IROS 2015 论文列表

We consider the problem of a human-following robot in which a human is equipped with a low-fidelity odometry sensor and a robot follows the human leader - often lagging well behind and out of visual contact with the human. The challenge is for the robot to determine the path taken by the human, despite the relatively noisy odometry data available. Such a system is useful in a “pack mule” application, where the robot carries a heavy load for the human. Our key idea is to equip the robot with sensors allowing it to build a map, and to use observations of the environment structure to constrain the path of the human. We propose and evaluate several approaches: a particle filter method that extends monte-carlo localization approaches, and a multi-hypothesis maximum-likelihood approach based on stochastic gradient descent optimization that efficiently clusters similar trajectories. We demonstrate that our proposed approaches are able to track human trajectories in several synthetic and real-world datasets.

This work demonstrates the use of depth panoramas in the construction of detailed 3D models of extended environments. The paper describes an approach to the acquisition of such panoramic images using a robotic platform that collects sequences of depth images with a commodity depth sensor. These sequences are stitched into panoramic images that efficiently capture scene information while reducing noise in the captured imagery. Scene structure is extracted from the panoramas and used to register a collection of panoramas to obtain coverage of extended areas. The presented approach maintains fine geometry detail in 3D reconstructions, uses a modest amount of memory, and enjoys excellent scaling characteristics across large environments.

This paper proposes a method of map merging using cycle consistency check and spanning tree selection by RANSAC. Key issues in map merging are the transformation of the coordinate frame of each submap to a global coordinate frame and the rejection of outliers (false arcs) in a pose graph. The proposed method reduces outliers using cycle consistency constraints. To cope with the exponential increase of cycles, cycle consistency check is applied iteratively with a limited number of chordless cycles. Then, spanning trees having only inlier arcs are selected by RANSAC to find the transformation of each submap to a global coordinate frame. Experiments using public datasets show that the proposed method successfully obtained good spanning trees at high outlier ratio.

In this paper, we propose a hybrid FUZZY P + ID controller to stabilize the towing cable tension between a cable ship and a burying system. First, we develop the model of a winch system driven by valve-controlled hydraulic motors and evaluate the step responses yielded by the conventional PID and the proposed FUZZY P+ID controllers using simulations. The comparative studies show that the control performance yielded by the FUZZY P+ID controller is superior. We replace the existing PID controller implemented on the towing winch with the FUZZY P + ID controller for burying cable tasks at speed up to 1 knot under sea waves with significant variations (peak-to-peak) from 1.5 to 2.5 meters. The real applications demonstrate that the FUZZY P + ID controller is much more robust than the conventional PID controller.

This paper presents an optimal design strategy for magnetic targeting of therapeutic drugs. In this study, to maximize the effect of the treatment and minimize adverse effects on the patient, mathematical models have been developed to find the number and the size of the boluses with respect to the growth of a tumor. Using these models, control strategies are developed to establish a schedule that allows the physician to administer the medication while respecting borne by the patient doses. To transport the drugs, we use therapeutic magnetic boluses composed of magnetic particle aggregates as navigable agents controlled by magnetic gradients. Based on the optimal design of the bolus, an experimental investigation is carried out in millimeter-sized fluidic artery vessels to demonstrate the steerability of the magnetic bolus under different velocity, shear-stress and trajectory constraints with a laminar viscous fluidic environment.

We propose a machine learning (ML)-inspired approach to estimate the relevant region of a motion planning problem during the exploration phase of sampling-based path-planners. The algorithm guides the exploration so that it draws more samples from the relevant region as the number of iterations increases. The approach works in two steps: first, it predicts if a given sample is collision-free (classification phase) without calling the collision-checker, and it then estimates if it is a promising sample, i.e., if it has the potential to improve the current best solution (regression phase), without solving the local steering problem. The proposed exploration strategy is integrated to the RRT# algorithm. Numerical simulations demonstrate the efficiency of the proposed approach.

We present a system for determining a global solution for the relative poses between multiple sensors with different modalities and varying fields of view. The final calibration result produces a tree of transforms rooted at a single sensor that allows the fusion of the sensor streams into a shared coordinate frame. The method differs from other approaches by handling any number of sensors with only minimal constraints on their fields of view, producing a global solution that is better than any pairwise solution, and by simplifying the data collection process through automatic data association.

For intelligent robots to solve real-world tasks, they need to manipulate multiple objects, and perform diverse manipulation actions apart from rigid transfers, such as pushing and sliding. Planning these tasks requires discrete changes between actions, and continuous, collision-free paths that fulfill action-specific constraints. In this work, we propose a multi-modal path planner, named MOPL, which accepts generic definitions of primitive actions with different types of contact manifolds, and randomly spans its search trees through these subspaces. Our evaluation shows that this generic search technique allows MOPL to solve several challenging scenarios over different types of kinematics and tools with reasonable performance. Furthermore, we demonstrate MOPL by solving and executing plans in two real-world experimental setups.

Sensorimotor coupling is ubiquitous in living organisms. Sensory and motor systems are utterly useless if left without the presence of the other. One crucial faculty that organisms have developed with tremendous ecological advantages is the ability to discern between the origins of perceptual input as being originated by the environment or the organism itself, provided by resource efficient sensor and motor systems. This ability has been shown to be implemented through a specialized circuit (forward model) receiving a copy of the motor command (corollary discharge). We propose a fast method to derive a resource constrained forward model by framing sensorimotor coupling as a low-rank approximation of an overly detailed forward model. By framing the problem as a factorization approach we can resort to currently available off-the-shelf solvers for matrix factorization. We experimentally show that by solving the problem as a low-rank approximation we obtain more than an order of magnitude speed up relatively to minimizing the objective function with gradient descent methods. The development of resource constrained and ecologically adapted sensorimotor systems is essential for the deployment of low-cost energy efficient autonomous robots for the execution of specific tasks in particular environments.

While humanoid robots and various associated algorithms for gait control and motion planning have been considered in the prior literature, the possibility of utilizing a low-profile crawling gait to operate in tight spaces (e.g., going under obstacles) has not been considered previously and enables new capabilities of the humanoid robot. In this paper, a new gait for humanoid robots is presented that enables humanoid motion in tight spaces that are vertically constrained. The gait is laterally symmetric and utilizes a cooperative motion of both the hands and feet. The addition of this gait expands the set of environments that can be handled by the humanoid robot. The efficacy of the proposed gait is demonstrated experimentally on a NAO humanoid robot.

This paper presents a work on mapping the use of space by humans in long periods of time. Daily geometric maps with the same coordinate frame were generated with SLAM, and in a similar manner, daily affordance density maps (places people use) were generated with the output of a human tracker running on the robot. The contribution of the paper is two-fold: an approach to detect geometric changes to cluster them in similar geometric configurations and the building of geometric and affordance composite maps on each cluster. This approach avoids the loss of long term retrieved information. Geometric similarity was computed using a normal distance approach on the maps. The analysis was performed on data collected by a mobile robot for a period of 4 months accumulating data equivalent to 70 days. Experimental results show that the system is capable of detecting geometric changes in the environment and clustering similar geometric configurations.

Traversable region detection is important for autonomous visual navigation of mobile robots. Only short range traversable regions can be detected using traditional methods based on stereo vision because of the limited image resolution and baseline of stereo vision. In this paper, we propose a novel method to detect long range traversable regions without using any supervised or self-supervised learning process. Superpixels are clustered using an improved spectral clustering algorithm to segment the image effectively, and after integrating short range traversable region detection based on u-v-disparity, the traversable region can be extended to long range naturally. The experimental results show that the proposed method works well in different outdoor/field environments, and the detecting range can be improved greatly in comparison with traditional methods. Furthermore, the proposed superpixels clustering algorithm can also be applied in other robot vision tasks like road detection and object recognition.

Jumping is one among successful locomotion gaits for small scale robots to overcome obstacles and traverse rough terrains. Design of a jumping robot features many engineering challenges; some of them may be solved by seeking inspiration and solutions in nature. This paper entails the conceptual and detailed design of TAUB, a miniature jumping robot whose design was inspired by the Desert Locust's legs. The jumping mechanism consist a pair of legs each with two equal segments, mimicking the femur and tibia, joined by torsion springs, mimicking the Semi-Lunar Process (SLP) of the Locust's legs. The jumping legs enable relatively large opening angle of 150 degrees, similar to the locust, allowing using softer springs and increase the energy storing and jumping performance. The 22. 6 gr autonomous robot prototype constructed outperforms other existing robots, by jumping to a height of 3. 1 m covering a distance of 3 m.

We focus on a mobile robot who must keep itself localized while closely following another robot or human. This problem has many real-world applications including that of a co-bot engaged in a follow-the-leader behavior or a robot that is participating in a convoy. If the robot is expected to eventually break away and reach its own goal, then the robot must stay self-localized. A key challenge for localization while tailing another is the extreme and persistent occlusion of the robot's sensors by the dynamic obstacle in front of it that is not modeled in its map. Current Monte Carlo localization (MCL) methods use sensor models with random noise, which are inadequate under such occlusion. We utilize a particle filter that simultaneously tracks the subject robot and the leader. We introduce novel particle weighting and adaptive sampling schemes that significantly improve the follower's localization. The result is a robust and adaptive MCL for applications involving persistent occlusion.

Link elasticity is commonly understood to be a detrimental side-effect of imperfect mechanical designs of robotic arms and comparable machinery. In contrast to this notion, this paper demonstrates a novel approach to exploit intrinsic robot link compliance in order to estimate a priori unknown payload masses, measure and also control end effector forces. In this way, the intrinsic link elasticity can be seen as an enabler for new sensing and control capabilities instead of a purely detrimental effect.

This paper presents the design and experimental validation of a low weight compliant arm for aerial manipulation, which uses a linear actuator to move the elbow joint. The proposed arm design allows for estimating payload mass, as well as detecting collisions against obstacles in the forearm. The compliant arm design is intended to be applied in aerial manipulation and transportation, being mounted on the base of an Unmanned Aerial Vehicle (UAV) like a multirotor or a small autonomous helicopter. The payload mass estimation can be useful in the adaptation and stabilization of UAV altitude and attitude controllers when objects of unknown weight are grasped by the arm, while compliance reduces the effect of physical interaction when lifting an object on UAV stability. The paper addresses the mechanical design and construction of a three DoF arm prototype (elbow, wrist roll and pitch), deriving the equations for estimating the payload mass. A set of experiments for payload mass estimation and collision detection and reaction have been performed to test the validity of the approach.

The fidelity with which series elastic actuators (SEAs) render desired impedances is important. Numerous approaches to SEA impedance control have been developed under the premise that high-precision actuator torque control is a prerequisite. Indeed, the design of an inner torque compensator has a significant impact on actuator impedance rendering. The disturbance observer (DOB) based torque control implemented in NASA's Valkyrie robot is considered here and a mathematical model of this torque control, cascaded with an outer impedance compensator, is constructed. While previous work has examined the impact a disturbance observer has on torque control performance, little has been done regarding DOBs and impedance rendering accuracy. Both simulation and a series of experiments are used to demonstrate the significant improvements possible in an SEA's ability to render desired dynamic behaviors when utilizing a DOB. Actuator transparency at low impedances is improved, closed loop hysteresis is reduced, and the actuator's dynamic response to both commands and interaction torques more faithfully matches that of the desired model. All of this is achieved by leveraging DOB based control rather than increasing compensator gains, thus making improved SEA impedance control easier to achieve in practice.

Manipulation of deformable objects is a widely applicable but challenging task in robotics. One promising nonparametric approach for this problem is trajectory transfer, in which a non-rigid registration is computed between the starting scene of the demonstration and the scene at test time. This registration is extrapolated to find a function from ℝ3 to ℝ3, which is then used to warp the demonstrated robot trajectory to generate a proposed trajectory to execute in the test scene. In prior work [1] [2], only depth information from the scenes has been used to compute this warp function. This approach ignores appearance information, but there are situations in which using both shape and appearance information is necessary for finding high quality non-rigid warp functions. In this paper, we describe an approach to learn relevant appearance information about deformable objects using deep learning, and use this additional information to improve the quality of non-rigid registration between demonstration and test scenes. Our method better registers areas of interest on deformable objects that are crucial for manipulation, such as rope crossings and towel corners and edges. We experimentally validate our approach in both simulation and in the real world, and show that the utilization of appearance information leads to a significant improvement in both selecting the best matching demonstration scene for a given test scene, and finding a high quality non-rigid registration between those two scenes.

This paper examines the design of a parallel spring-loaded actuated linkage intended for dynamically dexterous legged robotics applications. Targeted at toe placement in the sagittal plane, the mechanism applies two direct-drive brushless dc motors to a symmetric five bar linkage arranged to power free tangential motion and compliant radial motion associated with running, leaping, and related agile locomotion behaviors. Whereas traditional leg design typically decouples the consideration of motor sizing, kinematics and compliance, we examine their conjoined influence on three key characteristics of the legged locomotion cycle: transducing battery energy to body energy during stance; mitigating collision losses upon toe touchdown; and storing and harvesting prior body energy in the spring during stance. This analysis leads to an unconventional design whose “knee” joint rides above the “hip” joint. Experiments demonstrate that the resulting mechanism can deliver more than half again as much kinetic energy to the body (or more than double the kinetic energy if the full workspace is used), and offers a five-fold increase in energy storage and collision efficiency relative to the conventional design.