ICRA 2015 论文列表

Colonoscopy is the gold standard screening test for colorectal cancer. However, it can miss up to 22% of lesions. One way to reduce this misrate is by guiding the endoscopist's attention towards suspicious areas (red-flagging). In this paper, we present the proof-of-concept of a novel endoscopic imaging device for the scanning of tubular organs such as the colon and the esophagus. The envisaged concept works as an accessory for any conventional flexible endoscope using it as a rail. It can generate a red-flagged map of the whole organ by orbiting and sliding a radial array of optical sensors around and along the endoscope. This concept paves the way for a semi-automated concurrent red-flagging technique for colonoscopy that enhances the endoscopist's situational awareness and reduces misrate.

Co-robots, i.e. robots that work close to people, will need to account for the preferences and expectations of their human co-workers in executing trajectories or actions. Consistent, legible and predictable trajectories are a key factor in making humans comfortable around robots. In this work, we take a data-driven approach towards designing robot trajectories that are more acceptable to human co-workers and observers. We use an online survey to ask people to rate multiple robot trajectories generated in a variety of environments. We compute a large set of features for each trajectory, also taking into account environment information. We use a combination of the features and the survey ratings to learn a classifier that predicts the rating for a new trajectory based on the learned human-observer preferences. The classifier also helps identify and highlight the most important features that influence people's ratings of the trajectories. Finally, we discuss how a data-driven approach using the results of this analysis can be used to help design better trajectories that are more acceptable to people.

MRI-based navigation and propulsion of millirobots is a new and promising approach for minimally invasive therapies. The strong central field inside the scanner, however, precludes torque-based control. Consequently, prior propulsion techniques have been limited to gradient-based pulling through fluid-filled body lumens. This paper introduces a technique for generating large impulsive forces that can be used to penetrate tissue. The approach is based on navigating multiple robots to a desired location and using self-assembly to trigger the conversion of magnetic potential energy into sufficient kinetic energy to achieve penetration. The approach is illustrated through analytical modeling and experiments in a clinical MRI scanner.

Many control policies developed for legged robots are based on control of an underlying, simplified version of the dynamics of the robot. A good example is the Linear Inverted Pendulum Model (LIPM) which has become the standard control template for ZMP-based rigid robots. For compliant robots, this reduced order model is naturally the Spring-Loaded Inverted Pendulum (SLIP), which has proven to have many interesting traits that are potentially useful for control of full order robots. The methods proposed so far for this purpose are mainly focused on either matching the dynamics of the robot to those of SLIP, or following a SLIP-produced trajectory. These methods can be problematic, especially for underactuated systems. In the present work, we explore an opposite approach, by starting from SLIP and step-by-step constructing toward the full order robot. The goal is to detect and capture the essential stabilizing variables in the reduced order model that can potentially maintain their stabilizing effect in the full order robot, as well. Our initial investigations show that the proposed method provides excellent potentials for synthesizing stable gaits for underactuated compliant robots by use of a much simpler and more robust approach compared to the ones previously presented in the literature.

Minimally invasive robotic surgery techniques are becoming popular thanks to their enhanced patient benefits, including shorter recovery time, better cosmetic results and reduced discomforts. Less invasive procedures would be achieved with the use of Medical Capsule Robots (MCRs). These devices are characterized by low power requirements and small dimensions as well as uncompromising safety. MCRs operate wirelessly in abdominal Minimally Invasive Surgery (MIS) and Natural Orifice Transluminal Endoscopic Surgery (NOTES) or in the Gastrointestinal (GI) tract. The design process of MCRs, however, is expensive and time consuming. A platform for rapid prototyping MCRs is needed so that MCR researchers can reduce development costs and spend more time in studying innovative MCR applications. In this work, we introduce an open source modular platform geared toward rapid prototyping MCRs. To speed up the prototyping process, the MCR is programmed using TinyOS instead of bare-bone C. We present the hardware architecture of the platform, and the motivation for using TinyOS. To show the viability of TinyOS, we present results from an experiment involving sensing, actuation and wireless communication. This work lays the foundation for our future goal of building an integrated design environment for the design, analysis and simulation of MCRs.

This paper proposes a new visual controller to endow flying manipulators with the capability of cooperatively and automatically positioning an assembly part on a visual target. We consider two VToL UAVs each equipped with a robot manipulator and a video camera. The two manipulators are rigidly connected to an assembly part, consisting of bar in this case study, that needs to be automatically placed on its target pose. The control system we propose uses the images relayed from the two onboard cameras to simultaneously control the motions of the UAVs and the manipulators to achieve the task: positioning the bar as well as respecting the closed-chain constraint. The controller evolves in the image space, thus inheriting robustness with respect to calibration errors. It takes into account the under-actuation inherent to rotary-wing aerial vehicles as well as the redundancy of the whole system. Furthermore, by leveraging image moments the 3D relative motion between the carried assembly part and the environment is controlled, thus yielding to natural and practically-relevant movements. Numerical simulations have been carried out to verify the validity of the proposed approach.

This paper addresses the problem of ice-relative navigation of underwater robotic vehicles for oceanographic missions under moving sea ice. We define the ice-relative navigation problem and report a novel underwater robot vehicle, Nereid-UI, and its navigation system designed for under-ice oceanographic operations in ice-covered seas. The goal of Nereid-UI is to enable deployments in polar ocean regions traditionally considered difficult or impossible to access such as the ice-ocean interface in marginal ice zones, the water column of ice-covered seas, and the water column underlying ice shelves. This paper also reports the navigation results of Nereid-UI's first under-ice sea trials and science operations conducted under moving multi-year sea ice off the ice breaker Polarstern in the Arctic Ocean at 83° N 6° W - about 200 km NE of Greenland.

We present an approach for asymptotically optimal motion planning for kinodynamic systems with arbitrary nonlinear dynamics amid obstacles. Optimal sampling-based planners like RRT*, FMT*, and BIT* when applied to kinodynamic systems require solving a two-point boundary value problem (BVP) to perform exact connections between nodes in the tree. Two-point BVPs are non-trivial to solve, hence the prevalence of alternative approaches that focus on specific instances of kinodynamic systems, use approximate solutions to the two-point BVP, or use random propagation of controls. In this work, we explore the feasibility of exploiting recent advances in numerical optimal control and optimization to solve these two-point BVPs for arbitrary kinodynamic systems and how they can be integrated with existing optimal planning algorithms. We combine BIT* with a two-point BVP solver that uses sequential quadratic programming (SQP). We consider the problem of computing minimum-time trajectories. Since the duration of trajectories is not known a-priori, we include the time-step as part of the optimization to allow SQP to optimize over the duration of the trajectory while keeping the number of discrete steps fixed for every connection attempted. Our experiments indicate that using a two-point BVP solver in the inner-loop of BIT* is competitive with the state-of-the-art in sampling-based optimal planning that explicitly avoids the use of two-point BVP solvers.

In this paper we propose a framework combining techniques from sampling-based motion planning, machine learning, and trajectory optimization to address the kinodynamic motion planning problem in real-time environments. This framework relies on a look-up table that stores precomputed optimal solutions to boundary value problems (assuming no obstacles), which form the directed edges of a precomputed motion planning roadmap. A sampling-based motion planning algorithm then leverages such a precomputed roadmap to compute online an obstacle-free trajectory. Machine learning techniques are employed to minimize the number of online solutions to boundary value problems required to compute the neighborhoods of the start state and goal regions. This approach is demonstrated to reduce online planning times up to six orders of magnitude. Simulation results are presented and discussed. Problem-specific framework modifications are then discussed that would allow further computation time reductions.

In this paper we propose the fastest converging control policy (also known as deadbeat control) for walking with the bipedal spring-mass model, which serves as an abstraction of a robot on compliant legs. To fully leverage the passive dynamics of the system, the touchdown angle of the swing-leg is assigned as the only control input of the system. We show that two steps (or one stride) are necessary and sufficient to converge to target walking gaits. We first analyze the dynamics of the system to identify the limit cycles as well as the limitations of the control authority within the definition of walking. Then, we present the two-step deadbeat control policy that guarantees stability with the fastest possible convergence rate for the system. For each equilibrium gait, the basin of attraction in which this two-step control exists is a measure of the robustness of the system. The simulation results show that human-like walking gaits (double hump ground reaction force profile) have relatively large basins of attraction. Finally, we extend the policy to various energy levels to accommodate walking on uneven ground that has height changes. We show in simulation that the system indeed rejects various disturbances and converges to the desired equilibrium gait in two steps.

The Mechanically Adjustable Compliance and Controllable Equilibrium Position Actuator (MACCEPA) is a Variable Stiffness Actuator (VSA) where both equilibrium position and stiffness of the actuator can be controlled independently. It uses only one linear spring and has a simple design but its compactness is limited by the spring. For this reason a MACCEPA utilizing torsion spiral springs was designed, reducing the planar dimensions of the actuator. Torsion spiral springs are placed around the joint axis, allowing a more compact VSA in comparison to previous designs. To the authors' best knowledge, this is the first VSA based on torsion springs. This paper firstly presents the design of the actuator as the static equations and secondly discusses the design and production of the torsion spiral springs. The newly presented actuator is built and experiments are conducted to validate the model and feasibility of the torsion MACCEPA.

In this paper we address the problem of dense depth map estimation from sparse noisy range data to reconstruct large heterogeneous outdoor scenes. We propose a surface inpainting solution through energy minimisation with an adaptive selection of surface regularisers among a set of well known convex and non-convex regularisers. In fact, the selection of norm is pivotal with respect to the intrinsic surface characteristics. Our goal is to show how dense interpolation of sparse range data can be leveraged of more exotic and non-convex regularisers such as the log and logTGV [1] which can better capture the scene geometry. In contrast to state of the art solutions, we do not restrict ourselves to this set of norms, instead we search for the most apt norm for each semantically segmented part of the scene. Our energy model selection use Bayesian optimisation to learn the best choice of free parameters. This results in an adaptive model selection and the generalisation of well studied regularisation norms. We conclude with a detailed experimental analysis of our approach using a basis of four norms over a set of challenging outdoor scenes.

One of the major challenge in spine surgery is successfully cutting the desired bony structure while avoiding injury to the vital anatomy such as the spinal cord. This paper presents a vibration signal processing method to discriminate different types of tissue in robot-assisted spine surgery. During bone milling process, the tissue vibration signal measured by a laser displacement sensor is decomposed into some frequency sub-bands through the wavelet packet transform, and the harmonic component whose frequency is an integer times of the spindle frequency is obtained. The wavelet energy of the 1st, 2nd and 3rd harmonics is then used as input vector to an artificial neural network for discriminating tissue. To verify the effectiveness of the proposed method, extensive milling experiments are carried out on porcine spines. The experimental results indicate that the method produces up to 100% discrimination accuracy for the vertebra being cut and the spinal cord, and yields the success discrimination rate of more than 80% for the adjacent bony structure and the muscle, so the safety of robot-assisted spine surgery is improved.

In this paper, based on the apparent stiffness analysis of a dual tendon-sheath system, we propose a method for using a surgical instrument of the da Vinci surgical robotic system (from Intuitive Surgical Inc.) for estimation of mechanical properties of tissues. The performance of the method is experimentally evaluated by comparing tissues with different stiffnesses and by localizing tumors in an artificial tissue sample.

We describe an on-line observer that allows to monitor the soil-tire contact longitudinal stiffness of off-road vehicles equipped with low cost sensors. The knowledge of such a value is a milestone in the development of active security devices, or at least a mean to inform the driver about the longitudinal stability of his vehicle. Indeed, these vehicles are frequently subject to ground changes and high slopes, causing fatal losses of control - slipping, rollover, etc. - every year. The observer has been tested and validated on a grape harvester in various conditions and its output evaluated with respect to an estimation based on expensive but precise sensors. These experimental results show that the driver can be informed in real-time about the longitudinal stability of his vehicle and alerted about critical situations.

Skid-steered, tracked, tele-operated robots are used to perform high-risk critical missions such as bomb disposal under conditions deemed too risky to send a human. Often the robots carry heavy payloads that raise their centers of mass, increasing the risk of tip-over. Since it is often not feasible to send a human to right a toppled robot, tip-over is equivalent to mission failure. Hence, an autonomous behavior to prevent robot tip-over is desired. In this research, a simplified model of mobile robot dynamics permits separation of pitch and roll stabilization. Adaptive control is used to stabilize the appropriate angle based on the normalized tip-over measure. Experimental validation of this control is successfully demonstrated on an iRobot Packbot and a Segway RMP 440.

As an important branch of production facilities, automated manufacturing systems are a promising perspective of modern factories which become increasingly complex and flexible. From the perspective of development, time variable is a significant concern. In this paper, we propose a class of time-varying systems, Tv-S4R. After involving a global time information loop, it can represent realistic system whose manufacturing process varies with time variables. Event-based control is introduced for serving two important purposes. First, it exerts time-varying specifications on the system, i.e., different processes are executed at different time stages. Second, it prevents the system from reaching any first-met bad marking which will eventually result in deadlocks. We will show that the event-based control can enhance our specifications' expressivity capability compared to their state-based counterparts in terms of generalized mutual exclusion constraints. At last, supervisors are simplified according to the identification of dependence and independence of inequalities. An experimental study illustrates the application of Tv-S4R and the effectiveness of event-based control method.

There have been increasing interests in the robotics community in building smaller and more agile autonomous micro aerial vehicles (MAVs). In particular, the monocular visual-inertial system (VINS) that consists of only a camera and an inertial measurement unit (IMU) forms a great minimum sensor suite due to its superior size, weight, and power (SWaP) characteristics. In this paper, we present a tightly-coupled nonlinear optimization-based monocular VINS estimator for autonomous rotorcraft MAVs. Our estimator allows the MAV to execute trajectories at 2 m/s with roll and pitch angles up to 30 degrees. We present extensive statistical analysis to verify the performance of our approach in different environments with varying flight speeds.

Due to the poor controllability in hydrogels, Hydrogels-based assembly to form larger 3D complex shapes is still a big challenge. In this paper, we have reported a novel “bottom-up” method to fabricate three-dimensional (3D) magnetic alginate microfibers (MAMs) assemblies with complex shapes. Specifically, Alginate microfibers encapsulating Fe3O4 magnetic nanoparticles (MNs) and fibroblasts (NIH/3T3) have been spun using microfluidic method with “pinch-off” scheme. Experimental results show that the MAMs can respond quickly to the magnetic field, enabling their enhanced controllability. The magnetic assembly system is constructed by PDMS microfluidic device, 3D supporting model adhered on the bottom surface of dish filled with deionized water and magnet. The microfluidic “printing“ and magnetic deposition in magnetic assembly process are experimentally demonstrated, respectively. Because of magnetic field, the complex assembly shapes can be fabricated just by moving the microfluidic device in a plane. To match well with the shape of supporting model and to keep a stable assembly structure, the secondary cross-linking method is employed. From the LIVE/DEAD assay, cells can survive well during the magnetic assembly process.