We consider the kinematics of a mobile manipulator that consists of a nonholonomic (or holonomic) mobile platform and a holonomic manipulator mounted atop of the platform. At any fixed time instant and with fixed initial posture of the platform these kinematics are viewed as a map from a Hilbert space of controls and manipulator joint positions into a task-space of the mobile manipulator, and called the instantaneous kinematics. A derivative of this map is referred to as the instantaneous analytic Jacobian of the mobile manipulator. We derive an instantaneous dexterity matrix and a dexterity measure of the mobile manipulator. The dexterity matrix is used in order to introduce an instantaneous dexterity ellipsoid of the mobile manipulator. Concepts are illustrated with examples of mobile manipulators equipped with both holonomic and nonholonomic mobile platforms.

We present a normal form approach to solving the singular inverse kinematic problem for non-redundant robot kinematics. A standing assumption is made that singular configurations of the kinematics have corank 1 and that the kinematics are equivalent to the quadratic normal form. The approach has been illustrated with a simulation case study of 2R kinematics. A comparison with the singularity-robust inverse technique and the null-space based approach has demonstrated advantages of the normal form approach.

In this article, a new method of solving the singular inverse kinematic problem for nonredundant robotic manipulators is proposed, based on normal forms of the manipulator kinematics. A complete solution to the inverse problem is presented under assumption that the kinematics are equivalent to the quadratic normal form. Comparisons with other existing methods, including the singularity-robust inverse, are made that reveal advantages of the normal form approach. Experimental results concerned with singular tracking in the AdeptOne manipulator are reported.

This paper gives the concept of an autonomous robotic agent that is capable of showing both machine learning and reactive behavior. The first methodology is used to collect information about the environment and to plan robot actions based on this information while the robot is performing tasks. Processing and storing information obtained during several task executions is called lifelong learning. Reactive behavior, the second desirable feature of an autonomous robot, is needed to execute the actions in a dynamically changing environment. This paper presents the main components of the machine learning part of the autonomous robotic agent: the conceptual state generalization module that forms conceptual states as categories (clusters) of observation vectors obtained from the sensor system; and the robot behavior effects modeller that calculates the next conceptual state from the current state and an action obtained from the action planner.