Comparison of position and force-based techniques for environment stiffness estimation in robotic tasks

Environment stiffness estimation techniques traditionally rely on position data. Unfortunately, obtaining the required information with sufficient accuracy and precision can be difficult, motivating the search for alternative approaches. COBA is an online stiffness estimation algorithm based on force data only, that does not require position information. In this paper, we compare the results of position-based stiffness estimation algorithms with those obtained by COBA. Both simulation and experimental results were collected and analysed. We find that, by not requiring position data, COBA avoids some problems that can negatively affect the performance of position-based algorithms. This demonstrates the usefulness of force-based stiffness estimation techniques for adaptive force control techniques.