Optimal Joint Torque Research Articles

Torque-balancing algorithm for the redundantly actuated parallel mechanism

In this work, a joint torque distribution algorithm of parallel mechanism (PM) with one redundant actuator is suggested. The algorithm is effective in avoiding the joint torque saturations. In the algorithm, the force equations are modified into the line equations representing the joint torque solutions. Then, the optimal joint torque distribution is found only by searching the intersections among the line equations. A comparative study of the suggested algorithm with the other typical actuator redundancy resolution algorithms, such as the minimum-norm (MN) torque algorithm, the weighted MN torque algorithm, and a task-priority algorithm, is conducted through simulations and experiments with a 3-DOF planar PM with one actuator redundancy. The simulations and experiments confirm that the suggested redundant actuation algorithm shows an enhanced performance in terms of joint torque limit avoidance and minimum output force/torque distortion over other preexisting algorithms.

Optimal synthesisversusdynamic analysis for an acrobatic aerial movement

A predictive method: the optimal synthesis, based on an optimal dynamic problem, was solved for generating the aerial phase of a backward somersault. Optimal joint torques were compared with those resulting from inverse dynamics. A gymnast was modeled as a planar seven-segment multibody system with six internal degrees of freedom. The aerial motion was generated using a parametric optimization technique. The performance criterion minimized the time integral of quadratic joint torques. Optimal synthesis has produced a somersault closely mimicking the experimental one. Optimal joint torques revealed the motor control at the knee after takeoff unattainable by inverse dynamics analysis because of the measurement uncertainty propagation. Moreover, our study suggests that minimization of joint torques might be the primary control objective of a backward somersault.

A Modular and High-Precision Motion Control System With an Integrated Motor

This paper describes the design and control of an integrated direct-drive joint suitable for applications that require a high-precision motion control system. The joint integrates a direct-drive synchronous motor with axial air gap, a torque sensor, and a high-resolution position sensor. The key design aspects of the integrated joint, like the motor's armature, cooling system, motor housing, bearing arrangement, sensors, and power electronics are detailed. We also present a number of advanced implementations in motor torque control, optimal joint torque sensory feedback, and motion control using positive joint torque feedback. Experimental results illustrate an outstanding performance regarding thermal response, torque ripple, reference trajectory tracking, torque disturbance rejection, and joint stiffness.

Algorithms for Computing Optimal Joint Torques for Power Grasps.

Power grasp or enveloping grasp enables multifingered robot hands to grasp an object firmly by constraining it with multiple contacts with the surfaces of the finger links and palm. Most of power grasps can maintain equilibrium against external force/moment in all directions with constant joint torques as long as their contacts are unbroken. This paper proves that this condition is equivalent to the condition that the grasp force in power grasps is statically indeterminate. We propose two algorithms for computing optimal joint torques for power grasps taking the indeterminacy into consideration. The first algorithm computes the most robust joint torques against changes in external force/moment. The second one computes the minimum magnitude of joint torques in order to maintain equilibrium against a given set of external force/moment. These algorithms are independent of computation of the indeterminate grasp force if it is linear for joint torques and external force/moment.

Motions With Minimal Joint Torques for Redundant Manipulators

The article presents a method for obtaining motions with minimal joint torques for redundant mechanical manipulators. The existing methods based on the local resolution of redundancy do not guarantee that singular configurations, which lead to large joint torques and joint accelerations, will also be avoided while optimizing joint torques. For the present method, the degrees of redundancy are resolved at the global level. It is observed that motions with optimal joint torques are usually associated with large joint speeds and joint accelerations, compared with the joint speeds and joint accelerations for motions with minimal joint speeds or joint accelerations. To overcome this problem, the optimization criterion for minimal joint torques is modified by combining it separately with the optimization criteria for both minimal joint speeds and minimal joint accelerations. The resulting optimal motions, with smaller joint speeds and joint accelerations, are achieved at the expense of having larger joint torques.