A hybrid algorithm for real-time frame-to-frame pose estimation during monocular vision-only SLAM/SFM is presented. The algorithm combines concepts from two existing approaches to pose tracking, Bayesian estimation methods and measurement inversion techniques, to achieve in real-time a feasible, smooth estimate of the relative pose between a robotic platform and a tumbling target. It is assumed that no a priori information about the target is available, and that only a monocular camera is available for measuring the relative motion of the target with respect to the robotic platform. The rationale for a hybrid approach is explained, and an algorithm is presented. A specific implementation using a modified Rao-Blackwellised particle filter is described and tested. Results from both numerical simulations and field experiments are included which demonstrate the performance and viability of the hybrid approach. The hybrid approach to pose estimation described here is applicable regardless of the method by which the map/reconstruction is estimated.

Limits on the technology available to marine scientists to study jellyfish in situ motivate the development of an automated robotic tracking system for deployment in the deep ocean. This paper synthesizes a control strategy for robotic jellyfish tracking with a remotely operated vehicle (ROV). The control strategy employs three feedback loops tailored to the jelly-tracking task. A primary loop provides moderate proportional-derivative feedback to track animal swimming motion without producing excessive hydrodynamic disturbances, which might impact the behavior of the animal under study. A second boundary-control loop provides aggressive thrust, in a direction away from the target and only when needed to prevent loss of the target outside the boundaries of the vision sensor. A third disturbance-accommodation loop counters low-frequency bias forces without fighting commands issued by the human pilot. The complete system was implemented and tested in the Monterey Bay using the Monterey Bay Aquarium Research Institute's ROV Ventana. The control system autonomously tracked six animals for durations longer than 15 min, including a Plychogena medusa which was tracked for 89 min

This paper presents a general framework for planning the quasi-static motion of a three-limbed climbing robot in vertical natural terrain. The problem is to generate a sequence of continuous one-step motions between consecutive holds that will allow the robot to reach a particular goal hold. A detailed algorithm is presented to compute a one-step motion considering the equilibrium constraint only. The overall framework combines this local planner with a heuristic search technique to generate a complete plan. An online implementation of the algorithm is demonstrated in simulation.

A new motion planning framework is presented that enables multiple mobile robots with limited ranges of sensing and communication to maneuver and achieve goals safely in dynamic environments. To combine the respective advantages of centralized and de-centralized planning, this framework is based on the concept of centralized planning within dynamic robot networks. As the robots move in their environment, localized robot groups form networks, within which world models and robot goals can be shared. Whenever a network is formed, new information then becomes available to all robots in this network. With this new information, each robot uses a fast, centralized planner to compute new coordinated trajectories on the fly. Planning over several robot networks is decentralized and distributed. Both simulated and real-robot experiments have validated the approach.

The performance of monocular vision based target tracking is a strong function of camera motion. Without motion, the target estimation problem is unsolvable. By designing the camera path, the best possible estimator performance can be achieved. This paper describes a trajectory design method based on the predicted target state error covariance. This method uses a pyramid, breadth-first search algorithm to generate real-time paths that achieve a minimum uncertainty bound in fixed time or a desired uncertainty bound in minimum time.

A method of planning paths for formations of mobile robots with nonholonomic constraints is presented. The kinematics equations presented in this paper allow a general geometrical formation of mobile robots to be maintained while the group as a whole travels an arbitrary path. It is possible to represent a formation of mobile robots by a single entity with the same type of nonholonomic constraint as the individual members. Thus, any path-planner or control method may be used with the formation as would be applied to an individual robot. Equations are developed for changing the geometrical formation and hardware results are presented from the Stanford MARS Testbed.

Marine biologists desire an automated system for observing gelatinous zooplankton. This study represents the first attempt at real-time robotic tracking of a gelatinous animal in the open ocean. Challenges stem from uneven lighting, encounters with other marine life and frequent out-of-frame events. Analysis of a large video library of human-controlled jelly tracking supported selection of an appropriate vision algorithm. The vision architecture was implemented as an ROV pilot aid and tested at 600 m below the surface of Monterey Bay.

This paper presents a randomized motion planner for kinodynamic asteroid avoidance problems, in which a robot must avoid collision with moving obstacles under kinematic, dynamic constraints and reach a specified goal state. Inspired by probabilistic-roadmap (PRM) techniques, the planner samples the state x time space of a robot by picking control inputs at random in order to compute a roadmap that captures the connectivity of the space. However, the planner does not precompute a roadmap as most PRM planners do. Instead, for each planning query, it generates, on the fly, a small roadmap that connects the given initial and goal state. In contrast to PRM planners, the roadmap computed by our algorithm is a directed graph oriented along the time axis of the space. To verify the planner's effectiveness in practice, we tested it both in simulated environments containing many moving obstacles and on a real robot under strict dynamic constraints. The efficiency of the planner makes it possible for a robot to respond to a changing environment without knowing the motion of moving obstacles well in advance.

This paper derives a decentralized object impedance controller (DOIC) especially suited for use by a team of robots and/or multiple manipulators. In contrast to the original object impedance controller (OIC), a separate DOIC operates on each robot to provide local computation of the manipulator force commands, thus removing the need for expensive high-bandwidth communication of force signals between the robots. The key advance of the DOIC that provides this benefit is a decentralized algorithm for estimating the external force on the object. This paper derives the DOIC, provides stability analysis, and verifies the theory with definitive physical experiments.

Underwater robotic systems require a control architecture that handles automatic control aspects and discrete-event management while considering real-time issues within a formalized framework in order to achieve a high degree of autonomy and robustness while performing missions. To extract generic characteristics of working implementations, we examine two methodologies developed independently by French and American research organizations to control their underwater robots. In particular, we study how a specific reactive and complex mission is encoded under the two different methodologies. Details of the methodologies along with mission programs created by each team to accomplish the evaluation mission are presented.

This paper presents an approach for the symbolic dynamic modelling of object/robot-team systems composed of an object manipulated by a team of r robots. The modelling approach merges the dynamic models of the object and robots into a system model. Derivations show that the system acceleration can be computed with complexity proportional to r. This paper demonstrates in a detailed example with physical experiments how the modelling approach can be used for symbolic analysis of a closed-loop object/robot-team system.

This paper examines the stability of the flexible-object impedance controller when coupled to an arbitrary passive environment. A simple representative system is developed to study the problem based on phenomena observed in a more complex, experimental system. Analysis of this representative system leads to several conclusions regarding factors that limit the achievable impedances of a multi-arm system grasping a common object. The filtering of the external force signals and the effective mass contributed by the manipulators prove to be two of the important factors. Some general guidelines are developed for the successful application of flexible-object impedance control to insure coupled system stability. These guidelines are applied to both the representative and experimental systems and verified.

This paper extends the operational space formulation to the important class of serial-to-parallel chain (branching) manipulators. The various models and concepts developed in operational space, such as dynamically consistent force/torque decomposition for control of redundant manipulators and the augmented object model for cooperative manipulator systems, are shown to extend directly. Dynamic modeling and experimental results for a free-flying, two-arm space robot are presented to validate this extension.

This paper describes a collaborative research effort initiated by Monterey Bay Aquarium Research Institute (MBARI), Stanford Aerospace Robotics Laboratory (ARL), and NASA Ames Research Center. The goal of this joint effort was to develop an experimental system which demonstrates real-time supervisory control of an underwater vehicle from an interactive, 3-D graphical interface.

This paper presents a control strategy for manipulation of flexible objects by multiple robot arms. The control policy developed in this paper for flexible objects is based on a controller developed previously for rigid objects: the object impedance controller. The controller compensates for the dynamics of both the arms and the object and responds to environmental forces with a fully programmable impedance relationship. For a class of flexible objects, the required extensions to object impedance control are outlined in this paper. The paper presents experimental results for a dual arm robotic system manipulating an object with a single flexible degree of freedom in both free-space and environmental contact tasks.>