ICRA 2018 论文列表

The structural features in Manhattan world encode useful geometric information of parallelism, orthogonality and/or coplanarity in the scene. By fully exploiting these structural features, we propose a novel monocular SLAM system which provides accurate estimation of camera poses and 3D map. The foremost contribution of the proposed system is a structural feature-based optimization module which contains three novel optimization strategies. First, a rotation optimization strategy using the parallelism and orthogonality of 3D lines is presented. We propose a global binding method to compute an accurate estimation of the absolute rotation of the camera. Then we propose an approach for calculating the relative rotation to further refine the absolute rotation. Second, a translation optimization strategy leveraging coplanarity is proposed. Coplanar features are effectively identified, and we leverage them by a unified model handling both points and lines to calculate the relative translation, and then the optimal absolute translation. Third, a 3D line optimization strategy utilizing parallelism, orthogonality and coplanarity simultaneously is proposed to obtain an accurate 3D map consisting of structural line segments with low computational complexity. Experiments in man-made environments have demonstrated that the proposed system outperforms existing state-of-the-art monocular SLAM systems in terms of accuracy and robustness.

In this paper we present the design of a fin-like dielectric elastomer actuator (DEA) that drives a miniature autonomous underwater vehicle (AUV). The fin-like actuator is modular and independent of the body of the AUV. All electronics required to run the actuator are inside the 100 mm long 3D-printed body, allowing for autonomous mobility of the AUV. The DEA is easy to manufacture, requires no pre-stretch of the elastomers, and is completely sealed for underwater operation. The output thrust force can be tuned by stacking multiple actuation layers and modifying the Young's modulus of the elastomers. The AUV is reconfigurable by a shift of its center of mass, such that both planar and vertical swimming can be demonstrated on a single vehicle. For the DEA we measured thrust force and swimming speed for various actuator designs ran at frequencies from 1 Hz to 5 Hz. For the AUV we demonstrated autonomous planar swimming and closed-loop vertical diving. The actuators capable of outputting the highest thrust forces can power the AUV to swim at speeds of up to 0.55 body lengths per second. The speed falls in the upper range of untethered swimming robots powered by soft actuators. Our tunable DEAs also demonstrate the potential to mimic the undulatory motions of fish fins.

Legged robots have many potential applications in real-world scenarios where the tasks are too dangerous for humans, and compliance is needed to protect the system against external disturbances and impacts. In this paper, we propose a model-based controller for hierarchical tasks of legged systems subject to external disturbance. The control framework is based on projected inverse dynamics controller, such that the control law is decomposed into two orthogonal subspaces, i.e., the constrained and the unconstrained subspaces. The unconstrained component controls multiple desired tasks with impedance responses. The constrained space controller maintains the contact subject to unknown external disturbances, without the use of any force/torque sensing at the contact points. By explicitly modelling the external force, our controller is robust to external disturbances and errors arising from incorrect dynamic model information. The main contributions of this paper include (1) incorporating an impedance controller to control external disturbances and allow impedance shaping to adjust the behaviour of the motion under external disturbances, (2) optimising contact forces within the constrained subspace that also takes into account the external disturbances without using force/torque sensors at the contact locations. The techniques are evaluated on the ANYmal quadruped platform under a variety of scenarios.

How to divide floor plans or navigation maps into semantic representations, such as rooms and corridors, is an important research question in fields such as human-robot interaction, place categorization, or semantic mapping. While most works focus on segmenting robot built maps, those are not the only types of map a robot, or its user, can use. We present a method for segmenting maps from different modalities, focusing on robot built maps and hand-drawn sketch maps, and show better results than state of the art for both types. Our method segments the map by doing a convolution between the distance image of the map and a circular kernel, and grouping pixels of the same value. Segmentation is done by detecting ripple-like patterns where pixel values vary quickly, and merging neighboring regions with similar values. We identify a flaw in the segmentation evaluation metric used in recent works and propose a metric based on Matthews correlation coefficient (MCC). We compare our results to ground-truth segmentations of maps from a publicly available dataset, on which we obtain a better MCC than the state of the art with 0.98 compared to 0.65 for a recent Voronoi-based segmentation method and 0.70 for the DuDe segmentation method. We also provide a dataset of sketches of an indoor environment, with two possible sets of ground truth segmentations, on which our method obtains an MCC of 0.56 against 0.28 for the Voronoi-based segmentation method and 0.30 for DuDe.

We propose a method for optimizing control policies for assistive lower-limb devices. The method frames parameter selection as a dueling bandits problem in which a user indicates his or her qualitative preferences between pairs of parameter sets chosen from a library. We generate the library through an offline optimization procedure that seeks to reproduce the varied gaits of healthy human subjects. By separating the parameter selection process into online and offline portions, the method can handle high-dimensional parameter spaces and produces policies that can generalize to different gait scenarios such as speed variation. We evaluate the method on five subjects walking on a powered knee and ankle prosthesis governed by a neuromuscular control policy that has 43 parameters. We find the five subjects preferred four different parameter sets from the library and that the resulting optima resemble intact subject gait data. This result suggests the offline portion of the optimization method indeed produces control parameters that can adapt to different gaits. Moreover, we find that for three out of the four parameter sets we tested, the procedure also generates parameters that improve the ability of the prosthesis to adapt to increasing gait speed by increasing ankle net work production. The results encourage further research and exploration in clinical settings toward advanced prosthesis controls that employ online learning.

Accurate yield estimation is very important for improving the vineyard management, the quality of the grapes and the health of the vines. The most common systems use RGB image processing for achieving a good estimation. In order to collect images, robots or farming vehicles can be equipped with a RGB camera. In this paper, we propose a low-cost autonomous system which can navigate through a vineyard while collecting grape pictures in order to provide a yield estimation. Our system uses only a laser scanner to detect the row and follows it until its end, then it navigates towards the next one, exploiting the knowledge of the vineyard. The navigation algorithm was tested both in simulation and in a real environment with good results. Furthermore, a yield estimation of two different grape varieties is presented.

Powered lower-limb orthoses and prostheses are attracting an increasing amount of attention in assisting daily living activities. To safely and naturally collaborate with human users, the key technology relies on an intelligent controller to accurately decode users' movement intention. In this work, we proposed an innovative locomotion recognition system based on depth images. Composed of a feature extraction subsystem and a finite-state-machine based recognition subsystem, the proposed approach is capable of capturing both the limb movements and the terrains right in front of the user. This makes it possible to anticipate the detection of locomotion modes, especially at transition states, thus enabling the associated wearable robot to deliver a smooth and seamless assistance. Validation experiments were implemented with nine subjects to trace a track that comprised of standing, walking, stair ascending, and stair descending, for three rounds each. The results showed that in steady state, the proposed system could recognize all four locomotion tasks with approximate 100% of accuracy. Out of 216 mode transitions, 82.4% of the intended locomotion tasks can be detected before the transition happened. Thanks to its high accuracy and promising prediction performance, the proposed locomotion recognition system is expected to significantly improve the safety as well as the effectiveness of a lower-limb assistive device.

Pose Graph Optimization (PGO) is an important optimization problem arising in robotics and machine vision applications like 3D reconstruction and 3D SLAM. Each node of pose graph corresponds to an orientation and a location. The PGO problem finds orientations and locations of the nodes from relative noisy observation between nodes. Recent investigations show that well-known iterative PGO solvers need good initialization to converge to good solutions. However, we observed that state-of-the-art initialization methods obtain good initialization only in low noise problems, and they fail in challenging problems having more measurement noise. Consequently, iterative methods may converge to bad solutions in high noise problems. In this paper, a new method for obtaining orientations in the PGO optimization problem is presented. Like other well-known methods the initial locations are obtained from the result of a least-squares problem. The proposed method iteratively approximates the problem around current estimation and converts it to a least-squares problem. Therefore, the method can be seen as an iterative least-squares method which is computationally efficient. Simulation results show that the proposed initialization method helps the most well-known iterative solver to obtain better optima and significantly outperform other solvers in some cases.

Sensors to measure quadcopter ground effect are often relatively large, heavy, and require significant power, which restricts their applicability when it comes to small quadcopters and other aircraft that require lightweight and non-obstructive solutions. This paper presents the design of an elastomeric contact-resistance-based airflow sensor to measure ground effect with a mass of approximately 0.04 g and a power draw of 42 μW when in operation. It uses a rigid flap attached to the top of a flexible conductive pillar (CNT/PDMS), which deflects with varying winds speeds. A simple model is presented to describe expected trends between air flow speeds and sensor deflection and is compared with a wind tunnel characterization of the sensor for varying airflows. The sensor is characterized in a wind tunnel to identify a minimum airflow necessary for sensor functionality. Finally, sensors are attached to a Crazyflie 2.0 (Bitcraze) quadcopter and tested for performance in detecting ground effect, where there is a clear trend between sensor output and the intensity of the turbulent flow, related to proximity to ground and thrust level.

Hemiparetic gait after stroke is typically asymmetric and energetically inefficient. A major contributor to walking deficits is impaired paretic ankle function. Impaired paretic ankle plantarflexion (PF) reduces forward propulsion symmetry and impaired paretic ankle dorsiflexion (DF) diminishes ground clearance during swing. We have developed soft wearable robots (soft exosuits) to assist paretic PF and DF during walking after stroke. Through experimental studies with poststroke patients, we have demonstrated that exosuits can improve forward propulsion symmetry and ground clearance in walking, ultimately reducing the metabolic cost of walking. This paper presents an optimized soft exosuit aimed at use in clinical gait training for patients poststroke. The optimized exosuit is lightweight, easy to don and doff, and capable of efficiently delivering mechanical assistance to the paretic ankle. This paper focuses on the optimized controller that can deliver well-timed consistent ankle assistance to patients. A preliminary study was performed using this exosuit with three poststroke patients with heterogeneous gait patterns. Results showed that compared to a previously published controller, more consistent assistive force profiles could be delivered to individuals poststroke while consuming 50% less electrical power. Additionally, a preliminary biomechanical assessment was performed during overground walking.

Increasing research efforts have been made to improve the energy efficiency of variable impedance actuators (VIAs) through reduction of energy consumption. However, the harvesting of dissipated energy in such systems remains under-explored. This study proposes a novel variable damping module design enabling energy regeneration in VIAs by exploiting the regenerative braking effect of DC motors. The proposed damping module uses four switches to combine regenerative and dynamic braking, in a hybrid approach that enables energy regeneration without reduction in the range of damping achievable. Numerical simulations and a physical experiment are presented in which the proposed module shows an optimal trade-off between task-performance and energy efficiency.

This article presents the design and experimental evaluation of a prototype robotic platform for minimally invasive surgical procedures. The platform utilizes a hybrid actuation scheme, consisting of a 5 Degree-of-Freedom (DoF) servo-actuated manipulator for extra-operative and pivoting motion and a 4 DoF shape memory alloy actuated probe at the distal end, for intra-operative dexterity. The architecture targets thoracic and abdominal operations, with low interaction forces at the probe's end-effector. The system, runs under the Robot Operating System framework for easier deployment and development. Additional accompanying software is developed to aid the surgeon during deployment. Specifically, a Graphical User Interface employing modules controls for online parameter reconfiguration, operation mode switching while custom viewports for stereo imaging are implemented. Teleoperation is feasible with the integration of a haptic device. In-vitro evaluation of the robot is presented, to assess the maneuvering efficiency and further potential exploitation of the design.

As increasing attention is paid on human action recognition from skeleton data, this paper focuses on such tasks by proposing a hierarchical model to discover the structure information of body-parts involved in human actions. Considering human actions as simultaneous motions of different body-parts of the human skeleton, we propose a hierarchical model to simultaneously apply discriminative body-parts selection at a same scale and group coupling of bundles of body-parts at different scales, while we decompose the human skeleton into a hierarchy of body-parts of varying scales. To represent such hierarchy of body-parts, we accordingly build a hierarchical RRV (Rotation and Relative Velocity) descriptors. The hierarchical representations encoded by Fisher vectors of the hierarchical RRV descriptors are properly formulated into the hierarchical model via the proposed hierarchical mixed norm, to apply sparse selection of body-parts and regularize the structure of such hierarchy of body-parts. The extensive evaluations on three challenging datasets demonstrate the effectiveness of our proposed approach, which achieves superior performance compared to state-of-the-art results on different sizes of datasets, showing it is more widely applicable than existing approaches.

In this paper, a geometric and unified model of soft-rigid multi-body systems is presented, based on a discrete Cosserat approach of the soft-body dynamics. The model is in fact a generalization to soft and hybrid systems of the geometric theory of rigid robotics characterized by the exponential map. A generalization of the recursive Newton-Euler algorithm is also presented, able to solve inverse and forward dynamic problems with linear O(N) complexity. The proposed model provides several improvements with respect to the existing flexible multi-body models, which make it particularly suitable to study the dynamics of modern soft robots as shown for a multi-body system inspired by motile bacteria.

We present a new general framework for disassembly sequence planning. This framework is versatile allowing different types of search schemes (exhaustive vs. preemptive), various part separation techniques, and the ability to group parts, or not, into subassemblies to improve the solution efficiency and parallelism. This enables a truly hierarchical approach to disassembly sequence planning. We demonstrate two different search strategies using this framework that can either yield a single solution quickly or provide a spectrum of solutions from which an optimal may be selected. We also develop a method for subassembly identification based on collision information. Our results show improved performance over an iterative motion planning based method for finding a single solution and greater functionality through hierarchical planning and optimal solution search.

Autonomous driving requires 3D perception of vehicles and other objects in the in environment. Much of the current methods support 2D vehicle detection. This paper proposes a flexible pipeline to adopt any 2D detection network and fuse it with a 3D point cloud to generate 3D information with minimum changes of the 2D detection networks. To identify the 3D box, an effective model fitting algorithm is developed based on generalised car models and score maps. A two-stage convolutional neural network (CNN) is proposed to refine the detected 3D box. This pipeline is tested on the KITTI dataset using two different 2D detection networks. The 3D detection results based on these two networks are similar, demonstrating the flexibility of the proposed pipeline. The results rank second among the 3D detection algorithms, indicating its competencies in 3D detection.

The ability to build maps is a key functionality for the majority of mobile robots. A central ingredient to most mapping systems is the registration or alignment of the recorded sensor data. In this paper, we present a general methodology for photometric registration that can deal with multiple different cues. We provide examples for registering RGBD as well as 3D LIDAR data. In contrast to popular point cloud registration approaches such as ICP our method does not rely on explicit data association and exploits multiple modalities such as raw range and image data streams. Color, depth, and normal information are handled in an uniform manner and the registration is obtained by minimizing the pixel-wise difference between two multi-channel images. We developed a flexible and general framework and implemented our approach inside that framework. We also released our implementation as open source C++ code. The experiments show that our approach allows for an accurate registration of the sensor data without requiring an explicit data association or model-specific adaptations to datasets or sensors. Our approach exploits the different cues in a natural and consistent way and the registration can be done at framerate for a typical range or imaging sensor.

Recent advances in probe-based Confocal Laser Endomicroscopy (pCLE) enable real-time, in situ and in vivo tissue assessment at the micro scale. The limited field-of-view offered by pCLE necessitates the use of mosaicking to allow for accurate tissue characterization from the incoming image stream. However, mosaicking requires a series of contiguous good-quality images, which is particularly challenging because probe-tissue distance must be maintained within a very narrow working range at all times and probe-tissue contact force must be kept to a minimum so that tissue deformation is avoided. Robotic manipulation of the endomicroscopy probe has provided partial solution to these challenges, but sensorless approaches have not been thoroughly investigated up to date. This paper proposes a novel sensorless framework that uses a single non-reference image-quality metric to learn an approximation of tissue motion and subsequently track it. Moreover, a pCLE robotic tool for autonomous endomicroscopy scanning is designed and used for testing and validation purposes. Experiments on lens paper and ex vivo porcine tissue validate the philosophy of the framework.

We present a novel flying modular platform capable of grasping and transporting objects. It is composed of four cooperative identical modules where each is based on a quadrotor within a cuboid frame with a docking mechanism. Pairs of modules are able to fly independently and physically connect by matching their vertical edges forming a hinge. Four one degree of freedom (DOF) connections results in a one DOF four-bar linkage that can be used to grasp external objects. In this paper, we propose a decentralized method that allows the Flying Gripper to control its position, attitude and aperture angle. In our experiments, we tested the hovering performance for different aperture angles and with a grasped object. The performance for a closing and opening motion was also verified.