Using an electronic gyroscope with real-time microcontroller; design and implementation of an impedance type kinesthetic interface with high pose, force and timing fidelity

Date

2019

Author

Sızlayan, Seyit Yiğit

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

102
views

0
downloads

In this thesis, we propose the design of a 1-DOF kinaesthetic (kinesthetic) force feedback interface for haptic and other robotic applications. The developed system is based on direct-drive principle. Thus, it provides a mechanically robust (and easy-to-build), and affordable solution. The major advantage of the direct-drive approach is that the coupling between the actuator and human motion is transparent, a desired feature for human-machine interfaces. The main drawback of a direct-drive actuator interface is that the effective motion measurement resolution is very low compared to other systems that utilize capstan-drive-like mechanisms. To overcome this low motion measurement resolution problem, we used an electronic gyroscope to measure the angular velocity from the hand grip transparently. We integrate the asynchronous encoder and synchronous gyroscope measurements to get high fidelity motion information using a statistical fusion filter implemented on an embedded microcontroller board. We developed three different filters, all based on Kalman estimator. Two of the filters fuse the synchronous gyroscope measurements with the asynchronous encoder measurements. One of the two filters uses a classical second-order constant velocity model where the gyroscope and the encoder measurements are treated as observations/measurements. The latter model is a first-order filter where the gyroscope measurements drive the system from the input terminal. The last filter is also a second order filter where we only use the encoder for both position and velocity estimation. We tested and compared the designed filters on encoders with resolution down to 400CPR. Results show that the filters that utilize the gyroscope measurements substantially outperform the encoder-only solution in terms of estimation precision. We did not observe significant differences between the filters that use the gyroscope. The integrated system provides us 1KHz hard-real-time operation and programmable interfaces for a considerably low cost. Another critical sub-system of an haptic interface is the torque control module whose bandwidth needs to be high enough for accurate impedance control. In the literature, PWM based closed-loop torque control is commonly used for this purpose. However, in this thesis, we designed a high-bandwidth, OPAMP-based current controller for rendering force commands. We compared a PWM based solution with our driving circuit and showed that our force control policy provides a much better torque tracking performance. The most important performance metric of a haptic interface is the Z-width of the system, which directly depends on the measurement accuracy, sampling rate, and the force-fidelity. Due to extremely low effective resolution caused by direct-drive mechanisms, the vast majority of the haptic interfaces uses the capstan-like transmission to satisfy the high Z-width requirements. Indeed, our system achieves high measurement and force rendering accuracy, and a quite fast 1KHz hard-real-time operation, while still enjoying the benefits of a direct drive approach. We believe that our approach in this thesis can be used to design and develop direct-drive based (high DOF) haptic systems with great Z-width performance.

Subject Keywords

Gyroscopes., Keywords: Gyroscope, Encoder, Fusion, Kalman Filter, Haptic Interface, Pose Fidelity, Embedded Control, Human-Machine Interface.

URI

http://etd.lib.metu.edu.tr/upload/12624265/index.pdf
https://hdl.handle.net/11511/44186

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Design of Programmable, High-Fidelity Haptic Paddle Sizlayan, Seyit Yigit; Ankaralı, Mustafa Mert (2019-01-01) In this study, we propose the design of a 1-DOF haptic interface for kinesthetic learning applications and educational platforms in some engineering and science courses. The proposed system is a directly-driven, high fidelity, easy-to-build, easy-to-program haptic device and yet still affordable for use in education. The major advantage of the system is that there is no transmission between the motor and human (i.e. direct-drive) motion which greatly reduces the mechanical complexity, improves robustness, r...
Exploring and implementing pleasant touch in the interface of products for design purposes : the case of a Bang & Olufsen remote control Fennis, Tirza Johanna Maria; Şener Pedgley, Bahar; Sonneveld, Marieke; Department of Industrial Design (2012) This thesis proposes a design strategy for pleasant touch. Literature is reviewed on the importance of pleasant touch, existing implementations in products and design for tactility. A lack of competence is found on how to design for pleasant touch in the interface of products: functional pleasant tactility. Therefore, a design vision is created by the author as a designer, on how to design functional pleasant tactility. The envisioned design strategy is then implemented through a study where a Bang & Olufse...
Image generation by back-propagation on input using a discriminator network Taplı, Merve; Akbaş, Emre; Department of Computer Engineering (2021-9-08) In this thesis, we propose an image generation method that only involves a discriminator network; no generator or decoder networks are required. To generate an image, we iteratively apply an adversarial attack on the discriminator by updating the input image, which is noise at the beginning, to maximize the discriminator's output score. Generated images are then used as negative examples, together with the real images as positive examples, to fine-tune the discriminator. After several rounds of generation a...
Supervised and unsupervised models of brain networks for brain decoding Alchihabi, Abdullah; Yarman Vural, Fatoş Tunay; Önal Ertuğrul, Itır; Department of Computer Engineering (2018) In this thesis, we propose computational network models for human brain. The models are estimated from fMRI measurements, recorded while subjects perform a set of cognitive tasks. We employ supervised and unsupervised machine learning techniques to represent high level cognitive tasks of human brain by dynamic networks. In the first part of this thesis, we propose an unsupervised multi-resolution brain network model. First, we decompose the signal into multiple sub-bands using Wavelet transform and estimate...
Extending design cognition with computer vision and generative deep learning Çakmak, Başak; Mennan, Zeynep; Department of Architecture (2022-2) This thesis proposes to scrutinize the project of extending cognition in the field of design by providing the exploration of alternative procedures of formalization. It aims at investigating the possible extensions of design cognition with a theoretical framework based on the concept of Extended Cognition, traced within two different contexts: One is the phenomenological context, which aims at a sensual and cognitive extension in the design studies of the 20th century searching for a counter-cultural expres...

Citation Formats

S. Y. Sızlayan, “Using an electronic gyroscope with real-time microcontroller; design and implementation of an impedance type kinesthetic interface with high pose, force and timing fidelity,” Thesis (M.S.) -- Graduate School of Natural and Applied Sciences. Electrical and Electronics Engineering., Middle East Technical University, 2019.