Robotic systems for manipulation tasks are increasingly expected to be easy to configure for new tasks or unpredictable environments, while keeping a transparent policy that is readable and verifiable by humans. We propose the method BEhavior TRee eXPansion with Large Language Models (BETR-XP-LLM) to dynamically and automatically expand and configure Behavior Trees as policies for robot control. The method utilizes an LLM to resolve errors outside the task planner's capabilities, both during planning and execution. We show that the method is able to solve a variety of tasks and failures and permanently update the policy to handle similar problems in the future.

The control system of industrial robots is often model-based, and the quality of the model of high importance. Therefore, a fast and easy-to-use process for finding the model parameters from a combination of prior knowledge and measurement data is required. It has been shown that the experiment design can be improved in terms of short experiment times and an accurate parameter estimate if the robot configurations for the identification experiments are selected carefully. Estimates of the information matrix can be generated based on simulations for a number of candidate configurations, and an optimization problem can be solved for finding the optimal configurations. This work shows that the proposed method for improved experiment design works with a real manipulator, i.e. it is demonstrated that the experiment time is reduced significantly and the accuracy of the parameter estimate can be maintained or reduced if experiments are conducted only in the optimal manipulator configurations. It is also shown that the model improvement is relevant for realizing accurate control. Finally, the experimental data reveals that, in order to further improve the model accuracy, a more advanced model structure is needed for taking into account the commonly present nonlinear transmission stiffness of the robotic joints.

Industrial robots can solve tasks in controlled environments, but modern applications require robots able to operate also in unpredictable surroundings. An increasingly popular reactive policy architecture in robotics is Behavior Trees (BTs) but as other architectures, programming time drives cost and limits flexibility. The two main branches of algorithms to generate policies automatically, automated planning and machine learning, both have their own drawbacks and have not previously been combined for generation of BTs. We propose a method for creating BTs by combining these branches, inserting the result of an automated planner into the population of a Genetic Programming algorithm. Experiments confirm that the proposed method performs well on a variety of robotic assembly problems and outperforms the base methods used separately. We also show that this high level learning of Behavior Trees can be transferred to a real system without further training.

Gearbox failures cost thousands of lost production hours in plants that use industrial robots. In this context, an automated monitoring system that can warn the user of an impending failure can save precious resources. This problem has been addressed in many other domains through the use of machine learning approaches. However, standard machine learning algorithms are limited in their ability to detect gearbox failures, mainly due to task variability arises from robot-specific data. To improve detection performance of machine learning approaches, in this paper we propose techniques to curate the data prior to building a classification model. In a systematic hypothesis-driven study exploring the effect of different preprocessing techniques, we evaluate training data augmentation with estimated measurements, data differencing to suppress task dependence, inclusion of local variation, and selection of principal components on data collected from 26 industrial robots from the field. Our results show that preprocessing techniques improve the failure detection performance.

An estimation-based iterative learning control (ILC) algorithm is applied to a realistic industrial manipulator model. By measuring the acceleration of the end-effector, the arm angular position accuracy is improved when the measurements are fused with motor angle observations. The estimation problem is formulated in a Bayesian estimation framework where three solutions are proposed: one using the extended Kalman filter (EKF), one using the unscented Kalman filter (UKF), and one using the particle filter (PF). The estimates are used in an ILC method to improve the accuracy for following a given reference trajectory. Since the ILC algorithm is repetitive no computational restrictions on the methods apply explicitly. In an extensive Monte Carlo simulation study it is shown that the PF method outperforms the other methods and that the ILC control law is substantially improved using the PF estimate.

A sensor fusion method for state estimation of a flexible industrial robot is presented. By measuring the acceleration at the end-effector, the accuracy of the arm angular position is improved significantly when these measurements are fused with motor angle observation. The problem is formulated in a Bayesian estimation framework and two solutions are proposed; one using the extended Kalman filter (EKF) and one using the particle filter (PF). The technique is verified on experiments on the ABB IRB4600 robot, where the accelerometer method is showing a significant better dynamic performance, even when model errors are present.

An adaptive iterative learning control (ILC) algorithm based on an estimation procedure using a Kalman filter and an optimization of a quadratic criterion is presented. It is shown that by taking the measurement disturbance into consideration the resulting ILC filters become iteration-varying. Results from experiments on an industrial robot show that the algorithm is successful also in an application.

This paper gives an overview of classical iterative learning control algorithms. The presented algorithms are also evaluated on a commercial industrial robot from ABB. The paper covers implicit to explicit model-based algorithms. The result from the evaluation of the algorithms is that performance can be achieved by having more system knowledge.