Cutting forces prediction is very important in micromilling for cutting tool's design and process planning. This paper presents a new model for uncertainty estimation of dynamic cutting forces in micromilling using a type-2 fuzzy rule-based system. The type-2 fuzzy estimation not only filters the noise and estimates the instantaneous cutting force in micromilling using observations acquired by sensors during cutting experiments, but also assesses the uncertainties associated with the prediction caused by the manufacturing errors and signal processing. Moreover, the interval output of the type-2 fuzzy system gives very useful information to machine tool controllers in order to maximize material removal while controlling tool wear or tool failure to maintain part quality specifications.

Because of the difficulty in understanding the physics of the machining process, several different intelligence methods, which employ cutting forces for estimation tool wear, have been developed in the past few years. Unfortunately, none of them can overcome the difficulty to estimate the errors of approximation during tool wear monitoring. This paper aimed at presenting a tool wear monitoring method using type-2 Takagi-Sugeno-Kang (TSK) fuzzy approach. This innovative method not only provides high reliability of the tool wear prediction over a wide range of cutting conditions, but also assesses the uncertainties associated with the modeling process and with the outcome of the model itself. The magnitude and direction of uncertainties in the machining process are described explicitly to increase the credibility of assessments.

In this paper, the problem of multi-objective trajectory planning is studied for redundant planar serial manipulators using a data-driven hybrid neuro-fuzzy system. A first pre-processing step involves an offline planning generating a large dataset of multi-objective trajectories, covering mostly the robot workspace. The optimized criteria are travelling time, consumed energy, and singularity avoidance. The offline planning is initialized through a cycloidal minimum time parameterized trajectory in joint space. This trajectory is then optimized using an augmented Lagrangian technique. The outcomes of this pre-processing step allow building a Tsukamoto neuro-fuzzy inference system to learn and capture the robot multi-objective dynamic behavior. Once this system is trained and optimized, it is used in a generalization phase to achieve online planning. Simulation results showing the effectiveness of the proposal are presented and discussed.

We study, for each of the possible joint-sensor layouts, the subspaces into which the motion of the hip-attachment points of parallel manipulators are completely measured. The projection of the motion of these points onto their subspaces allows us to write the underlying direct kinematics as a linear algebraic system constrained by the proper orthogonality of the rotation matrix. Although the solution of this problem requires a nonlinear technique, we propose a linear procedure that provides what we term a polar least square estimate. The resulting procedure is fast, robust to measurement noise, and produces estimates with about the same accuracy as a nonlinear procedure.

In this paper we decouple the translational and rotational degrees of freedom of the end-effector of parallel manipulators, and hence, decompose the direct kinematics problem into two simpler subproblems. Most of the redundant joint-sensor layouts produce a linear decoupling equation expressing the least-square solution of position for a given orientation of the end-effector. The resulting orientation problem can be cast as a linear algebraic system constrained by the proper orthogonality of the rotation matrix. Although this problem is nonlinear, we propose a procedure that provides what we term a decoupled polar least-square estimate. The resulting procedure is fast, robust to measurement noise, and produces estimates with about the same accuracy as a procedure for nonlinear systems if sufficient redundancy is used.

This paper presents the results of the implementation of a genetic algorithm for the geometrical synthesis of star-like topology parallel manipulators. The algorithm allows one to use several design criteria while considering a population of potential designs. The algorithm has proven to be simple and effective. Although the Y-star parallel manipulator is the design that possesses the greatest percentage of its workspace having a good dexterity, it is not the one that produce the largest workspace volume or the largest cubic volume inscribed within its workspace.

Using redundant sensors, the direct kinematics of general parallel manipulators is here decoupled and reduced to an orientation problem only. The accuracy of the decoupling depends on the condition number of a decoupling matrix. The conditions under which this matrix is isotropic are derived and illustrated with a numerical example.

The effect of redundant sensors in the solution of the direct kinematics of general parallel manipulators is studied here, for purposes of online implementation. We describe three measurement conditions under which the kinematic relations are linear and hence, a linear decoupling of the position and orientation of the direct kinematics is possible. For each condition, a decoupling equation is derived and used to reduce the direct kinematics to an orientation problem only. The latter is solved by resorting to the polar-decomposition theorem. The numerical accuracy and efficiency of the procedure are also investigated.>