An enhanced edge tracking method using a low resolution tactile sensor

Abstract

Edge detection plays an important role in object recognition and exploration. In this paper, we propose an efficient tracking algorithm, which uses a coarse resolution tactile sensor set, for edge detection of a 2D shape object. Although many researchers have often used a tactile sensor with high resolution such as a 10x10 or higher for edge detection, in this research, we use a 2x2 tactile sensor to track the edges of an object. Using this type of low resolution sensor, we can reduce the manufacturing cost of the sensor and simplify the calculation and control process. In our algorithm, we only need to control the sensor to make it move along the vertical axis, Oy, or the horizontal axis, Ox. The sensor is moved along a suitable direction according to the sensing signals receiving from it. A set of simulations for both convex and concave shapes has been done to verify the algorithm. In addition, methods to increase the accuracy of the algorithm are also discussed.

Exploratory Tactile Servoing With Active Touch

“…a ridge or edge) around an object. Robotics work on tactile edge following dates back a quarter of a century [16], [17], and has focussed on edge following with image processing techniques such as median filters, Hu transforms and, more recently, geometric moments [18] and image dilation [19]. Image processing techniques have also been integrated into a control framework for tactile servoing to follow extended features such as planar wires [4].…”

Active contour following to explore object shape with robot touch

“…Although some techniques exist that allow robots to follow contours, none have used the probabilistic active perception approach developed here. For instance, in [11,12,13], tactile images were used for edge recognition, applying image processing techniques such as median filter, Hu transform and geometric moments. Meanwhile, another approach for contour following was based on contact force and orientation with respect to the surface [14,15].…”

A Local Filtering Technique for Robot Skin Data

“…The main advantage of this method is that algorithms developed for images can be directly applied to the tactile sensing domain when the shape of the contact distribution needs to be processed [2]. In the literature, several works proposed the use of tactile images generated from planar tactile sensing devices for contact shape classification [2], [10]- [13] or processing and features extraction [14]- [18]. The major limitation of tactile images is that they can only be directly generated from sensors distributed on a planar surface.…”

“…In particular, within the scope of this paper, the specific problem of processing the contact distribution to extract edges is addressed. The detection of this geometric feature is important for dexterous manipulation, as demonstrated by previous works addressing this problem with the use of tactile images [14], [15], [17], [18], by using optical tactile sensing technologies which directly provide an image of the contact distribution [16] or by using raw sensor measurements [20], [21]. However, the latter are based on Bayesian active exploration and the processing of contact shape (which is the focus of this paper) is not considered.…”