mechatronics and Robotics. IEEE Industrial Electronics Society APS, Aachen, Germany, 2004 [4] N. Ruangpayoongsak, H. Roth, J. Chudoba, Mobile robots for search and rescue, IEEE International Safety, Security and Rescue Robotics Workshop, Kobe, Japan, Jun [5] P. Ridao, M.Carreras, E. Hernandez, N. Palomeras, Underwater Telerobotics for Collaborative Research, in Advances in Telerobotics (Heidelberg, Germany: Springer, 2007) [6] H. Pongraz, Gestaltung und Evaluation von virtuellen und Telepräsenzsystemen an Hand von Aufgabenleistung und Präsenzempfinden, Dissertation, Human Factors Institute, Universität der Bundeswehr, 2008 [7] D.Waltz On-Orbit Servicing of Space Systems (Malabar, USA: Krieger Publishing. Co., 1993) [8] Choset, H., and Kortenkamp, D., Path Planning and Control for AERCam a Free-Flying Inspection Robot in Space, Journal of Aerospace Engineering, 12(2), 1999, [9] T.M. Davis, XSS-10 Micro-Satellite Flight Demonstration, proc. of Georgia Institute of Technology Space Systems Engineering Conference, paper no. GT-SSEC.D.3, Nov [10] XSS-11 micro satellite fact sheet, Air Force Research Laboratory, Space Vehicles Directorate, doc. no. AFD , Apr [11] T.E. Rumford, Demonstration of Autonomous Rendezvous Technology (DART) Project Summary, proc. of Space Systems Technology and Operations Conference, Orlando, USA, Apr [12] Microsatellite Launch, Crosslink- The Aerospace Corporation magazine of advances in aerospace technology, 7(2), 2006, 2 [13] G. Hirzinger, K. Landzettel, C. Fagerer, Telerobotics with large time delays-the ROTEX experience, proc. of the IEEE/RSJ/GI International Conference on Intelligent Robots and Systems, Munich, Germany, Sep [14] T. Imaida, Y. Yokokohji, T. Doi, M. Oda, and T. Yoshikawa, Ground-Space Bilateral Teleoperation Experiment Using ETS-VII Robot Arm with Direct Kinesthetic Coupling, proc. of IEEE International Conference on Robotics and Automation, Seoul, Korea, [15] K. Landzettel et al., Robotic On-Orbit Servicing DLR s Experience and Perspective, proc. of. the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems, Beijing, China, Oct [16] C. Preusche, D. Reintsema, K. Landzettel, G. Hirzinger, Robotics Component Verification on ISS ROKVISS Preliminary Results for Telepresence, proc. of. the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems, Oct. 2006, Beijing, China [17] J. Shoemaker, M. Wright, Orbital Express Space Operations Architecture Program, Proceedings of SPIE Spacecraft Platforms and Infrastructure Symposium, Orlando, USA, [18] E. Stoll, Ground Verification of Telepresence for On-Orbit Servicing, Dissertation, Lehrstuhl für Raumfahrttechnik, Technische Universität München, (Munich, Germany: Verlag Dr. Hut 2008). [19] T. Jono et al., Demonstrations of ARTEMIS-OICETS Inter-Satellite Laser Communications, proc. of 24th AIAA International Communications Satellite Systems Conference, San Diego, USA, [20] E. Stoll et al., Ground Verification of the Feasibility of Telepresent On-Orbit Servicing, Journal of Field Robotics, 26(3), 2009, [21] M.E. Polites, An Assessment of the Technology of Automated Rendezvous and Capture in Space, NASA/TP , 1998 [22] C. McGhan, R. Besser, R. Sanner, and E. Atkins, Semi- Autonomous Inspection with a Neutral Buoyancy Free-Flyer, proc. of Guidance, Navigation, and Control Conference, Keystone, USA, Aug [23] S. Fredrickson, S. Duran, J. Mitchel, Mini AERCam Inspection Robot for Human Space Missions, proc. of AIAA Space, San Diego, USA, Sep [24] Miller, D.W., Kong, E.M.C., & Saenz-Otero, A., Overview of the SPHERES Autonomous Rendezvous & Docking Laboratory on the International Space Station, 26th Annual AAS Guidance & Control Conference, Breckenridge, CO, Feb, [25] Shahravi, M., and Kabganian, M., Attitude tracking and vibration suppression of flexible spacecraft using implicit adaptive control law, American Control Conference, 2005, Proceedings of the 2005, pp , vol. 2, June [26] Novint Technologies, Inc., cited Jun [27] M. de Pascale, D. Prattichizzo, The Haptik Library: a component based architecture for uniform access to haptic devices, IEEE Robotics and Automation Magazine, 14(4), 2007, [28] Nolet, S., The SPHERES Navigation System: from Early Development to On-Orbit Testing, AIAA Guidance, Navigation and Control Conference and Exhibit, Hilton Head, USA, Aug [29] G. Niemeyer, C. Preusche, G. Hirzinger, Telerobotics in Handbook of Robotics (Berlin, Germany: Springer, 2008) [30] K. Hashtrudi-Zaad, S.E. Salcudean, Transparency in Time-Delayed Systems and the Effect of Local Force Feedback for Transparent Teleoperation, IEEE Transactions on Robotics and Automation, 18(4), Feb [31] J.E. Colgate, J.M. Brown, Factors Affecting the Z-Width of a Haptic Display, proc. of the IEEE In-ternational Conference on Robotics & Automation, San Diego, USA, May [32] B. Khademian, K. Hashtrudi-Zaad, Kinesthetic performance analysis of dual-user teleoperation systems, proc. of IEEE International Conference on Systems, Man, and Cybernetics, Montreal, Canada, Oct