Hybrid force/position control for quasi continuum manipulators

The inherent compliant and safe structure of fluid driven continuum manipulators makes them a promising solution for various tasks. Despite their cheap production costs these robots have yet not found their way into industrial applications. This is due to the lack of precise models as well as control strategies which are both open fields of research.A basic industrial task is to control the force the manipulator exerts at its tool center point on a given object. In this work we present a hybrid force/position controller (HFPC) for the Bionic Soft Arm (BSA). It is assumed that contact is only established at the tool center point where the contact force can be measured. Further, we show how to extend the basic HFPC approach in order to overcome model inaccuracies. Experimental results are provided for the BSA where the HFPC is incorporated into an existing structure.

Variable Compliance Control for Robotic Peg-in-Hole Assembly: A Deep-Reinforcement-Learning Approach

Industrial robot manipulators are playing a significant role in modern manufacturing industries. Though peg-in-hole assembly is a common industrial task that has been extensively researched, safely solving complex, high-precision assembly in an unstructured environment remains an open problem. Reinforcement-learning (RL) methods have proven to be successful in autonomously solving manipulation tasks. However, RL is still not widely adopted in real robotic systems because working with real hardware entails additional challenges, especially when using position-controlled manipulators. The main contribution of this work is a learning-based method to solve peg-in-hole tasks with hole-position uncertainty. We propose the use of an off-policy, model-free reinforcement-learning method, and we bootstraped the training speed by using several transfer-learning techniques (sim2real) and domain randomization. Our proposed learning framework for position-controlled robots was extensively evaluated in contact-rich insertion tasks in a variety of environments.

Analysis of Human Computer Interaction (HCI) Model in SMEs

In Small and Medium-sized Enterprises (SMEs), Human-Computer Interaction (HCI) is considered as a cross-disciplinary segment applied in ergonomics, psychology and the engineering departments. HCI deals with the evaluation, implementation, designing and theoretical evaluation of means in which humans utilize and relate with computing applications. The term ‘Interaction’ is differentiated from other terminologies in the same application interface. The term refers to the abstract system which allows humans to interact with devices of computing for a particular industrial task. An application interface in this case applies to the selection of the technical (software and hardware) realization of a specified interaction system. Because of extensive research to incorporate diversified HCI into an understandable model, this paper evaluates HCI model in SMEs to provide the projected guidance to designers of the system using Information Technology (IT). The choice of a good model provides the recommendable direction for presentation languages e.g., Task Action Grammar (TAG) and the design actions determine the feel and look of the system. In this contribution, critical design projects in every discipline are identified alongside the present study trends and future research directions.

Electromyography Assessment of the Assistance Provided by an Upper-Limb Exoskeleton in Maintenance Tasks

In this paper, the analysis of the intensity of muscle activations in different subjects when they perform an industrial task in a repetitive way assisted by a robotic upper-limb exoskeleton is presented. To do that, surface electromyography (EMG) signals were monitored with and without a robotic upper-limb exoskeleton for 10 subjects during a drilling task, a typical tedious maintenance or industrial task. Our results show that wearing the upper-limb exoskeleton substantially reduces muscle activity during a drilling task above head height. Specifically, there is statistically significant differences in the pectoralis major and rhomboids muscles between the groups wearing or not wearing the robotic upper-limb exoskeleton.

Incorporating artificial skin signals in the constraint-based reactive control of human–robot collaborative manipulation tasks

Purpose The purpose of this paper is to develop a control strategy for human–robot collaborative manipulation tasks that can deal with proximity signals from 373 interconnected cells of an artificial skin. Design/methodology/approach The robot and the operator accomplish an industrial task while interacting in a shared workspace. The robot controller detects and avoids collisions based on the information from the artificial skin. Conflicting constraints can be handled by prioritizing between hard and soft constraints or by weighing the different constraints. Findings Weak soft constraints (low weight) are specified to command the robot to move along a nominal path with constant velocity. Stronger soft constraints (higher weight) prevent collisions by means of either moving the end effector backward along the path or circumventing an obstacle. The proposed approach is validated experimentally. Originality/value As a first contribution, this paper proposes a discrete optimization algorithm activates an a priori selected maximum number of cells. The algorithm selects the appropriate distribution based on the amplitude of each signal and the spatial distribution of the proximity measurements. A second contribution is the specification of a human–robot collaborative application as an optimization problem using eTaSL (expression graph-based task specification language), which provides reactive control.

Natural-Language-Instructed Industrial Task Execution

To effectively cooperate with a human, advanced manufacturing machines are expected to execute the industrial tasks according to human natural language (NL) instructions. However, NL instructions are not explicit enough to be understood and are not complete enough to be executed, leading to incorrected executions or even execution failure. To address these problems for better execution performance, we developed a Natural-Language-Instructed Task Execution (NL-Exe) method. In NL-Exe, semantic analysis is adopted to extract task-related knowledge, based on what human NL instructions are accurately understood. In addition, logic modeling is conducted to search the missing execution-related specifications, with which incomplete human instructions are repaired. By orally instructing a humanoid robot Baxter to perform industrial tasks “drill a hole” and “clean a spot”, we proved that NL-Exe could enable an advanced manufacturing machine to accurately understand human instructions, improving machine’s performance in industrial task execution.