ALBERT

Description: The Kennedy Space Center has significant infrastructure for research using controlled environment plant growth chambers. Such research supports development of bioregenerative life support technology for long-term space missions. Most of the existing chambers in Hangar L and Little L will be moved to the new Space Experiment Research and Processing Laboratory (SERPL) in the summer of 2003. The impending move has created an opportunity to update the control system technologies to allow for greater flexibility, less labor for set-up and maintenance, better diagnostics, better reliability and easier data retrieval. Part of these improvements can be realized using hardware which communicates through an ethernet connection to a central computer for supervisory control but can be operated independently of the computer during routine run-time. Both the hardware and software functionality of an envisioned system were tested on a prototype plant growth chamber (CEC-4) in Hangar L. Based upon these tests, recommendations for hardware and software selection and system design for implementation in SERPL are included.

Description: This paper presents two probabilistic developments for use with Electromyograms (EMG). First described is a new-electric interface for virtual device control based on gesture recognition. The second development is a Bayesian method for decomposing EMG into individual motor unit action potentials. This more complex technique will then allow for higher resolution in separating muscle groups for gesture recognition. All examples presented rely upon sampling EMG data from a subject's forearm. The gesture based recognition uses pattern recognition software that has been trained to identify gestures from among a given set of gestures. The pattern recognition software consists of hidden Markov models which are used to recognize the gestures as they are being performed in real-time from moving averages of EMG. Two experiments were conducted to examine the feasibility of this interface technology. The first replicated a virtual joystick interface, and the second replicated a keyboard. Moving averages of EMG do not provide easy distinction between fine muscle groups. To better distinguish between different fine motor skill muscle groups we present a Bayesian algorithm to separate surface EMG into representative motor unit action potentials. The algorithm is based upon differential Variable Component Analysis (dVCA) [l], [2] which was originally developed for Electroencephalograms. The algorithm uses a simple forward model representing a mixture of motor unit action potentials as seen across multiple channels. The parameters of this model are iteratively optimized for each component. Results are presented on both synthetic and experimental EMG data. The synthetic case has additive white noise and is compared with known components. The experimental EMG data was obtained using a custom linear electrode array designed for this study.

Description: Smart SPHERES is a prototype free-flying space robot based on the SPHERES platform. Smart SPHERES can be remotely operated by astronauts inside a spacecraft, or by mission controllers on the ground. We developed Smart SPHERES to perform a variety of intravehicular activities (IVA), such as operations inside the International Space Station (ISS). These IVA tasks include environmental monitoring surveys (radiation, sound levels, etc.), inventory, and mobile camera work. In this paper, we first discuss the motivation for free-flying space robots. We then describe the development of the Smart SPHERES prototype, including avionics, software, and data communications. Finally, we present results of initial flight tests on-board the ISS.

Description: Robonaut, the humanoid robot developed at the Dexterous Robotics Laboratory at NASA Johnson Space Center serves as a testbed for human-robot collaboration research and development efforts. One of the primary efforts investigates how adjustable autonomy can provide for a safe and more effective completion of manipulation-based tasks. A predictive algorithm developed in previous work was deployed as part of a software interface that can be used for long-distance tele-operation. In this paper we provide the details of this algorithm, how to improve upon the methods via optimization, and also present viable alternatives to the original algorithmic approach. We show that all of the algorithms presented can be optimized to meet the specifications of the metrics shown as being useful for measuring the performance of the predictive methods. Judicious feature selection also plays a significant role in the conclusions drawn.

Description: Astrobee is a new class of free-flying robots that operate in the interior of the International Space Station (ISS). In addition to providing a research platform for zero-g free-flying robotics (replacing the Synchronized Position Hold Engage and Reorient Experimental Satellites (SPHERES)), Astrobee improves the efficiency of ISS operations by providing flight and payload controllers with a mobile camera/sensor platform. The Astrobee system consists of three robots, a docking station, and a ground data system. Development began in late 2014, and Astrobee will launch to ISS in late 2018. This paper provides an overview of the capabilities of the Astrobee system, the concept of operations for ISS operations support, and the challenges of operating a free-flying robot within the constraints of the ISS environment.

Description: The FootFall Planning System is a ground-based planning and decision support system designed to facilitate the control of walking activities for the ATHLETE (All-Terrain Hex-Limbed Extra-Terrestrial Explorer) family of robots. ATHLETE was developed at NASA's Jet Propulsion Laboratory (JPL) and is a large six-legged robot designed to serve multiple roles during manned and unmanned missions to the Moon; its roles include transportation, construction and exploration. Over the four years from 2006 through 2010 the FootFall Planning System was developed and adapted to two generations of the ATHLETE robots and tested at two analog field sites (the Human Robotic Systems Project's Integrated Field Test at Moses Lake, Washington, June 2008, and the Desert Research and Technology Studies (D-RATS), held at Black Point Lava Flow in Arizona, September 2010). Having 42 degrees of kinematic freedom, standing to a maximum height of just over 4 meters, and having a payload capacity of 450 kg in Earth gravity, the current version of the ATHLETE robot is a uniquely complex system. A central challenge to this work was the compliance of the high-DOF (Degree Of Freedom) robot, especially the compliance of the wheels, which affected many aspects of statically-stable walking. This paper will review the history of the development of the FootFall system, sharing design decisions, field test experiences, and the lessons learned concerning compliance and self-awareness.

Description: We address the development of predictive tele-operator interfaces for humanoid robots with respect to two basic challenges. Firstly, we address automating the transition from fully tele-operated systems towards degrees of autonomy. Secondly, we develop compensation for the time-delay that exists when sending telemetry data from a remote operation point to robots located at low earth orbit and beyond. Humanoid robots have a great advantage over other robotic platforms for use in space-based construction and maintenance because they can use the same tools as astronauts do. The major disadvantage is that they are difficult to control due to the large number of degrees of freedom, which makes it difficult to synthesize autonomous behaviors using conventional means. We are working with the NASA Johnson Space Center's Robonaut which is an anthropomorphic robot with fully articulated hands, arms, and neck. We have trained hidden Markov models that make use of the command data, sensory streams, and other relevant data sources to predict a tele-operator's intent. This allows us to achieve subgoal level commanding without the use of predefined command dictionaries, and to create sub-goal autonomy via sequence generation from generative models. Our method works as a means to incrementally transition from manual tele-operation to semi-autonomous, supervised operation. The multi-agent laboratory experiments conducted by Ambrose et. al. have shown that it is feasible to directly tele-operate multiple Robonauts with humans to perform complex tasks such as truss assembly. However, once a time-delay is introduced into the system, the rate of tele\ioperation slows down to mimic a bump and wait type of activity. We would like to maintain the same interface to the operator despite time-delays. To this end, we are developing an interface which will allow for us to predict the intentions of the operator while interacting with a 3D virtual representation of the expected state of the robot. The predictive interface anticipates the intention of the operator, and then uses this prediction to initiate appropriate sub-goal autonomy tasks.

Description: We describe the development and usage of a streaming data analysis software framework. The framework is used for three different applications: Earth science hyper-spectral imaging analysis, Electromyograph pattern detection, and Electroencephalogram state determination. In each application the framework was used to answer a series of science questions which evolved with each subsequent answer. This evolution is summarized in the form of lessons learned.

Description: This paper describes and evaluates one method of modeling compliance in a wheel-on-leg walking robot. This method assumes that all of the robot s compliance takes place at the ground contact points, specifically the tires and legs, and that the rest of the robot is rigid. Optimization is used to solve for the displacement of the feet and of the center of gravity. This method was tested on both robots of the ATHLETE family, which have different compliance. For both robots, the model predicts the sag of points on the robot chassis with an average error of about one percent of the height of the robot.

Description: The President s Vision for Space Exploration, laid out in 2004, relies heavily upon robotic exploration of the lunar surface in early phases of the program. Prior to the arrival of astronauts on the lunar surface, these robots will be required to be controlled across space and time, posing a considerable challenge for traditional telepresence techniques. Because time delays will be measured in seconds, not minutes as is the case for Mars Exploration, uploading the plan for a day seems excessive. An approach for controlling dexterous robots under intermediate time delay is presented, in which software running within a ground control cockpit predicts the intention of an immersed robot supervisor, then the remote robot autonomously executes the supervisor s intended tasks. Initial results are presented.