Abstract | ||
---|---|---|
For the majority of robots, the environment around them is built from rigid bodies. However, we know that there are many elastic and deformable objects around us, so the robots should have a way of perceiving them. The best way to achieve this is through direct interaction with the environment - haptic sensing. In our work, we are addressing a problem of haptic material classification. In order to do that we used novel Deep Neural Network architecture which is able to deal with long signal sequences. For the challenging dataset from the legged robot, we achieved a 97.96 % classification accuracy. The paper begins with a short introduction followed by the related work section. Afterwards, the details of our method are given. Next, the datasets used in the research are presented. Then, the obtained results are discussed. Finally, concluding remarks and future work plans are given. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/ROBOSOFT.2019.8722819 | 2019 2nd IEEE International Conference on Soft Robotics (RoboSoft) |
Keywords | DocType | ISBN |
Legged locomotion,Robot sensing systems,Haptic interfaces,Task analysis,Windows | Conference | 978-1-5386-9260-8 |
Citations | PageRank | References |
2 | 0.39 | 0 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jakub Bednarek | 1 | 4 | 2.12 |
Michał Bednarek | 2 | 3 | 3.45 |
Piotr Kicki | 3 | 8 | 2.20 |
Krzysztof Walas | 4 | 22 | 4.23 |