Simon Nolet

Flight Testing of Multiple-Spacecraft Control on SPHERES During Close-Proximity Operations

A, B, C = state-space matrices a = acceleration due to linear-quadratic-regulatorand artificial-potential-field-determined control effort aAPF = acceleration due to artificial-potential-fielddetermined control effort aLQR = acceleration due to linear-quadratic-regulatordetermined control effort am = maximum acceleration aobs = acceleration of chaser spacecraft toward an obstacle ax;y;z = acceleration due to the control effort Do = obstacle region of influence da = goal acceleration decay constant dg = goal exponential decay constant do = stopping distance constant JLQR = linear quadratic regulator cost function KLQR = linear quadratic regulator state feedback gain ka = acceleration shaping parameter kd = docking safety parameter kg = velocity shaping function ko = obstacle function ks = safety function kv = velocity shaping parameter Lo = obstacle exterior surface N = linear quadratic regulator gain matrix Q = linear quadratic regulator state gain matrix R = linear quadratic regulator control effort gain matrix r = Euclidean norm distance or relative range r = relative distance vector rc = position vector of the chaser spacecraft rg = position vector of the chaser spacecraft from the goal rinit = initial distance of the chaser spacecraft from the goal rm = maximum allowable distance of the chaser spacecraft from the goal ro = position vector of the chaser spacecraft from the obstacle rt = position vector of the target spacecraft with respect to the Earth S = solution of the Riccati equation u = control effort vector V = potential function Vg = goal potential function Vo = obstacle potential function vm = maximum relative velocity vo = desired velocity of chaser spacecraft toward an obstacle vobs = velocity of chaser spacecraft toward an obstacle x = state vector x, y, z = positions, or states, along the Cartesian axis Q = linear quadratic regulator state performance gain R = linear quadratic regulator control effort gain t = time increment = standard deviation for the obstacle’s region of influence ! = orbital angular velocity

Autonomous docking experiments using the SPHERES testbed inside the ISS

This paper presents recent results regarding the research on autonomous docking with tumbling targets performed by the MIT Space Systems Laboratory (SSL). The objective of this research is to develop a guidance, navigation and control (GN&C) architecture that enables safe and fuel efficient docking of a thruster-based spacecraft with a tumbling target in the presence of obstacles and contingencies. Over the calendar year 2006, experiments were performed inside the International Space Station (ISS) using the SPHERES nano-satellites to validate a GN&C architecture on hardware in microgravity. A series of attitude slews, an autonomous docking maneuver with a fixed beacon and a station-keeping maneuver were among the experiments carried out in May to validate subsets of the architecture with only a fraction of the SPHERES hardware. The second set of experiments occurred in August and involved two satellites and the remaining navigation hardware. The global estimator allowing the SPHERES to navigate within the US Laboratory was validated. Multiple successful docking maneuvers between two satellites were also accomplished. In November, more complex docking scenarios were experimented, leading to the first successful autonomous docking with a tumbling target ever performed in microgravity. Results collected during key ISS experiments are presented in this paper.

Ground and Space Testing of Multiple Spacecraft Control During Close-Proximity Operations

A multiple spacecraft close-proximity control algorithm was implemented and tested with the Synchronized Position Hold Engage and Reorient Experimental Satellites (SPHERES) facility onboard the International Space Station (ISS). During flight testing, a chaser satellite successfully approached a virtual target satellite, while avoiding collision with a virtual obstacle satellite. This research contributes to the control of multiple spacecraft for emerging missions, which may require simultaneous gathering, rendezvous, and docking. The unique control algorithm was developed at NPS and integrated onto the MIT SPHERES facility. The control algorithm implemented combines the efficiency of the Linear Quadratic Regulator (LQR), and the robust collision avoidance capability of the Artificial Potential Function method (APF). The LQR control effort serves as the attractive force toward goal positions, while the APF-based repulsive functions provide collision avoidance for both fixed and moving obstacles. The amalgamation of these two control methods into a multiple spacecraft close-proximity control algorithm yielded promising results as demonstrated by simulations performed at NPS. Comprehensive simulation evaluation enabled implementation and testing of the spacecraft control algorithm on the SPHERES facility at MIT. Finally, successful ground testing enabled execution of flight testing onboard the ISS. The NPS’s Spacecraft Robotics Laboratory (SRL) and MIT’s Space Systems Laboratory (SSL) simulations, the MIT’s SSL SPHERES ground testing, and the SPHERES flight testing results are all presented in this paper.

Development and verification of algorithms for spacecraft formation flight using the SPHERES testbed: application to TPF

The MIT Space Systems Laboratory and Payload Systems Inc. has developed the SPHERES testbed for NASA and DARPA as a risk-tolerant medium for the development and maturation of spacecraft formation flight and docking algorithms. The testbed, which is designed to operate both onboard the International Space Station and on the ground, provides researchers with a unique long-term, replenishable, and upgradeable platform for the validation of high-risk control and autonomy technologies critical to the operation of distributed spacecraft missions such as the proposed formation flying interferometer version of Terrestrial Planet Finder (TPF). In November 2003, a subset of the key TPF-like maneuvers has been performed onboard NASAs KC-135 microgravity facility, followed by 2-D demonstrations of two and three spacecraft maneuvers at the Marshall Space Flight Center (MSFC) in June 2004. Due to the short experiment duration, only elements of a TPF lost in space maneuver were implemented and validated. The longer experiment time at the MSFC flat-floor facility allows more elaborate maneuvers such as array spin-up/down, array resizing and array rotation be tested but in a less representative environment. The results obtained from these experiments are presented together with the basic estimator and control building blocks used in these experiments.

Design of an algorithm for autonomous docking with a freely tumbling target

For complex unmanned docking missions, limited communication bandwidth and delays do not allow ground operators to have immediate access to all real-time state information and hence prevent them from playing an active role in the control loop. Advanced control algorithms are needed to make mission critical decisions to ensure safety of both spacecraft during close proximity maneuvers. This is especially true when unexpected contingencies occur. These algorithms will enable multiple space missions, including servicing of damaged spacecraft and missions to Mars. A key characteristic of spacecraft servicing missions is that the target spacecraft is likely to be freely tumbling due to various mechanical failures or fuel depletion. Very few technical references in the literature can be found on autonomous docking with a freely tumbling target and very few such maneuvers have been attempted. The MIT Space Systems Laboratory (SSL) is currently performing research on the subject. The objective of this research is to develop a control architecture that will enable safe and fuel-efficient docking of a thruster based spacecraft with a freely tumbling target in presence of obstacles and contingencies. The approach is to identify, select and implement state estimation, fault detection, isolation and recovery, optimal path planning and thruster management algorithms that, once properly integrated, can accomplish such a maneuver autonomously. Simulations and demonstrations on the SPHERES testbed developed by the MIT SSL will be executed to assess the performance of different combinations of algorithms. To date, experiments have been carried out at the MIT SSL 2-D Laboratory and at the NASA Marshall Space Flight Center (MSFC) flat floor.

Docking and reconfiguration of modular spacecraft: preliminary SWARM testing at MSFC

On-orbit servicing and assembly is a critical enabling technology for the advancement of large scale structures in space. The goal of the SWARM project (Synchronized Wireless Autonomous Reconfigurable Modules) is to develop and mature algorithms for autonomous docking and reconfiguration, to be used as the building blocks for autonomous servicing and assembly. Algorithms for approach, docking, and reconfiguration have been implemented and tested through a demonstration of the assembly of two telescope sub-apertures at Marshall Space Flight Center (MSFC) in July 2006. The algorithms developed for reconfiguration set the mass properties based on the configuration. Updatable parameters include the location of sensors and receivers with respect to the geometric center, thruster locations, and control gains specific to each configuration. To test these algorithms in a 2D environment, a ground testbed was developed to provide multiple docking ports and modular payload attachments. Hardware components include nodes, Universal Docking Ports, posts, sub-aperture mirrors, and a SPHERES satellite as the assembler tug. Testing at MSFC successfully demonstrated relative docking and reconfiguration. Valuable information was gained about the performance of the docking under friction, sensitivity to estimator initialization, thrust authority needed for different phases of the test, and control when CM changes during the test.

Development of the miniature video docking sensor

To perform realistic demonstrations of autonomous docking maneuvers using micro-satellites, the MIT Space Systems Laboratory (SSL) developed a miniature universal docking port along with an optical sensing system for relative state estimation. The docking port has an androgynous design and is universal since any two identical ports can be connected together. After a rigid connection is made, it is capable of passing electrical loads between the connected micro-satellites. The optical sensor uses a set of infrared LEDs, a miniature CCD-based video camera, and an Extended Kalman Filter to determine the six relative degrees of freedom of the docking satellite. The SPHERES testbed, also developed by the MIT SSL, was used to demonstrate the integrated docking port and sensor system. This study focuses on the development of the optical docking sensor, and presents test results collected to date during fully autonomous docking experiments performed at the MIT SSL 2-D laboratory. Tests were performed to verify the validity of the docking sensor by taking measurements at known distances. These results give an estimate of the sensor accuracy, and are compared with a theoretical model to understand the sources of error in the state measurements.