Learning grasp affordance densities

By Renaud Detry, Dirk Kraft, Oliver Kroemer, Leon Bodenhagen, Jan Peters, Norbert Krüger, Justus H. Piater. 2011

95 citations recorded by the source corpus.

We address the issue of learning and representing object grasp affordance models. We model grasp affordances with continuous probability density functions (grasp densities) which link object-relative grasp poses to their success probability. The underlying function representation is nonparametric and relies on kernel density estimation to provide a continuous model. Grasp densities are learned and refined from exploration, by letting a robot “play” with an object in a sequence of grasp-and-drop actions: the robot uses visual cues to generate a set of grasp hypotheses, which it then executes and records their outcomes. When a satisfactory amount of grasp data is available, an importance-sampling algorithm turns it into a grasp density. We evaluate our method in a largely autonomous learning experiment, run on three objects with distinct shapes. The experiment shows how learning increases success rates. It also measures the success rate of grasps chosen to maximize the probability of success, given reaching constraints.

Subfields: Artificial Intelligence, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

Citations in the corpus

Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection (2016, methodology): “Our approach is most closely related to recent work on self-supervised learning of grasp poses by Pinto and Gupta (2016), as well as earlier work on learning from autonomous experimentation (Detry et al., 2011; Montesano and Lopes, 2009).”
Learning to Manipulate Unknown Objects in Clutter by Reinforcement (2015, background): “Second, the robot needs to learn grasp affordances from the executed actions and their features (Detry et al. 2011).”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, background): “Representing and Synthesizing Skill Parameters: Similar to the predicate representations, these policy parameter arguments can be modeled as additional input features for the precondition classifier or as distributions over valid and invalid argument values [86, 180].”
Acquiring visual servoing reaching and grasping skills using neural reinforcement learning (2013, methodology): “Along the same lines of the implicit grasp success prediction used here, Detry et al. [26] learned affordance densities of objects, which were explored through pick-and-drop movements.”
Dex-Net 1.0: A cloud-based network of 3D objects for robust grasp planning using a Multi-Armed Bandit model with correlated rewards (2016, background): “Our work is also closely related to research on actively sampling grasps to build a statistical model of grasp quality from fewer examples [10], [25], [36].”