Classification of Household Materials via Spectroscopy

  title={Classification of Household Materials via Spectroscopy},
  author={Zackory M. Erickson and Nathan Luskey and S. Chernova and Charles C. Kemp},
  journal={IEEE Robotics and Automation Letters},
Recognizing an object's material can inform a robot on the object's fragility or appropriate use. To estimate an object's material during manipulation, many prior works have explored the use of haptic sensing. In this letter, we explore a technique for robots to estimate the materials of objects using spectroscopy. We demonstrate that spectrometers provide several benefits for material recognition, including fast response times and accurate measurements with low noise. Furthermore… 

Pregrasp Object Material Classification by a Novel Gripper Design with Integrated Spectroscopy

It is concluded that spectroscopy is a promising sensing modality for enabling robots to not only classify grasped objects but also understand their underlying material composition.

Robotic Touch: Classification of Materials for Manipulation and Walking

This work is addressing a problem of haptic material classification by using novel Deep Neural Network architecture which is able to deal with long signal sequences and achieving a 97.96 % classification accuracy.

Object Description Using Visual and Tactile Data

Convolutional neural networks are used to separately extract visual and haptic features and then fuse these two types of features to form a multitask-multilabel classification method that produces the most accurate object descriptions with the smallest number of parameters.

Study of Reflection-Lass-Based Material Identification from Common Building Surfaces

Compared to existing material identification methods, the proposed reflection-loss-based method is capable of identifying materials from a significant distance without requiring any contact with the object and without requiring dedicated sensors from the infrastructure point of view.

The Visual Segmentation of Scene Information and Applications in Predictive Haptics

This work develops a way to segment a scene for the various objects in the scene using a pre-trained neural network that takes in spectroscopy measurements and pictures of a small patch of each object so that the PR2 can infer interactions with these objects.

Deeply Supervised Subspace Learning for Cross-Modal Material Perception of Known and Unknown Objects

A depth-supervised subspace cross-modal material retrieval model is trained to learn a common low-dimensional feature representation to capture the clustering structure among different modal features of the same class of objects.

A Framework for Sensorimotor Cross-Perception and Cross-Behavior Knowledge Transfer for Object Categorization

The results indicate that sensorimotor knowledge about objects can be transferred both across behaviors and across sensory modalities, such that a new robot can bootstrap its category recognition models without having to exhaustively explore the full set of objects.

Design of an Object Scanning System and a Calibration Method for a Fingertip-Mounted Dual-Modal and Dual Sensing Mechanisms (DMDSM)-based Pretouch Sensor for Grasping

The system design and ranging accuracy have been verified by physical experiments, and the collected data from seven types of common household objects have shown promising prospects of using DMDSM sensors in grasping.

Friction from Reflectance: Transfer Learning Approach

  • Piotr KickiK. Walas
  • Computer Science
    2019 4th International Conference on Robotics and Automation Engineering (ICRAE)
  • 2019
This paper is proposing a new approach of estimating friction coefficient from vision, i.e. reflectance images, based on transfer learning as well as the use of pre-trained networks to solve the friction estimation task.

Programmable Spectrometry: Per-pixel Material Classification using Learned Spectral Filters

This paper proposes the concept of programmable spectrometry for per-pixel material classification, where instead of sensing the HSI of the scene and then processing it, this is achieved using a computational camera with a programmable spectral response.



Material classification based on thermal properties — A robot and human evaluation

A system that classifies materials based on their thermal properties alone, minimising the amount of manipulation required and was compared with human performance in the task of classifying materials and was found to perform better.

Bayesian Exploration for Intelligent Identification of Textures

Performance of 99.6% in correctly discriminating pairs of similar textures was found to exceed human capabilities, and the method of Bayesian exploration developed and tested in this paper may generalize well to other cognitive problems.

Semi-Supervised Haptic Material Recognition for Robots using Generative Adversarial Networks

This work presents a semi-supervised learning approach for material recognition that uses generative adversarial networks (GANs) with haptic features such as force, temperature, and vibration and explores how well this approach can recognize the material of new objects.

Material recognition using tactile sensing

Material Recognition from Heat Transfer given Varying Initial Conditions and Short-Duration Contact

This work modeled the thermodynamics of the sensor in contact with a material as contact between two semi-infinite solids in order to enable robots to more efficiently sense their environments and take advantage of brief contact events over which they lack control.

Tactile-Data Classification of Contact Materials Using Computational Intelligence

Experimental results indicate that the CI tools are effective in dealing with the challenging problem of material classification, and the comparative analysis shows that SVM provides the best tradeoff between classification accuracy and computational complexity of the classification algorithm.

Deep learning for tactile understanding from visual and haptic data

This work proposes and explores a purely visual haptic prediction model that takes advantage of recent advances in deep neural networks by employing a unified approach to learning features for physical interaction and visual observations.

Vibrotactile Recognition and Categorization of Surfaces by a Humanoid Robot

This paper proposes a method for interactive surface recognition and surface categorization by a humanoid robot using a vibrotactile sensory modality. The robot was equipped with an artificial

Material classification by tactile sensing using surface textures

An application of machine learning to distinguish between seven different materials, based on their surface texture, including quality assurance and estimating surface friction during manipulation tasks is described.

Grounding semantic categories in behavioral interactions: Experiments with 100 objects