Pose-independent recognition of convex objects from sparse tactile data

Beccari, G.; Caselli, Stefano; Zanichelli, Francesco

doi:10.1109/robot.1997.606861

Cited by 6 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Shape recognition has long been based on the notion of "interpretation tree," which represents correspondences between features extracted from the tactile data and those on the model [17], [21], [22]. A volumetric approximation [3] can be built over tactile data to enhance feature selection and prune incompatible models.…”

Section: B Tactile Shape Recognitionmentioning

confidence: 99%

“…In [3], internal and external volumetric approximations of a convex object were built over sparse tactile data. The main disadvantage of this method was its applicability-limited to convex objects only.…”

Section: B Tactile Shape Recognitionmentioning

confidence: 99%

“…In every instance, the real location of q is drawn as a circular dot and its estimated location q 3 as the closest cross. 2) Recognition Tests: In an instance of recognizing, say, the crossed trough, we generate three data curves through a randomly selected point on the surface, and try to register them onto the four surfaces 3 in Table I. The registration error (2) on the crossed trough itself is 0.000393.…”

Section: B Simulationmentioning

confidence: 99%

“…Past work on object recognition based on tactile data has been limited to certain classes of surfaces, including polyhedra [21], [22], convex objects [3], quadrics [31], and superquadrics [1].…”

mentioning

confidence: 99%

See 3 more Smart Citations

Recognition of Curved Surfaces From “One-Dimensional” Tactile Data

Ibrayev

Jia

2012

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

This paper presents a method for recognition of 3D objects with curved surfaces from linear tactile data. For every surface model in a given database, a lookup table is constructed to store principal curvatures precomputed at points of discretization on the surface. To recognize an object, a robot hand with touch sensing capability obtains data points on its surface along three concurrent curves. The two principal curvatures estimated at the curve intersection point are used to look up the table associated with each model to locate surface discretization points that have similar local geometries. Local searches are then performed starting at these points to register the tactile data onto the model. The model with the best registration result is recognized. The presented method can recognize closed-form surfaces as well as triangular meshes, as demonstrated through simulation and robot experiments. Potential applications include recognition of (manufactured) home items that are grasped daily, and dexterous manipulation during which recognition is simultaneously performed.Note to Practitioners-In automation, the eventual purpose of sensing is for assembly and other operations. Recognition by touch has several advantages: efficiency, robustness to disturbance, and immunity to occlusion (unlike a vision system). We demonstrate that 3D objects can be recognized with small amounts of data acquired via robot touch. In particular, recognition can be as simple as dragging one or more fingers across a part of an object's surface.

show abstract

Section: B Tactile Shape Recognitionmentioning

confidence: 99%

Section: B Tactile Shape Recognitionmentioning

confidence: 99%

Section: B Simulationmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Recognition of Curved Surfaces From “One-Dimensional” Tactile Data

Ibrayev

Jia

2012

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

show abstract

“…For instance, in [13] a method using a network of triangular B-spline patches based on arbitrary topological domain maintaining tangent plane continuity is used. Techniques for recognizing convex polyhedra from sparse tactile data, using volumetric representations, are used in [14]. The performance of an hybrid force/velocity controller for automatic edge following is analyzed in [15].…”

Section: Introductionmentioning

confidence: 99%

A 2D unknown contour recovery method immune to system non-linear effects

Juan¹,

Tur²

2006

2006 IEEE International Conference on Robotics and Biomimetics

View full text Add to dashboard Cite

Abstract-A method to recover general 2D a priori unknown contours using a kind of special optic sensor is described. Contour recovery is an important task for exploratory operations in unknown environments as well as for more practical applications such as grinding or deburring. It is not an easy task since the recovered contour (generally obtained using encoder data) is severely distorted due to errors in the kinematic model of the robot and to the non-linearities of its actuators. Some mathematical models have been presented to partially compensate for those effects, but they require a deep knowledge of both the robot and sensor models which are difficult to obtain accurately, and normally imply an adaptive non-linear control to estimate some of the unknown parameters of the model.Our approach, in despite of its simplicity, is intrinsically immune to non-linearities, which allows us to eliminate most of the distortions added to the sensor data. A simple algorithm to follow unknown planar contours is presented and used to test the performance of this approach in comparison to the one using encoder data. Experimental results and practical problems are also discussed.

show abstract