Haptic guidance is a promising method for assisting an operator in solving robotic remote operation tasks. It can be implemented through different methods, such as virtual fixtures, where a predefined trajectory is used to generate guidance forces, or interactive guidance, where sensor measurements are used to assist the operator in real-time. During the last years, the use of learning from demonstration (LfD) has been proposed to perform interactive guidance based on simple tasks that are usually composed of a single stage. However, it would be desirable to improve this approach to solve complex tasks composed of several stages or gestures. This paper extends the LfD approach for object telemanipulation where the task to be solved is divided into a set of gestures that need to be detected. Thus, each gesture is previously trained and encoded within a Gaussian mixture model using LfD, and stored in a gesture library. During telemanipulation, depending on the sensory information, the gesture that is being carried out is recognized using the same LfD trained model for haptic guidance. The method was experimentally verified in a teleoperated peg-in-hole insertion task. A KUKA LWR4+ lightweight robot was remotely controlled with a Sigma.7 haptic device with LfD-based shared control. Finally, a comparison was carried out to evaluate the performance of Gaussian mixture models with a well-established gesture recognition method, continuous hidden Markov models, for the same task. Results show that the Gaussian mixture models (GMM)-based method slightly improves the success rate, with lower training and recognition processing times.

Future human spaceflight exploration missions to the Moon and beyond are hypothesised to benefit from human-robotic integrated operations. The European Space Agency focuses on preparing these operations, following the objective stated in the Global Exploration Roadmap of the International Space Exploration Coordination Group. Currently, human-robotic operations aim at technology development and demonstration, yet essential questions that remain unanswered are: How can human performance be measured, and which metrics are used? This contribution aims at answering this by preparing a pilot phase focusing on human performance assessment for subjects controlling a rover. A recent study by Hosseini (2016) for ESA identified essential knowledge gaps that must be filled to assess astronaut performance for tele-operations, which is of great importance for future missions since they affect mission planning, task allocation and even tool selection. In this regard, this study proposes follow-on research in which a tele-operations experiment is conducted by driving a rover, in order to evaluate human performance in tele-operations. In the previous study, two space-to-ground and multiple ground-to-ground tele-operations experiments were analysed in which subjects controlled a rover. Data analysis studied the command time and execution time of the assigned tasks, i.e. the time it takes for the human to give the command to the rover and the time it takes for the rover to execute its tasks, respectively. Results showed that the main challenge for performance assessment is the lack of recorded parameters. The logged data is limited only to time values and success/failure results and it does not specify performance variations. Furthermore, the study concludes that many different sub-tasks were performed in a limited amount of time, resulting in scarce data per sub-task and limiting the statistical significance. Follow-on research is proposed that aims at solving for the two above mentioned issues. Firstly, the study introduces parameters, which are used to assess the performance of pilots regarding neuro-ergonomics and human factors. Secondly, an experiment is set up and is tested for its rigidity in a protocol rehearsal, prior to performing the experiment with a relatively large group of participants. It is hypothesised that this approach has the potential not only to increase the qualitative assessment of the performance, but also to increase the quantitative results essential for preparing crew training and future missions.

This paper introduces a new Learning from Demonstration (LfD)-based method that makes usage of robot effector forces and torques recorded during expert demonstrations, to generate force-based haptic guidance reference trajectories on-line, that are intended to be used during haptic shared control for additional operator 'guidance'. Derived haptic guidance trajectories are superimposed to master-device inputs and feedback forces within a bilateral control experiment, to assist an operator by the guidance during peg-in-hole insertion. We show that 96 peg-in-hole expert demonstrations were sufficient to obtain a good model of the task, which was used on-line to generate haptic guidance trajectories in real-time with a 1kHz sampling rate.

On July 28th 2014, 23:47 UTC, the European Space Agency launched the Haptics-1 Kit to the International Space Station (ISS) on its last Automated Transfer Vehicle ATV-5. The Kit reached the station two weeks later, marking the first haptic master device to enter the ISS. The first force-feedback and human perceptual motor performance tests started to take place on December 30th 2014, and are the first of their kind in the history of spaceflight. Three astronauts participated in the Haptics-1 experiment until November 2015, allowing the investigation of the effects of microgravity on various psycho-motor performance metrics related with the usage of haptic feedback. Experiments are conducted following full adaptation to the space environment (after 3 months in space). This paper introduces the Haptics-1 experiment and associated hardware. Detailed experimental results are reported from a first stiffness just noticeable difference (JND) experimental study in space, carried out on the ISS and pre-flight on ground with 3 astronauts. The first findings from the experiment show no major alterations in-flight, when compared to on-ground data, if the manipulandum is secured in flight against a sufficiently stiff reference structure.

In this paper, we propose a closed-loop force sensor based nested admittance/impedance control strategy to actively estimate and minimize the effects of geometric misalignment that naturally occur during assembly tasks with compliant robots. The method allows the robot to be used with a stiff impedance control setting, which is beneficial for free air motion performance, yet allows to adjust for large misalignment errors between parts that need be assembled.

This paper introduces a new Learning from Demonstration (LfD)-based method that makes usage of robot effector forces and torques recorded during expert demonstrations, to generate force-based haptic guidance reference trajectories on-line, that are intended to be used during haptic shared control for additional operator 'guidance'. Derived haptic guidance trajectories are superimposed to master-device inputs and feedback forces within a bilateral control experiment, to assist an operator by the guidance during peg-in-hole insertion. We show that 96 peg-in-hole expert demonstrations were sufficient to obtain a good model of the task, which was used on-line to generate haptic guidance trajectories in real-time with a 1kHz sampling rate.

In a typical space teleoperation task, mismatches between the viewing direction of the operator and the direction of their required control input are often unavoidable. To execute these tasks, the operator is then required to perform mental rotations. Recent studies have shown that the task performance can thereby significantly decrease. In this paper, for the first time, the influence of mental rotations on task performance is studied if hap tic feedback is provided to the operator. A human factors experiment is conducted which analyses the influence of two different hap tic feedback control methods via various visual missmatch angles. The rotation is thereby set to the extreme cases of 0. and 180.To clearly analyze the effects. The first hap tic feedback method consists of direct, scaled force and torque feedback to the operator as measured by a force/torque sensor at the slave robot. The second method consists of hap tic shared control which provides artificially generated guidance forces to the operator. It is shown that mental rotations decrease teleoperation performance despite the addition of direct force feedback. In contrast, hap tic shared control provides lower increase in the operator mental workload and also less between-operator variability of errors made due to the mental rotations.