IROS 2016 Papers

In this paper, we propose a scene clustering algorithm which uses straight line features. Scenes are represented as nodes in the graph, and each connectivity between nodes is calculated by a pre-trained vocabulary tree. By applying a spectral clustering algorithm to the constructed graph, the scenes are partitioned into k groups where k is determined by the proposed estimation method. Instead of using the standard eigenvalue analysis, the optimal k is computed so that the partitioned graph becomes to have strong intra-class correlations while inter-class correlations are relatively weak. As a result, scenes are adaptively clustered according to the environmental changes. The clustering performance of the proposed method is quantitatively evaluated with three image sequences captured in challenging environments. Experimental comparisons demonstrate that our line-based algorithm outperforms existing algorithms utilizing other types of features as well as the produced scene clustering results are more human-like.

The crank wheel mechanism, consisting of a wheel mechanism and parallel links connected to each wheel, achieved high mobility and efficiency because it has both wheels and legs in a simple structure. However, each prior model of crank wheel mechanism has had shortcomings such as mass oscillation or fragile structure. In this paper, we propose a novel crank wheel mechanism, the “Eccentric Crank Rover”(ECR), which is an enhanced crank wheel mechanism with eccentric wheels. The eccentric wheels increase the under-body clearance, and change the body trajectory from straight to trochoid curve, which has the same shape as the crank legs but opposite phase trajectory. Thus, the body itself acts as a “second” crank leg. We experimentally confirmed higher step climbability, larger clearance, and lower cost of transport than other models such as normal wheel model, eccentric wheel model, and crank wheel model without eccentric wheel.

We consider the problem of generating dynamically feasible and safe plans for teams of aerial robots (quadrotors) while holding a fixed relative formation as well as transitioning between a sequence of formations. We extend the existing assignment and planning approaches for quadrotor teams to find minimal-time trajectories to enable team transition between non-rest initial and ending states while ensuring dynamic feasibility with respect to predefined kinematic, dynamic, and collision constraints. This work also presents a method for safe splitting and merging of robot formations according to input specification. The proposed methodology is capable of generating dynamically feasible and safe plans for teams of quadrotors in real time. We validate the performance of the proposed approach through various trials and scenarios conducted in simulation.

This work aims to design a Dynamic Surface Controller (DSC) in order to stabilize the nth order chained systems able to express several kinds of nonholonomic, nonlinear and underactuated systems as unicycle, car-like, be-steerable car, car with trailer, etc. These latter are difficult to handle because of the complexity of their kinematic model. We purpose the use of the DSC in order to avoid the explosion of terms yielded by the backstepping algorithm and leading to compute the high order derivatives of variables. From the Lyapunov stability analysis, we prove that the designed controls render the closed loop system ultimately Uniformly Bounded (UUB). The applicability and the effectiveness of the control method are shown through simulation results when applying the control method on a car-like robot expressed by a fourth order chained system.

When tracking multiple targets with autonomous cameras for 3D scene reconstruction, e.g., in sports, a significant challenge is handling the unpredictable nature of the targets' motion. Such a monitoring system must reposition according to the targets' movements and maintain satisfactory coverage of the targets. We propose an approximate, centralized approach for maximizing the visible boundary of dynamic targets using mobile cameras in a bounded 2D environment. Targets and obstacles translate, rotate, and deform independently, and cameras are only aware of the current position and shape of the targets and obstacles. Using current information, the environment is searched for better viewing positions, then cameras navigate to those positions while avoiding collisions with targets and obstacles. We present a benchmark and metrics to evaluate the performance of our method, and compare our approach to a simple gradient-based local method in several real-time simulations.

Mobility On Demand (MOD) systems are revolutionizing transportation in urban settings by improving vehicle utilization and reducing parking congestion. A key factor in the success of an MOD system is the ability to measure and respond to real-time customer arrival data. Real time traffic arrival rate data is traditionally difficult to obtain due to the need to install fixed sensors throughout the MOD network. This paper presents a framework for measuring pedestrian traffic arrival rates using sensors onboard the vehicles that make up the MOD fleet. A novel distributed fusion algorithm is presented which combines onboard LIDAR and camera sensor measurements to detect trajectories of pedestrians with a 90% detection hit rate with 1.5 false positives per minute. A novel moving observer method is introduced to estimate pedestrian arrival rates from pedestrian trajectories collected from mobile sensors. The moving observer method is evaluated in both simulation and hardware and is shown to achieve arrival rate estimates comparable to those that would be obtained with multiple stationary sensors.

Dynamic Movement Primitives (DMPs) are a generic approach for trajectory modeling in an attractor land-scape based on differential dynamical systems. DMPs guarantee stability and convergence properties of learned trajectories, and scale well to high dimensional data. In this paper, we propose DMP+, a modified formulation of DMPs which, while preserving the desirable properties of the original, 1) achieves lower mean square error (MSE) with equal number of kernels, and 2) allows learned trajectories to be efficiently modified by updating a subset of kernels. The ability to efficiently modify learned trajectories i) improves reusability of existing primitives, and ii) reduces user fatigue during imitation learning as errors during demonstration may be corrected later without requiring another complete demonstration. In addition, DMP+ may be used with existing DMP techniques for trajectory generalization and thus complements them. We compare the performance of our proposed approach against DMPs in learning trajectories of handwritten characters, and show that DMP+ achieves lower MSE in position deviation. We demonstrate in a second experiment that DMP+ can efficiently update a learned trajectory by updating only a subset of kernels. The update algorithm achieves modeling accuracy comparable to learning the adapted trajectory with the original DMPs.

We present a Visual Inertial Odometry system that enables the autonomous flight of Micro Aerial Vehicles in GPS denied and unstructured environments. The system relies on commercially available and affordable hardware both for sensing and computation. The algorithm runs in real time on an ARM based embedded micro-computer on-board an MAV. In experiments, we demonstrate the performance of the system both indoors and outdoors, in hand held an in-flight scenarios. The achieved accuracy of the experiments is competitive with other research which uses custom designed hardware and desktop-grade processors.

Performing manipulation tasks in human environments often requires coordinated motions. The Extended-Cooperative-Task Space (ECTS) presents a unified representation describing any type of coordinated motions of two end-effectors. This paper focuses on the specification of dual-arm motion tasks based on the ECTS representation and the evaluation of the performance of the specified tasks. We first examine how the ECTS motion variables can be used effectively to specify coordinated motions to manipulate different types of objects. Then we derive new performance indices that can be used to evaluate and optimize the configurations of dual-arm robot systems performing any type of coordinated-motion tasks. The experimental results of coordinated-motion tasks performed by a Baxter robot demonstrate intuitiveness and efficiency of ECTS-based task specifications. We also show that the proposed ECTS performance indices can enable redundancies in a dual-arm system to be effectively utilized for a larger range of workspace as well as desired control objectives.

We propose a learning-from-demonstration approach for grounding actions from expert data and an algorithm for using these actions to perform a task in new environments. Our approach is based on an application of sampling-based motion planning to search through the tree of discrete, high-level actions constructed from a symbolic representation of a task. Recursive sampling-based planning is used to explore the space of possible continuous-space instantiations of these actions. We demonstrate the utility of our approach with a magnetic structure assembly task, showing that the robot can intelligently select a sequence of actions in different parts of the workspace and in the presence of obstacles. This approach can better adapt to new environments by selecting the correct high-level actions for the particular environment while taking human preferences into account.

To operate in human environments, robots must be able to withstand external disturbances. Small disturbances can be stabilized through momentum regulation, but larger ones require steps to prevent falling. This work presents two new techniques for disturbance rejection. The first is an extension of divergent component of motion (DCM) and capture point tracking controllers that augments a PI feedback control law with a disturbance observer. This is used to estimate transient disturbances through momentum-rate-of-change error. For larger disturbances, we present a novel optimization-based framework based on the DCM dynamics that uses a quadratic program to compute the desired ground reaction forces and recovery step location. Using optimization gives a flexibility that enables planning angular-momentum-rate-of-change trajectories to help reduce recovery step length. We then illustrate the effectiveness of these methods with hardware and simulation experiments of the THOR humanoid.

In many multi-robot applications such as target search, environmental monitoring and reconnaissance, the multi-robot system operates semi-autonomously, but under the supervision of a remote human who monitors task progress. In these applications, each robot collects a large amount of task-specific data that must be sent to the human periodically to keep the human aware of task progress. It is often the case that the human-robot communication links are extremely bandwidth constrained and/or have significantly higher latency than inter-robot communication links, so it is impossible for all robots to send their task-specific data together. Thus, only a subset of robots, which we call the knowledge leaders, can send their data at a time. In this paper, we study the knowledge leader selection problem, where the goal is to select a subset of robots with a given cardinality that transmits the most informative task-specific data for the human. We prove that the knowledge leader selection is a submodular function maximization problem under explicit conditions and present a novel distributed submodular optimization algorithm that has the same approximation guarantees as the centralized greedy algorithm. The effectiveness of our approach is demonstrated using numerical simulations.

In this paper, we present deformable configuration control — a fully distributed algorithm that allows multiple robots to deform their configuration to avoid various shapes of obstacles while maintaining connectivity. Robots contacted with an obstacle first estimate the width of an obstacle by sharing bumped status of individual robots, and choose an appropriate obstacle avoidance scenarios between obstacle detouring and bouncing off a wall. Second, for both scenarios, robots switch their motion model to a parent following motion to avoid the obstacle. Finally, robots keep sensing the maximum tree angle to estimate whether they are completely escaped from the obstacle, and return to flock formation motion model. We provide theoretical analysis about maintaining connectivity while robots are avoiding an obstacle. Simulation results show a group of 36 robots successfully avoid three different kinds of obstacles including 1) a small bar, 2) a narrow corridor, and 3) a large wall while maintaining connectivity.

The advent of advanced tactile sensing technology triggered the development of methods to employ them for grasp evaluation, online slip detection, and tactile servoing. In contrast to recent approaches to slip detection, distinguishing slip from non-slip conditions, we consider the more difficult task of distinguishing different types of slippage. Particularly we consider an object pushing task, where forces can only be applied from the top. In that case, the robot needs to notice when the object successfully moves vs. when the object gets stuck while the finger slips over its surface. As an example, consider the task of pushing around a piece of paper. We propose and evaluate three different convolutional network architectures and proof the applicability of the method for online classification in a robot pushing task.

We propose a method for generating a configuration space path that closely follows a desired task space path despite the presence of obstacles. We formalize closeness via two path metrics based on the discrete Hausdorff and Frechet distances. Armed with these metrics, we can cast our problem as a trajectory optimization problem. We also present two techniques to assist our optimizer in the case of local minima by further constraining the trajectory Finally, we leverage shape matching analysis, the Procrustes metric, to compare with respect to only their shape.

Modeling and control of soft robots is an up-to-date and exciting area of research which has been tackled with complementary approaches so far. In this paper, we modify the existing continuum Cosserat approach optimizing it for soft robot arms which can be discretized in a finite number of sections and degrees of freedom. The resulting new piece-wise constant strain model extends the existing piece-wise constant curvature model by allowing torsion and shears strains which are fundamental to cope with out-of-the-plane external forces as appearing for example during ground locomotion. A first experimental comparison has been also conducted using one fluidic actuated leg of the soft crawler FASTT.

Considering perception as an observation process only is the very reason for which robotic perception methods are to date unable to provide a general capacity of scene understanding. Related work in neuroscience has shown that there is a strong relationship between perception and action. We believe that considering perception in relation to action requires to interpret the scene in terms of the agent's own potential capabilities. In this paper, we propose a Bayesian approach for learning sensorimotor representations through the interaction between action and observation capabilities. We represent the notion of affordance as a probabilistic relation between three elements: objects, actions and effects. Experiments for affordances discovery were performed on a real robotic platform in an unsupervised way assuming a limited set of innate capabilities. Results show dependency relations that connect the three elements in a common frame: affordances. The increasing number of interactions and observations results in a Bayesian network that captures the relationships between them. The learned representation can be used for prediction tasks.

We propose a method to induce a realistic two-hand manipulation of virtual objects by representing their reactions to multiple contacts of hands and fingers in a unified way. One of the key problems in achieving physics-based virtual interaction is that there is no accurate haptic feedback to the user's hands and fingers, which must introduce a mismatch between the real hands and their virtual counterparts. This problem becomes more serious when multiple hands or fingers are involved to a single virtual object, for example, a hand-over situation from one hand to the other. The key contribution of this paper is to represent and describe the physical status of a virtual object in the local object space and to change the status in the object point of view in order to avoid its drastic changes by the physical-virtual mismatch. The experiments compare the performance of the proposed method with objects in different shapes when using 2D and 3D visualization devices.

In this paper we introduce a particularly relevant class of aerial manipulators that we name protocentric. These robots are formed by an underactuated aerial vehicle, a planar-Vertical Take-Off and Landing (PVTOL), equipped with any number of different parallel manipulator arms with the only property that all the first joints are attached at the Center of Mass (CoM) of the PVTOL, while the center of actuation of the PVTOL can be anywhere. We prove that protocentric aerial manipulators (PAMs) are differentially flat systems regardless the number of joints of each arm and their kinematic and dynamic parameters. The set of flat outputs is constituted by the CoM of the PVTOL and the absolute orientation angles of all the links. The relative degree of each output is equal to four. More amazingly, we prove that PAMs are differentially flat even in the case that any number of the joints are elastic, no matter the internal distribution between elastic and rigid joints. The set of flat outputs is the same but in this case the total relative degree grows quadratically with the number of elastic joints. We validate the theory by simulating object grasping and transportation tasks with unknown mass and parameters and using a controller based on dynamic feedback linearization.