A mobile manipulator is treated as a robotic system composed of a non-holonomic mobile platform and a holonomic manipulator mounted on the platform. The kinematics of the mobile manipulator can be represented as a driftless control system with outputs. By adopting the endogenous configuration space approach we propose two kinematic dexterity measures, called local and global dexterity. The local dexterity, modeled upon the manipulability of stationary manipulators, indicates how infinitesimal motions in the configuration space propagate to the taskspace of the mobile manipulator. The global dexterity corresponds to L2-norm of the local dexterity over a prescribed region of the configuration space. Advantages of the endogenous dexterity measures over traditional performance measures of mobile manipulators known from the literature are described. Both the dexterities are employed for determining optimal configurations and optimal geometries of an exemplary mobile manipulator.

In this paper distributed architectures for autonomous vehicles are addressed, with a special emphasis on its real-time control requirements. The interconnection of the distributed intelligent subsystems is a key factor in the overall performance of the system. To better understand the interconnection requirements, the main techniques and modules of a global navigation system are described. A special focus on fieldbuses properties and major characteristics is made in order to point out some potentialities, which make them attractive in autonomous vehicles real-time applications, either in terms of reliability as in terms of real-time restrictions.

With the fast development of the control theory and engineering, robotics and artificial intelligence have become the focus in the field of intelligent systems. But research in this field is based on a series of experiments and requires various robot platforms which are often unaffordable by the universities in developing countries. The intention of this paper is to present the idea of applying innovative control education and building a laboratory using low cost equipments. In this paper the architecture and control system of a new kind of low cost intelligent robot, “Ability Storm”, is introduced. It not only shows the technical aspect, but also illustrates the related applications in control education. The versatility and effectiveness of this affordable intelligent robot platform are demonstrated through a number of experiments, including both basic laboratory experiments and other innovative project-oriented design, such as robot fire fighting, which clearly show the practicality and robustness of this cost-effective robot platform.

This paper discusses robots that are operational within a human-inhabited environment. Specifically, we identify different roles that such robots can adopt, reflecting different human-robot relationships. We discuss three different roles of robots in a project where we develop a robot as a therapeutic tool for children with autism: the robot as a therapeutic playmate, the robot as a social mediator, and the robot as a model social agent. Implications of these roles that go beyond this particular project are discussed.

A binary walking robot moves as a result of bi-state actuator transitions. Because of the bi-state nature of binary joints, many research results about continuous walking robots cannot be applied to binary walking robots directly. In this paper, a new and simple model of rigid-link binary walking robot is proposed, around which related concepts are introduced, and formulas are derived. Based on this model, general characteristics and limitations of periodic gaits are discussed, and the stability qualities of several straight-line walking periodic gaits are studied in both pitch-greater-than-stroke and stroke-greater-than-pitch cases. Valuable results are obtained from the analysis, which should be followed in the design of rigid-link binary walking robots.

This paper presents the implementation of an explicit model reference adaptive control (MRAC) for position tracking of a dynamically unknown robot. An auto regressive exogenous (ARX) model is chosen to define the plant model and the control input is optimised in a H2 norm to reduce computational time and to simplify the algorithm. The theory of MRAC falls into a description of the various forms of controllers and parameter estimation techniques, therefore, applications may require very complicated solution methods depending on the selected laws. However, in this study, the proposed MRAC shows that applications may be as easy as classical control methods, such as PID, by guaranteeing the stability and achieving the convergency of the plant parameters. Despite the selected simple control model, simple optimisation method and drawbacks of the robot the experimental results show that MRAC provides an excellent position tracking compared with conventional control (PID). Many experimental implementations have been done on the robot and one of them is included in the paper.

Three joint space algorithms slow the Cartesian path motion when it appears that joint motion is approaching a joint, speed, or acceleration limit. All three algorithms use quadratic curve fitting to predict where the joint motion is heading, followed by a prediction as to how much time would elapse until a limit is reached.

If a joint motion limit is encountered in the time-to-stop the Cartesian motion, these algorithms reduce the Cartesian speed using pulsed speed settings so that the robot or machine tool will have the necessary time to come to a complete stop. The joint space velocity and acceleration control algorithms set the override Cartesian speed to either full or some reduced speed, several times a second. This allows the joints to reach, but not exceed, their maximum velocity and accelerations limit, while remaining within the physical joint limits.

This paper describes an adaptive S-curve used to recover a tool path upon a system crash in the Windows operating system (OS). For a mechanism such as a robot or machine tool, the joint values, being delivered as setpoints to the servo-controller, are dynamically recorded by the real-time operating system also residing on the computer. The real-time OS can control the abort and record pertinent motion data after Windows OS crash.

Upon system recovery, the recovery trajectory generator examines the setpoints intervals to determine the current slow joint. At every trajectory step, and for the current slow joint, the S-curve velocity profile applies the joint entry state (position, speed, acceleration, and jerk) to interpolate the motion between the setpoints in a reverse direction. The other joints are proportionally interpolated (slowed) so that they pass through each setpoint simultaneously with the slow joint, but in a reverse direction.

The trajectory algorithm is optimal since the slow joint always uses the maximum allowable jerk to change the profile speed and acceleration for each trajectory step.

In this paper, we try to find an efficient and accurate dynamic model for a robot with spatial compliant links using the screw theory. After making some reasonable assumptions for the system, we introduce the Ding-Holzer method for system simplification. Instead of looking at the beam at rest, we consider the spatial forces and deflections of a moving elastic arm. We focus our study on the dynamic modelling problems of a compliant arm with six dimensional forces and deformations at the tip.

In this paper the use of inexpensive standard hardware and software is proposed in place of high-cost commercial solutions to set up real-time control experiments with a human in the control loop, including graphical and haptic virtual reality (VR). For this purpose a generic simulation environment for the implementation of control experiments using VR with haptic feedback and their evaluation, has been developed. As an example, setups of a car simulator with a human in the control loop and an inverted pendulum as an experiment for student laboratories are presented.

Kinematic calibration is essential to improve the accuracy of the manipulator. This paper presents a complete description of the Gough platform modeling and a unified scheme to identify its kinematic parameters. The interest of this formulation is that it may be applied whatever information is available on the state of the robot (measurement or constraints) without using the kinematics to obtain the basic system of constraint equations. Moreover, the scheme may be applied for all parallel robots. We propose to experiment and compare three methods of calibration, using either (or both) external measurement and internal redundant sensors. Finally, we show how to reduce the initial error in pose determination by 99% for the Hexapode 300, CMW's machining center, and validate the choice of a self-calibration method in an industrial context.

This work introduces the concept of a brush-based tractor used for rescue in collapsed buildings or tunnels. The paper presents the bristle theory and traction experiments relating to a robot, which uses a sensor system for detecting the shapes of pipes or tunnel like voids within rubble. Traction experiments in the laboratory were used to investigate the characteristics of bristles and the performance of the brush units of different shapes. The experimental results are discussed in the paper and related to a single bristle theory with a view to giving guidance for the design of a future brush based rescue robot in debris.

In this paper, an analytical study of the kinematics and dynamics of Stewart platform-based machine tool structures is presented. The kinematic study includes the derivation of closed form expressions for the inverse Jacobian matrix of the mechanism and its time derivative. An evaluation of a numerical iterative scheme for an on-line solution of the forward kinematic problem is also presented. Effects of different configurations of the unpowered joints on the angular velocities and accelerations of the links are considered. The Newton-Euler formulation is used to derive the rigid body dynamic equations. Inclusion of models for actuator dynamics and joint friction is discussed.

This paper describes the use of a hybrid system comprising a multi-layer perceptron and a Petri Net model to control the navigation of multiple mobile robots in an unknown and cluttered environment. The multi-layer perceptron is trained with examples representing the typical static obstacles to be encountered by the robots and the evasive actions to be taken. The Petri Net model embodies the rules describing how the robots should move in order to avoid colliding against one another. The hybrid system has been implemented in simulation software as well as in actual mobile robots. The paper presents the results of simulation and practical tests that demonstrate the ability of groups of robots incorporating the proposed navigation system to negotiate various types of obstacles and find targets successfully.

This paper presents a novel Cobotic system with differential CVT. The new system is significantly cheaper, simpler to control and more efficient than Cobots with S-CVTs. Both path-guidance and power-assist functions can be simply realized with the new system. Basic structures, kinematic and dynamic models, as well as control algorithms, which are essential for design, control synthesis and control of the system, are briefly presented in the paper.

Though many research efforts have been devoted to developing robot controllers with uncertainties in the robot dynamics, most research so far has assumed that the model of actuator is known exactly. In implementing the robot controllers, calibration is necessary to identify the exact parameters of the actuator model. When there is an error in calibrating or modeling, it is unknown whether the stability of the system could still be guaranteed. In addition, the actuator model could change as the temperature varies due to overheating of the motor or changes in the ambient temperature. To overcome this problem, we propose in this paper, a task-space controller for the regulation of the robot with an approximate actuator model. We consider second-order dynamics of electric actuators with negligible armature inductances. Sufficient conditions for choosing the feedback gains are presented to guarantee stability.

The paper is aimed at generating optimal gait cycles in the sagittal plane of a biped, the locomotion system of which has anthropomorphic characteristics. Both single and double support phases are globally optimised, considering incompletely specified transition postural configurations from one phase to the other. An impactless heel-touch is prescribed. Full dynamic models are developed for both gait phases. They are completed by specific constraints attached to the unilaterality of contact with the supporting ground.

A parametric optimisation method is implemented. The biped joint coordinates are approximated by cubic splines functions connected at uniformly distributed knots along the motion time. The finite set of unknowns consists of the joint coordinate values at knots, some gait pattern parameters at phase transitions, and the motion time of each phase. The step length is adjusted to the prescribed gait speed by the optimisation process. Numerical simulations concerning slow and fast optimal gaits are presented and discussed.

A novel mobile robot FTIR (Finned Tube Inspection Robot) for finned tube inspecting is presented in this paper. FTIR can move in both x- and y-direction on the top layer of the tubing that is installed in layers. The snake-like arms mounted on the robot can be lowered into the narrow space between the tubes and the tubing can be inspected through the device installed in the end of the arms. The moving mechanisms, inspecting device applied in narrow space and control system are introduced.

The design of a parallel manipulator for a given workspace would be greatly facilitated, if the analytic expressions of the hypersurfaces bounding the workspace were available. In translational parallel manipulators (TPMs), the hypersurfaces bounding the workspace are actually threedimensional surfaces, which are the geometric locus of all the positions assumed by an end-effector point when the TPM reaches the workspace borders. These surfaces can be represented in a Cartesian reference system fixed to the frame. This paper studies the workspace of two TPMs, that have the same closure equations and workspace when a few geometric conditions are satisfied. The two TPMs are the translational 3-RUU and the DELTA robot. The kinematic analyses of these two TPMs are different from one another only when velocity and accelerations are considered. The result of this study is that the analytic expression of the surfaces bounding their workspace is a fourth degree polynomial equation, in the coordinates of a plattorm point, which contains all the manipulator geometric parameters. This analytic expression is given in an explicit form. The use of this expression is illustrated through a numerical example.

KNOWLEDGE REPRESENTATION, REASONING AND DECLARATIVE PROBLEM SOLVING, by Chitta Baral, University Press, Cambridge, 2003, xiv+530 pp., ISBN 0-521-81802-8 (Hbk, £60.00).