The combination of collaborative robots and end-to-end AI, promises flexible automation of human tasks in factories and warehouses. However, such promise seems a few breakthroughs away. In the meantime, humans and cobots will collaborate helping each other. For these collaborations to be effective and safe, robots need to model, predict and exploit human's intents for responsive decision making processes. Approximate Bayesian Computation (ABC) is approach to perform probabilistic predictions upon uncertain quantities. ABC includes priors conveniently, leverages sampling algorithms for inference and is flexible to benefit from complex models, e.g. via simulators. However, ABC is known to be computationally too intensive to run at interactive frame rates required for effective human-robot collaboration tasks. In this paper, we formulate human intent prediction as an ABC problem and describe two key performance innovations which allow computations at interactive rates. Our real-world experiments with a collaborative robot set-up, demonstrate the viability of our proposed approach. Experimental evaluations convey the advantages and value of human intent prediction for packing cooperative tasks. Qualitative results show how anticipating human's intent improves human-robot collaboration without compromising safety. Quantitative task fluency metrics confirm the qualitative claims.

This article introduces a minimal-cost fully automatic method to generate 3D object models with unlimited resolution from a single aligned image for arbitrary revolution objects. In this method, the scope is focused on objects with complex (non-lambertian) materials that have intricate free-form revolution surfaces. The method is solely based on a single 2D image extracting two cues: The object's external contour is robustly and precisely extracted by sub-pixel graph expansion followed by piecewise radial regression producing a continuous regularized function called ∞-resolution contour. The object's internal skeleton is efficiently approximated by pixel-wise MAT (medial-axis-transform). The MAT naturally exposes the object's rotational invariants. The integration of the ∞-resolution contour with the invariant rotation axis is done by a tensor product. When the resulting compact and implicit ∞-resolution 3D object model is tessellated at a given target resolution, the tensor product generates mesh vertices, normals and texture coordinates for realistic models. This method achieves unprecedented high quality 3D models with ultra-efficient serialization. These claims are supported by an extensive evaluation using data sets with ground-truth metrics for volume, area, file-size, and execution time.

We consider the problem of grasp and manipulation planning when the state of the world is only partially observable. Specifically, we address the task of picking up unknown objects from a table top. The proposed approach to object shape prediction aims at closing the knowledge gaps in the robot's understanding of the world. A completed state estimate of the environment can then be provided to a simulator in which stable grasps and collision-free movements are planned. The proposed approach is based on the observation that many objects commonly in use in a service robotic scenario possess symmetries. We search for the optimal parameters of these symmetries given visibility constraints. Once found, the point cloud is completed and a surface mesh reconstructed. Quantitative experiments show that the predictions are valid approximations of the real object shape. By demonstrating the approach on two very different robotic platforms its generality is emphasized.

The adoption of robots for service tasks in natural environments calls for the use of sensors to allow manipulation of objects under imperfect environment knowledge and the use of knowledge transfer from humans. This paper addresses these challenges by proposing a new abstraction architecture for embodiment independent sensor-based control of manipulation. The aim is to address three specific challenges: hardware independent control of manipulation, use of sensors to alleviate problems of complexity and uncertainty of the environment, and ease of transferring knowledge over different embodiments through a hierarchical abstract representation of manipulation skills. The proposed abstraction architecture is demonstrated for hardware independence and failure detection on two different manipulator platforms.

This paper addresses the problem of robot grasping in conditions of uncertainty. We propose a grasp controller that deals robustly with this uncertainty using feedback from different contact-based sensors. This controller assumes a description of grasp consisting of a primitive that only determines the initial configuration of the hand and the control law to be used. We exhaustively validate the controller by carrying out a large number of tests with different degrees of inaccuracy in the pose of the target objects and by comparing it with results of a naive grasp controller.