IROS 2016 Papers

The detection of consistent feature points in an image is fundamental for various kinds of computer vision techniques, such as stereo matching, object recognition, target tracking and optical flow computation. This paper presents an event-based approach to the detection of corner points, which benefits from the high temporal resolution, compressed visual information and low latency provided by an asynchronous neuromorphic event-based camera. The proposed method adapts the commonly used Harris corner detector to the event-based data, in which frames are replaced by a stream of asynchronous events produced in response to local light changes at μs temporal resolution. Responding only to changes in its field of view, an event-based camera naturally enhances edges in the scene, simplifying the detection of corner features. We characterised and tested the method on both a controlled pattern and a real scenario, using the dynamic vision sensor (DVS) on the neuromorphic iCub robot. The method detects corners with a typical error distribution within 2 pixels. The error is constant for different motion velocities and directions, indicating a consistent detection across the scene and over time. We achieve a detection rate proportional to speed, higher than frame-based technique for a significant amount of motion in the scene, while also reducing the computational cost.

Computing multiple contact points between geometric models evolved in a virtual 3D environment is central to many robotic simulation applications. While this task can be performed efficiently and robustly between complex polyhedra, using the exact analytic geometric models issued by CAD modelers still suffers from efficiency limitations. Yet models composed of smooth surfaces are required to ensure smooth contact constraints, thus avoiding possible numerical artifacts which may dramatically affect the behavior of the system in the case of functional contacts. This paper builds on the observation that industrial CAD models are mostly composed of simple surfaces to perform an off-line identification of similar features and build a bounding volume hierarchy in order to locate potential contacts. Those are then computed by dedicated analytic methods, or an iterative root-finder, depending on the actual geometric representations of the features. In the context of dynamic simulation of robotic tasks, our method exhibits interactive computation times while naturally providing better result accuracy than existing polyhedron-specific algorithms.

Matching 3D point clouds, a critical operation in map building and localization, is difficult with Velodyne-type sensors due to the sparse and non-uniform point clouds that they produce. Standard methods from dense 3D point clouds are generally not effective. In this paper, we describe a feature-based approach using Principal Components Analysis (PCA) of neighborhoods of points, which results in mathematically principled line and plane features. The key contribution in this work is to show how this type of feature extraction can be done efficiently and robustly even on non-uniformly sampled point clouds. The resulting detector runs in real-time and can be easily tuned to have a low false positive rate, simplifying data association. We evaluate the performance of our algorithm on an autonomous car at the MCity Test Facility using a Velodyne HDL-32E, and we compare our results against the state-of-the-art NARF keypoint detector.

We address the problem of rendezvous between two agents in urban street networks. Specifically, we consider the case where the agents have variable speeds and they need to schedule a rendezvous or a meeting under uncertainty in their travel times. Examples of such a scenario range from everyday life where two people would like to coordinate a meeting while going from office to home; to a futuristic case where automated taxis would like to meet each other for load balancing passengers. The scheduling for such scenarios can easily become challenging with uncertainties such as delayed departures, road blocks due to construction or traffic congestion. Any solution for such a task is required to minimize the waiting time and the planning overhead. In this paper, we propose an algorithm that optimizes the total travel time and the waiting time for two agents to complete their respective paths from start to rendezvous and from rendezvous to goal locations subject to delays along their paths. We validate our approach with a street network database which has a cost associated with every query made to the database server. Thus our algorithm intelligently optimizes for rendezvous trajectories that effectively mitigate the scourge of traffic delays, while simultaneously limiting the number of queries through careful analysis of the informative value of each potential query.

In this work, we address the problem of robust data association for feature cloud matching. For matching two feature clouds observed at two different poses, we discover that the covariance matrix of the measurement prediction error can be written as the sum of a low rank matrix and a block diagonal matrix, if we assume that the features are observed independently at each pose. This special structure of the covariance matrix allows us to compute its inverse analytically and efficiently. Together with a good bookkeeping strategy, the complexity of the Joint Compatibility (JC) test is reduced to O(1). Contrary to the approximated JC test, ours is both exact and fast. Based on the efficient JC test algorithm and a branch and bound search procedure, we devise an algorithm, called Fast Joint Compatibility Branch and Bound (FastJCBB), to quickly obtain robust data association. The FastJCBB algorithm is essentially modified from the conventional Joint Compatibility Branch and Bound (JCBB) algorithm and both of these algorithms are able to produce exactly the same data association results. However, with the substantial improvement in the efficiency of JC tests, our FastJCBB algorithm is much faster than the conventional JCBB, especially when matching two large feature clouds. It is reported that our FastJCBB algorithm is more than 740 times faster than the conventional JCBB in carrying out one million JC tests when matching two clouds with about 100 features each. Since both FastJCBB and JCBB share the same branch and bound procedure in exploring the interpretation tree, the search complexity remains exponential. Our main contribution is the significant improvement in the efficiency of exploring each node of the interpretation tree.

This paper presents a method for 6D pose estimation from a single RGB image for complex texture-less objects. This class of objects are common in any environment but still challenging to deal with. This is due to the fact that the distribution of surface brightness makes difficult to compute interest points or appearance-based descriptors. Here we propose a novel part-based method using an efficient template matching approach where each template independently encodes the similarity function using a Forest trained over the templates. Moreover, accuracy is even more incremented by using a cascade of the learned forest. These templates forests together with the simplicity of the computed image features allow a quick estimate of the pose achieving real-time performance. Performance are demonstrated both on synthetic and real images with known ground truth.

Human-in-the-loop task involving soft contact has become common in robotic application, especially in physical human-robot interaction. In this task, it is required that robot would sense interactions with human by touching, as well as assess possibility of human approaching by proximity sensation. In addition, it is also essential to fabricate an interface so that human does not feel uncomfortable during physical interaction with robot. This paper presents an attempt on fabrication of sensing elements that can be utilized for construction of a soft interface (or a robotic skin). Each element is made from fabrics and soft materials that can sense both proximity and applied force from human's touch. In addition, each sensing element can sense the relative distance of conductive object (or human body) that is approaching the sensing element's surface, and the 2×2 contact force distribution when the object makes contact with the sensing element. By exploiting simultaneous measurement of capacitance, each fabric sensing element can smoothly switch the proximity mode and tactile mode based on position of the object. We also constructed a model that can predict variation of capacitance measurement of proximity and tactile modes during operation for further analysis. The methods and results presented in this paper can be extended to construct a larger scale of robotic skin for robot's body, and act as a platform for study human-robot interaction.

Self-localization using visual markers is widely used for precise control of Unmanned aerial vehicles (UAVs). This paper presents a novel method of 6-DoF pose estimation using a monocular camera and non co-planar markers. We conducted numerical simulations comparing pose estimation using co-planar markers and using non co-planar markers, and designed a non co-planar marker feasible for robotics applications. Our method uses a single CPU core and a field programmable gate array (FPGA) and can conduct fast and precise 6-DoF pose estimation at the same frequency of the camera input (100 fps). In addition, to show the performance of our method, we applied our 6-DoF pose estimation algorithm to a palm-sized quadcopter and flew it autonomously using only an on-board system with a monocular camera and three-axis gyro sensor. To the authors' knowledge, this is the first study on fully on-board autonomous flight control of an MAV in 6-DoF space without an accelerometer.

Many autonomous vehicles require precise localization into a prior map in order to support planning and to leverage semantic information within those maps (e.g. that the right lane is a turn-only lane.) A popular approach in automotive systems is to use infrared intensity maps of the ground surface to localize, making them susceptible to failures when the surface is obscured by snow or when the road is repainted. An emerging alternative is to localize based on the 3D structure around the vehicle; these methods are robust to these types of changes, but the maps are costly both in terms of storage and the computational cost of matching. In this paper, we propose a fast method for localizing based on 3D structure around the vehicle using a 2D representation. This representation retains many of the advantages of “full” matching in 3D, but comes with dramatically lower space and computational requirements. We also introduce a variation of Graph-SLAM tailored to support localization, allowing us to make use of graph-based error-recovery techniques in our localization estimate. Finally, we present real-world localization results for both an indoor mobile robotic platform and an autonomous golf cart, demonstrating that autonomous vehicles do not need full 3D matching to accurately localize in the environment.

Expressive motion can situate a robot's attitude in its task motions, illustrating real-time reactions. Inspired by acting movement training, we construct path shape features that layer expression into a mobile robot's motion traversal. Our video-study results show that simple variations of path shape and orientation can influence human perceptions of a robot's task, focus, and confidence. We further find that sequencing path features is a useful way to create expressions that are pinpointed in time without requiring changes in velocity. Our quantitative features represent the Laban Space Effort: using path shape and orientation along the path to communicate the direct or indirect attitude of the robot toward its target destination (acting vocabulary italicized). These features illustrate expressive or stylistic aspects of the robot's inner state, filling a gap in the pre-existing literature that has mostly focused on task legibility. Our future work will evaluate temporal and spatial robot motion features in explicit interaction contexts.

This paper presents a novel Local-Path Planning approach for an independent four-wheel steering (I4WS) mobile base. The approach smoothly drives and rotates the platform while still following a given path and avoiding obstacles. I4WS vehicles are omnidirectional vehicles, similar to shopping carts, if one neglects the time needed to steer the wheels. However, due to the danger of mechanical parts breaking while actuating the wheels, they are commonly controlled in a stop-and-go fashion, where the vehicle stops to turn its wheels perpendicular to a desired instantaneous center of curvature (ICC) before starting to move again. The approach overcomes this limitation and ensures an ICC-based kinematic constraint during continuous motion using a flat-input controller running at 100 Hz. Therefore, the trajectory of the ICC is both predictable and suitable for model predictive control (MPC). The MPC implemented generates optimized collision-free trajectories of up to several meters ahead at 10 Hz, given a set of points along a path and laser contour readings. Furthermore, the planner realized here is able to deal with additional constraints such as the vehicle's view direction to focus on attention points while driving. Experimental results highlight the capabilities of the approach on a simulated robot, using GazeboSim, and demonstrate its applicability for the field of service robotics.

Allowing a human to express topological requirements to a robot in language enables untrained users to guide robot movement without requiring the human to understand sophisticated robot algorithms. By using a homotopy class or classes to represent one or more topological requirements, we build a framework that helps a robot understand a human's intent. This paper reviews a homotopic decomposition method that is used to convert any path into a string, which allows homotopic path equivalence to be performed by comparing strings. We then integrate the Homotopic Distributed Correspondence Graph (HoDCG) to infer the homotopic constraint in the format of strings from a language instruction. Finally, we use a homotopic path-planning algorithm that finds the optimal paths for a given objective and homotopic constraint. Experiment results show how a language instruction is converted into a path driven by an implicit topological requirement.

This paper presents an approach for selecting the next best base position for a climbing robot so as to observe the highest information gain about the environment. The robot is capable of adhering to and moving along and transitioning to surfaces with arbitrary orientations. This approach samples known surfaces, and takes into account the robot kinematics, to generate a graph of valid attachment points from which the robot can either move to other positions or make observations of the environment. The information value of nodes in this graph are estimated and a variant of A* is used to traverse the graph and discover the most worthwhile node that is reachable by the robot. This approach is demonstrated in simulation and shown to allow a 7 degree-of-freedom inchworm-inspired climbing robot to move to positions in the environment from which new information can be gathered about the environment.

We are developing a natural language interface for human robot interaction that implements reasoning about deep semantics in natural language. To realize the required deep analysis, we employ methods from cognitive linguistics, namely the modular and compositional framework of Embodied Construction Grammar (ECG) [18]. Using ECG, robots are able to solve fine-grained reference resolution problems and other issues related to deep semantics and compositionality of natural language. This also includes verbal interaction with humans to clarify commands and queries that are too ambiguous to be executed safely. We implement our NLU framework as a ROS package and present proof-of-concept scenarios with different robots, as well as a survey on the state of the art in knowledge-based language HRI.

Maps are an important component of most robotic navigation systems and building maps under uncertainty is often referred to as simultaneous localization and mapping or SLAM. Most SLAM approaches start from scratch and build a map only based on their own observations and odometry information. In this paper, we address the problem of how additional information can be exploited, for example from OpenStreetMap. We extend the standard graph-based SLAM formulation by relating the nodes of the pose-graph with an existing map. As this paper suggests, we can relate the newly built maps with information from publicly available maps with the laser range finder data from the robot and in this way improve the map quality. We implemented and evaluated our approach using real world data taken in urban environments. We illustrate that our extension to graph-based SLAM provides better aligned maps and adds only a marginal computational overhead.

Robotic handling inside the SEM is well known, but mostly a teleoperated task, but automation of in-situ handling would be highly advantageous and could enable serious production and assembly in the SEM. However, reliable techniques who facilitate an automation are still lacking. Especially the positioning along the optical axis of the SEM remains as a major challenge. In this paper, an universal approach addressing this particular issue is proposed, allowing to assist and improve handling and assembly task. The approach exploits typical charging effects in the SEM and derives information about the positions of objects and end-effectors as well as about assembly statuses. Assembly accuracies better than 100nm are achieved and controlled approaching speeds of 10 μm/s.

This paper presents a principled experimental analysis of a variable autonomy control approach to mobile robot navigation. A Human-Initiative (HI) variable autonomy system is investigated, in which a human operator is able to switch the Level of Autonomy (LOA) between teleoperation (joystick control) and autonomous control (robot navigates autonomously towards waypoints selected by the human) on-the-fly. Our hypothesis is that the HI system will enable superior navigation performance compared to either teleoperation or autonomy alone, especially in scenarios where the performance of both the human and the robot may at times become degraded. We evaluate our hypothesis through carefully controlled and repeatable experiments using a significant number of human test-subjects.

The demand for flexible industrial robotic solutions that are able to accomplish tasks at different locations in a factory is growing more and more. When deploying mobile robots in a factory environment, the predictability and reproducibility of their behaviors become important and are often requested. In this paper, we propose an easy-to-use motion planning scheme that can take into account user preferences for robot navigation. The preferences are extracted implicitly from the previous experiences or from demonstrations and are automatically considered in the subsequent planning steps. This leads to reproducible and thus better to predict navigation behaviors of the robot, without requiring experts to hard-coding control strategies or cost functions within a planner. Our system has been implemented and evaluated on a simulated KUKA mobile robot in different environments.

Design optimization and fabrication of concentric tube robots are time consuming because of the complexity of their workspaces and the characteristics of the superelastic materials used to make them. This paper presents a procedure for the expeditious design and fabrication of a concentric tube robot for applications that require rapid tube preparation but have less complex design constraints. This procedure reduces a 3D workspace optimization problem to a 2D problem. The continuum robot includes a heat-shrink tube to reduce fabrication time and to give it a small radius of curvature. Experimental results illustrate the feasibility of the proposed procedure.