Boosting human force: A robotic enhancement of a human operator’s force

Abstract

For a human operator, lifting and carrying a load can be a hazardous task with significant health-related risks. The collaboration between the human operator and a robotic device can provide an effective solution. This study proposes a control design for the robot such that it merely amplifies the force the human operator applies to the load. The algorithm is of a cyclic nature, where each cycle is subdivided into two stages. The first phase is the estimation of the human force, since the only connection between the robot and the human is via the load. In the second phase, the scaled force is applied to the load. The proposed approach is illustrated by means of simulation results for a two-link robot manipulator.

Robust disturbance estimation for human-robotic comanipulation

“…Applications with time‐varying mass can be encountered in the case of dispersing liquids from a container (e.g., painting or concrete pouring on building sites). This method extends a result presented in , which considered only the case with time‐invariant mass. The proposed algorithm is robust for large uncertainties in the mass of the load.…”

“…Theorem 2 shows that for a nonconstant external force, the estimation error remains bounded and that this bound can be related to the time derivatives of the external force via a linear ISS gain relation (see (21)). Moreover, the functionˇj .r, t/ as in (20) reflects a bound on the transient convergence rate of the estimation algorithm.…”

“…With the maximal human operation frequency, we can determine the period T of one cycle of the algorithm, which includes the estimation and amplification stages, see Figure 3. Using the analysis in Theorem 2, one can obtain an upper bound for the settling time for the estimator (from the expression of the functionˇj in (20)). If we consider that the output has settled if the error has dropped below 5% of the initial value, then the settling time for functionˇj as in (20) j max .P j / min .P j / .…”

“…The ISS property also provides some important insights for the design of the global controller because it allows to determine the period of the estimation cycle T 0 . The functionˇj , j D 1, 2, as in (20), is providing a bound on the convergence rate for the system:ˇ.r, t/ D re 1000t , and the settling time is T s D max j D1,2°3 4 j ± D 0.003 s. As a consequence, we have chosen T 0 D 0.005 s > T s .…”

Control scheme for human-robot co-manipulation of uncertain, time-varying loads

“…the mass of the load is considered known, or with very small uncertainty, and constant. The current study, which extends a result presented in [7], introduces a control strategy for a robotic device that amplifies the human force. As mentioned before, the robot does not have any This work is done in the frame of the TU/e Teleoperations project.…”