Computed torque control with nonparametric regression models

Abstract

Abstract-Computed torque control allows the design of considerably more precise, energy-efficient and compliant controls for robots. However, the major obstacle is the requirement of an accurate model for torque generation, which cannot be obtained in some cases using rigid-body formulations due to unmodeled nonlinearities, such as complex friction or actuator dynamics. In such cases, models approximated from robot data present an appealing alternative. In this paper, we compare two nonparametric regression methods for model approximation, i.e., locally weighted projection regression (LWPR) and Gaussian process regression (GPR). While locally weighted regression was employed for real-time model estimation in learning adaptive control, Gaussian process regression has not been used in control to-date due to high computational requirements. The comparison includes the assessment of model approximation for both regression methods using data originated from SARCOS robot arm, as well as an evaluation of the robot tracking performance in computed torque control employing the approximated models. Our results show that GPR can be applied for realtime control achieving higher accuracy. However, for the online learning LWPR is superior by reason of lower computational requirements.

Towards Motor Skill Learning for Robotics

“…While modern machine learning approaches such as Gaussian process regression (GPR) and support vector regression (SVR), yield significantly higher accuracy than traditional RBD models, their computational requirements can become prohibitively costly as they grow with number of data points. Thus, it is infeasible to simply use off-the-shelf regression techniques and the development of domain-appropriate versions of these methods is essential in order to make progress in this direction [7]. One possibility for reducing the computational cost is the partitioning of the data such that only the regionally interesting data is included in a local regression and, subsequently, combining these local predictions into a joint prediction.…”

Incremental online sparsification for model learning in real-time robot control

“…Due to the high complexity of modern robot systems such as humanoids or service robots, traditional analytical rigid-body model often cannot provide a sufficiently accurate inverse dynamics model. The lack of model precision has to be compensated by increasing the tracking gains K p and K v making the robot stiff and less safe for the environment [9]. Thus, to fulfill both requirements of compliant control, i.e., having low tracking gains and high tracking accuracy, more precise models are necessary.…”

“…For evaluation, we combine our sparsification framework with an incremental approach for Gaussian process regression (GPR) as described in [10]. The resulting algorithm is applied in online learning of the inverse dynamics model for robot control [17,9].…”

Estimating the non-linear dynamics of free-flying objects

“…The Hidden Markov model (HMM) is used for recognition and regeneration of human motion across various demonstrations [39,40], and to teach a robot to perform assembly tasks [41]. Locally Weighted Projection Regression (LWPR) is used to approximate the dynamic model of a robot arm for computed torque control [42,32], for teaching a robot to perform basic soccer skills [43], and is also applied for real-time motion learning for a humanoid robot [32]. Statistical approaches that find a locally optimal model to represent the demonstrations by maximizing the likelihood, are also possible.…”

“…They transform the demonstrations into a single or mixed distribution. To approximate the dynamic model of a robot arm for computed torque control [42,44], Gaussian Process Regression (GPR) is used. Additionally, Gaussian Mixture Regression (GMR) is used to model a robot motion from a set of human demonstrations, as a time-dependent model [45] or time-independent dynamical system [46].…”