IROS 2020 论文列表

In this paper, we investigate the perception-aware path finding, planning and following for a class of snake robots autonomously serpentining in an unmodeled and unknown environment. In the work, the onboard LiDAR sensor mounted on the head of the snake robot is utilized to reconstruct the local environment, by which and the modified rapidly-exploring random tree method, a feasible path from the current position of the robot to a local selected target position can be obtained. Next, the parametric cubic spline interpolation path-planning method and potential functions are applied to make the path more smooth so as to prevent the multi-link and elongated robot body from hitting obstacles. To steer, a time-varying line-of-sight control law is designed to ensure that the robot moves to the local target position along the generated path by the perception-aware method. The robot will repeatedly perform the above search-find-move strategy until it reaches the final predefined target point. Simulation and experimental results demonstrate a good performance of the proposed perception-aware approach, that is, the elongated and underactuated snake robot is capable of autonomously navigating in an unknown environment.

This paper presents a method to cope with autonomous assembly tasks in the presence of uncertainties. To this aim, a Peg-in-Hole operation is considered, where the target workpiece position is unknown and the peg-hole clearance is small. Deep learning based hole detection and 3D surface reconstruction techniques are combined for accurate workpiece localization. In detail, the hole is detected by using a convolutional neural network (CNN), while the target workpiece surface is reconstructed via 3D-Digital Image Correlation (3D-DIC). Peg insertion is performed via admittance control that confers the suitable compliance to the peg. Experiments on a collaborative manipulator confirm that the proposed approach can be promising for achieving a better degree of autonomy for a class of robotic tasks in partially structured environments.

Inertial Measurement Unit (IMU) has long been a dream for stable and reliable motion estimation, especially in indoor environments where GPS strength limits. In this paper, we propose a novel method for position and orientation estimation of a moving object only from a sequence of IMU signals collected from the phone. Our main observation is that human motion is monotonous and periodic. We adopt the Extended Kalman Filter and use the learning-based method to dynamically update the measurement noise of the filter. Our pedestrian motion tracking system intends to accurately estimate planar position, velocity, heading direction without restricting the phone's daily use. The method is not only tested on the self-collected signals, but also provides accurate position and velocity estimations on the public RIDI dataset, i.e., the absolute transmit error is 1.28m for a 59-second sequence.

This paper proposes a multiple stakeholder perspective model (MSPM) which predicts the future pedestrian trajectory observed from vehicle's point of view. For the vehicle-pedestrian interaction, the estimation of the pedestrian's intention is a key factor. However, even if this interaction is commonly initiated by both the human (pedestrian) and the agent (driver), current research focuses on developing a neural network trained by the data from driver's perspective only. In this paper, we suggest a multiple stakeholder perspective model (MSPM) and apply this model for pedestrian intention prediction. The model combines the driver (stakeholder 1) and pedestrian (stakeholder 2) by separating the information based on the perspective. The dataset from pedestrian's perspective have been collected from the virtual reality experiment, and a network that can reflect perspectives of both pedestrian and driver is proposed. Our model achieves the best performance in the existing pedestrian intention dataset, while reducing the trajectory prediction error by average of 4.48% in the short-term (0.5s) and middle-term (1.0s) prediction, and 11.14% in the long-term prediction (1.5s) compared to the previous state-of-the-art.

Many surgical robots with steerable surgical instruments have been proposed for endoscopic surgery. Surgical instruments should be small in size for insertion into the body and be able to handle large payloads such as tissue. Because the overall diameter and payload parameters are a trade-off, it is difficult to design an instrument with a large payload while reducing its diameter. In this paper, we optimize the payload of a rolling joint mechanism by deriving the moment equilibrium equation and constraints for endoscopic surgery. A scaled-up prototype was fabricated with the design variables obtained from the optimization, and the validity of the method for calculating the payload was confirmed by the experimentally measured payload. By plotting the distribution of payloads obtained from the moment equilibrium equation, we also confirmed that the payload obtained from the optimization is the maximum. In addition, optimizations with different numbers of joints confirm that the payload tends to decrease as the number of joints increases. This payload optimization method could also be extended to minimizing the deflection of the bending section against external forces and minimizing the diameter of the surgical instrument given the minimum required payload.

A robot in the future may initially has a good learning capability but an empty library of movements. It gradually enriches its library of movements through human demonstrations. Dynamic Movement Primitives (DMPs) has been proved to be an effective way to represent trajectories. Trajectories are classified into discrete and rhythmic ones, and parameters are set for each demonstrated trajectory. However, what kind of trajectory will be provided by robot users is sometimes unknown to robot developers, so trajectory pattern and the parameters can not be determined in advance. It's also impossible for non-technical robot users to set these parameters and determine the pattern of movements they are going to demonstrate. To make it easier for non-expert robot users to programme their robots by demonstration, this work presents an efficient way to deal with these two problems. The effectiveness of the proposed methodology is proved by teaching a robot to clean the whiteboard in different ways and stack a set of cubic boxes in specific order.

This paper presents a path planner, which enables a nonholonomic mobile manipulator to move its end-effector on an observed surface with a constrained orientation, given start and destination points. A partial point cloud of the environment is captured using a vision-based sensor, but no prior knowledge of the surface shape is assumed. We consider the multi-objective optimisation problem of finding robot paths which account for the nonholonomic constraints of the base, maximise the robot's manipulability throughout the motion, while also minimising surface-distance travelled between the two points. This work has application in industrial problems of rough robotic cutting, e.g. demolition of legacy nuclear plants, where dismantling does not require a precise path. We show how our approach embeds the nonholonomic constraints of the mobile platform into an extended Jacobian, while additionally encoding the constraint that the end-effector must remain in contact with the cut surface throughout the motion. We use this constrained Jacobian to plan a time-series of robot configurations. Additionally, we show how our novel cost function is suitable for combining with a variety of well-known path planners, such as RRT*. We present several empirical experiments in different scenarios, where a simulated non-holonomic mobile manipulator follows a trajectory, which is generated on noisy point clouds derived from real depth-camera images of real objects. Our planner (RRT*-CRMM) enables successful task completion by optimising the path over the travelled distance, the manipulability of the arm, and the movements of the mobile base.

In this paper, we propose a novel path planning algorithm for the nonholonomic multiple mobile robot system with applications to a robotic autonomous luggage trolley collection system at airports. We consider this path planning algorithm as a Multiple Traveling Salesman Problem (MTSP). Our path planning algorithm consists of three parts. First, we use the Minimum Spanning Tree (MSP) algorithm to divide the MTSP into a number of independent TSPs, which achieves the task assignment for each mobile robot. Secondly, we implement a closed-loop forward control policy based on the kinematic model of the mobile robot to get a feasible and smooth path. The control cost of the path is used as the new metric in solving the TSPs. Finally, in order to adapt to our case, we modify the TSP as an Open Dynamic Traveling Salesman Problem with Fixed Start (ODTSP-FS) and implement an ant colony algorithm to achieve the path planning for each mobile robot. We evaluate our algorithm with simulation experiments and the experimental results demonstrate that our algorithm can quickly generate feasible and smooth paths for each robot while satisfying the nonholonomic constraints.

Under line-of-sight (LOS) network conditions, multi-input multi-output (MIMO) wireless communications can increase the channel capacity between a team of robots and a multi-antenna array at a stationary base station. This increased capacity can result in greater data throughput, shortening the time necessary to complete channel-limited data aggregation tasks. To take advantage of this higher capacity channel, the robots in the team must be positioned to maximize complex channel orthogonality between each robot and receiver antenna. Using geometrically motivated assumptions, we derive transmitter spacing rules that can be easily be added on to existing path plans to improve backhaul throughput for data offloading from the robot team, with minimal impact on other system objectives. We demonstrate the effectiveness of the approach- both in ideal as well as realistic channels outside the domain of our simplifying assumptions-with numerical examples of robot-coordinated path plans in two example environments, achieving up to 42% improvement in task completion times.

This paper addresses the problem of path negotiation among self-interested multirobot operators in shared space. In conventional multirobot path planning problems, most of the research thus far has focused on the coordination and planning of collision-free paths for multiple robots with some common objectives. On the contrary, the recent progress of technologies in autonomous vehicles, including automated guidance vehicles, unmanned aerial vehicles, and manned autonomous cars, has increased demand for solving coordination and conflict avoidance in these autonomous and self-interested agents that pursue their own objectives. In this research, we tackle this problem from the operator perspective. We assume a problem setting where collisions between robots are avoided based on path reservation and negotiation. Under that circumstance, we propose a task-oriented utility function and a path negotiation algorithm for robot operators to maximize their task utility during path negotiation. The simulation and experiment results demonstrate the effectiveness of our task-based negotiation method over a simple path-based negotiation approach.

While impedance control is one of the most commonly used strategies for robot interaction control, variable impedance control is a more recent preoccupation. If designing impedance control with varying parameters allows increasing the system flexibility and dexterity, it is still a challenging issue, as it may result in a loss of passivity of the control system. This has an important impact on the stability and therefore on the safety of the interaction. In this paper, we propose methods to design passivity filters that guarantee passivity of the interaction. They aim at either checking whether a desired impedance profile is passive, or modifying it if required.

Robots working in human environments must be able to perceive and act on challenging objects with articulations, such as a pile of tools. Articulated objects increase the dimensionality of the pose estimation problem, and partial observations under clutter create additional challenges. To address this problem, we present a generative-discriminative parts-based recognition and localization method for articulated objects in clutter. We formulate the problem of articulated object pose estimation as a Markov Random Field (MRF). Hidden nodes in this MRF express the pose of the object parts, and edges express the articulation constraints between parts. Localization is performed within the MRF using an efficient belief propagation method. The method is informed by both part segmentation heatmaps over the observation, generated by a neural network, and the articulation constraints between object parts. Our generative-discriminative approach allows the proposed method to function in cluttered environments by inferring the pose of occluded parts using hypotheses from the visible parts. We demonstrate the efficacy of our methods in a tabletop environment for recognizing and localizing hand tools in uncluttered and cluttered configurations.

A novel identification method is developed which identifies the accumulated angular momentum (AAM) of spinning reaction wheels (RWs) of an uncooperative satellite captured by a robotic servicer. In contrast to other methods that treat captured satellite's RWs as non-spinning, the developed method provides simultaneously accurate estimates of the AAM of the captured satellite's RWs and of the inertial parameters of the entire system consisting of the robotic servicer and of the captured satellite. These estimates render the system free-floating dynamics fully identified and available to model-based control. Three-dimensional simulations demonstrate the method's validity. To show its usefulness, the performance of a model-based controller is evaluated with and without knowledge of the captured satellite's RWs AAM.

Directionality in path planning is essential for efficient autonomous navigation in a number of real-world environments. In many map-based navigation scenarios, the viable path from a given point A to point B is not the same as the viable path from B to A. We present a method that automatically incorporates preferred navigation directionality into a path planning costmap. This `preference' is represented by coloured paths in the costmap. The colourisation is obtained based on an analysis of the driving trajectory generated by the robot as it navigates through the environment. Hence, our method augments this driving trajectory by intelligently colouring it according to the orientation of the robot during the run. Creating an analogy between the vehicle orientation angle and the hue angle in the Hue-Saturation-Value colour space, the method uses the hue, saturation and value components to encode the direction, directionality and scalar cost, respectively, into a costmap image. We describe a costing function to be used by the A* algorithm to incorporate this information to plan direction-aware vehicle paths. Our experiments with LiDAR-based localisation and autonomous driving in real environments illustrate the applicability of the method.

Pursuit-Evasion Game (PEG) consists of a team of pursuers trying to capture one or more evaders. PEG is important due to its application in surveillance, search and rescue, disaster robotics, boundary defense and so on. In general, PEG requires exponential time to compute the minimum number of pursuers to capture an evader. To mitigate this, we have designed a parallel optimal algorithm to minimize the capture time in PEG. Given a discrete topology, this algorithm also outputs the minimum number of pursuers to capture an evader. A classic example of PEG is the popular arcade game, Pac-Man. Although Pac-Man topology has almost 300 nodes, our algorithm can handle this. We show that Pac-Man is overkill, i.e., given the Pac-Man game topology, Pac-Man game contains more pursuers/ghosts (four) than it is necessary (two) to capture evader/Pac-man. We evaluate the proposed algorithm on many different topologies.

Discrete graphs are commonly used to approximately represent configuration spaces used in robot motion planning. This paper explores a representation in which the costs of crossing local regions of the configuration space are represented using piecewise linear regression (PLR). We explore a few simple motion planning problems, and show that for these problems, the memory required to store the representation compares favorably to that required for standard discrete vertex-and-edge models, while preserving the quality of paths returned from searches.

Pose estimation of known objects is fundamental to tasks such as robotic grasping and manipulation. The need for reliable grasping imposes stringent accuracy requirements on pose estimation in cluttered, occluded scenes in dynamic environments. Modern methods employ large sets of training data to learn features in order to find correspondence between 3D models and observed data. However these methods require extensive annotation of ground truth poses. An alternative is to use algorithms that search for the best explanation of the observed scene in a space of possible rendered scenes. A recently developed algorithm, PERCH (PErception Via SeaRCH) does so by using depth data to converge to a globally optimum solution using a search over a specially constructed tree. While PERCH offers strong guarantees on accuracy, the current formulation suffers from low scalability owing to its high runtime. In addition, the sole reliance on depth data for pose estimation restricts the algorithm to scenes where no two objects have the same shape. In this work, we propose PERCH 2.0, a novel perception via search strategy that takes advantage of GPU acceleration and RGB data. We show that our approach can achieve a speedup of 100x over PERCH, as well as better accuracy than the state-of-the-art data-driven approaches on 6-DoF pose estimation without the need for annotating ground truth poses in the training data. Our code and video are available at https://sbpl-cruz.github.io/perception/.

The Next Best View (NBV) problem is important in the active robotic reconstruction. It enables the robot system to perform scanning actions in a reasonable view sequence, and fulfil the reconstruction task in an effective way. Previous works mainly follow the volumetric methods, which convert the point cloud information collected by sensors into a voxel representation space and evaluate candidate views through ray casting simulations to pick the NBV. However, the process of volumetric data transformation and ray casting is often time-consuming. To address this issue, in this paper, we propose a point cloud based deep neural network called PC-NBV to achieve efficient view planning without these computationally expensive operations. The PC-NBV network takes the raw point cloud data and current view selection states as input, and then directly predicts the information gain of all candidate views. By avoiding costly data transformation and ray casting, and utilizing powerful neural network to learn structure priors from point cloud, our method can achieve efficient and effective NBV planning. Experiments on multiple datasets show the proposed method outperforms state-of-the-art NBV methods, giving better views for robot system with much less inference time. Furthermore, we demonstrate the robustness of our method against noise and the ability to extend to multi-view system, making it more applicable for various scenarios.

Following considerable development in 3D scanning technologies, many studies have recently been proposed with various approaches for 3D vision tasks, including some methods that utilize 2D convolutional neural networks (CNNs). However, even though 2D CNNs have achieved high performance in many 2D vision tasks, existing works have not effectively applied them onto 3D vision tasks. In particular, segmentation has not been well studied because of the difficulty of dense prediction for each point, which requires rich feature representation. In this paper, we propose a simple and efficient architecture named point projection and back-projection network (PBP-Net), which leverages 2D CNNs for the 3D point cloud segmentation. 3 modules are introduced, each of which projects 3D point cloud onto 2D planes, extracts features using a 2D CNN backbone, and back-projects features onto the original 3D point cloud. To demonstrate effective 3D feature extraction using 2D CNN, we perform various experiments including comparison to recent methods. We analyze the proposed modules through ablation studies and perform experiments on object part segmentation (ShapeNet-Part dataset) and indoor scene semantic segmentation (S3DIS dataset). The experimental results show that proposed PBP-Net achieves comparable performance to existing state-of-the-art methods.