IROS 2020 论文列表

Many indoor spaces have constantly changing layouts and may not be mapped by an autonomous vehicle, yet maps such as floor plans or evacuation maps of these places are common. We propose a method for an autonomous robot to localize itself on such maps with inconsistent scale using Stochastic Gradient Descent (SGD) with scan matching using a 2D LiDAR. We also introduce a new scale state in 2D localization to manage the possible inconsistent scale of the input map. Experiments are conducted in an indoor corridor using three different input maps - a point cloud, a floor plan, and a hand-drawn map. The SGD localization algorithm is bench-marked to Adaptive Monte Carlo Localization (AMCL). In a point cloud mapped environment, our algorithm achieves 0.264m and 5.26° average position and heading error respectively. On the hand-drawn map, our SGD localization algorithm remains robust while AMCL fails. The role of the scale state in our SGD localization algorithm is demonstrated in poorly scaled maps.

Robotic exploration requires adaptively selecting navigation goals that result in the rapid discovery and mapping of an unknown world. In many real-world environments, subtle structural cues can provide insight about the unexplored world, which may be exploited by a decision maker to improve the speed of exploration. In sparse subterranean tunnel networks, these cues come in the form of topological features, such as loops or dead-ends, that are often common across similar environments. We propose a method for learning these topological features using techniques borrowed from topological image segmentation and image inpainting to learn from a database of worlds. These world predictions then inform a frontier-based exploration policy. Our simulated experiments with a set of real-world mine environments and a database of procedurally-generated artificial tunnel networks demonstrate a substantial increase in the rate of area explored compared to techniques that do not attempt to predict and exploit topological features of the unexplored world.

As AI becomes an integral part of our lives, the development of explainable AI, embodied in the decision-making process of an AI or robotic agent, becomes imperative. For a robotic teammate, the ability to generate explanations to justify its behavior is one of the key requirements of explainable agency. Prior work on explanation generation has been focused on supporting the rationale behind the robot's decision or behavior. These approaches, however, fail to consider the mental demand for understanding the received explanation. In other words, the human teammate is expected to understand an explanation no matter how much information is presented. In this work, we argue that explanations, especially those of a complex nature, should be made in an online fashion during the execution, which helps spread out the information to be explained and thus reduce the mental workload of humans in highly cognitive demanding tasks. However, a challenge here is that the different parts of an explanation may be dependent on each other, which must be taken into account when generating online explanations. To this end, a general formulation of online explanation generation is presented with three variations satisfying different "online" properties. The new explanation generation methods are based on a model reconciliation setting introduced in our prior work. We evaluated our methods both with human subjects in a simulated rover domain, using NASA Task Load Index (TLX), and synthetically with ten different problems across two standard IPC domains. Results strongly suggest that our methods generate explanations that are perceived as less cognitively demanding and much preferred over the baselines and are computationally efficient.

Wheeled-legged robots combine the efficiency of wheeled robots when driving on suitably flat surfaces and versatility of legged robots when stepping over or around obstacles. This paper introduces a planning and control framework to realise dynamic locomotion for wheeled biped robots. We propose the Cart-Linear Inverted Pendulum Model (Cart-LIPM) as a template model for the rolling motion and the under-actuated LIPM for contact changes while walking. The generated motion is then tracked by an inverse dynamic whole-body controller which coordinates all joints, including the wheels. The framework has a hierarchical structure and is implemented in a model predictive control (MPC) fashion. To validate the proposed approach for hybrid motion generation, two scenarios involving different types of obstacles are designed in simulation. To the best of our knowledge, this is the first time that such online dynamic hybrid locomotion has been demonstrated on wheeled biped robots.

Recent advancements in Bayesian likelihood-free inference enables a probabilistic treatment for the problem of estimating simulation parameters and their uncertainty given sequences of observations. Domain randomization can be performed much more effectively when a posterior distribution provides the correct uncertainty over parameters in a simulated environment. In this paper, we study the integration of simulation parameter inference with both model-free reinforcement learning and model-based control in a novel sequential algorithm that alternates between learning a better estimation of parameters and improving the controller. This approach exploits the interdependence between the two problems to generate computational efficiencies and improved reliability when a black-box simulator is available. Experimental results suggest that both control strategies have better performance when compared to traditional domain randomization methods.

We consider the task of underwater robot navigation for the purpose of collecting scientifically relevant video data for environmental monitoring. The majority of field robots that currently perform monitoring tasks in unstructured natural environments navigate via path-tracking a pre-specified sequence of waypoints. Although this navigation method is often necessary, it is limiting because the robot does not have a model of what the scientist deems to be relevant visual observations. Thus, the robot can neither visually search for particular types of objects, nor focus its attention on parts of the scene that might be more relevant than the pre-specified waypoints and viewpoints. In this paper we propose a method that enables informed visual navigation via a learned visual similarity operator that guides the robot's visual search towards parts of the scene that look like an exemplar image, which is given by the user as a high-level specification for data collection. We propose and evaluate a weakly supervised video representation learning method that outperforms ImageNet embeddings for similarity tasks in the underwater domain. We also demonstrate the deployment of this similarity operator during informed visual navigation in collaborative environmental monitoring scenarios, in large-scale field trials, where the robot and a human scientist collaboratively search for relevant visual content. Code: https://github.com/rvl-lab-utoronto/visual_search_in_the_wild.

This paper presents a proof-of-concept for a localization and mapping system for magnetic crawlers performing inspection tasks on structures made of large metal plates. By relying on ultrasonic guided waves reflected from the plate edges, we show that it is possible to recover the plate geometry and robot trajectory to a precision comparable to the signal wavelength. The approach is tested using real acoustic signals acquired on metal plates using lawn-mower paths and random-walks. To the contrary of related works, this paper focuses on the practical details of the localization and mapping algorithm.

Under the excitation of acoustic radiation, the amount of energy absorbed and rebounded by cells have the relationship with mechanical properties, e.g. stiffness, shape, weight and so on. In this paper, a femtosecond laser-activated micro-detector is designed to convert this relationship into an electrical signal. First, the acoustic radiation is generated by a femtosecond laser pulse in a microchannel and acts on neighbor cells / beads. Then, an ultra-thin glass sheet (UTGS)-based pressure sensor (cantilever) is fabricated at the bottom of the microfluidic chip to monitor changes in acoustic pressure during detection process. In this detection system, the pressure sensor is fabricated with a 10 μm UTGS in a shape of rectangular cantilever and functions like a detector to convert acoustic waves into shift response. Based on the amplitude of detected pulses, we can directly analyze the acoustic energy, coming from either femtosecond laser pulse or that remains after penetrating target cells. We have taken experiments on 10 μm beads and verified the applicability of this micro-detector, and the proposed method has great potential to be applied in label-free cell manipulation (i.e., sorting) as a detection mechanism.

New challenges arise when investigating the use of active prostheses for lower limb replacement, such as high motor power requirements, leading to increased weight and reduced autonomy. Series and parallel elasticity are often explored to reduce the necessary motor power but often the effect on the energy consumption of the prosthesis is not directly investigated, as the mechanical power properties are examined yet the motor and gearbox dynamics and efficiencies are not considered. This paper presents the investigation of a parallel elasticity compared to a series elastic actuation system used in an active ankle prosthesis. Using a matched electromechanical model of the actuator shows that the electrical efficiency can be influenced using parallel elasticity. The optimal configuration depends on the motor characteristics (dynamic behavior) and limitations, which should always be taken into account when designing optimal series and parallel springs. It has been shown that adding parallel elasticity allows to reduce the required gear ratio and thus associated friction and inertial losses. Allowing the parallel elasticity to be lockable can further influence the behavior and allow for a more versatile actuator.

In this paper, we propose a camera control system towards occasionally videoing preassigned objects. Based on the technique of real-time visual detection and tracking, using the Kalman filter and re-identification (ReID), we propose continuous composition of lens, based on the atomic rules of shots, and give the trajectory planning of the camera, to generate the PID controller to the pan-tilt. By both simulation and emulation by frame-wise cropping of video clips, we illustrate the efficiency of this method. Based on this model, we design and produce an AI automatic camera for lively photography and clip videoing.

Robot motion is controlled in the joint space whereas the robots have to perform tasks in their task space. Many tasks like carrying a glass of liquid, pouring liquid, opening a drawer requires constraints on the end-effector during the motion. The forward and inverse kinematic mappings between joint space and task space are highly nonlinear and multi-valued (for IK). Consequently, modeling task space constraints like keeping the orientation of the end-effector fixed while changing its position (which is required for carrying a cup of liquid without dropping it) is quite complex in the joint space. In this paper, we show that the use of screw linear interpolation to plan motions in the task space combined with resolved motion rate control to compute the corresponding joint space path, allows one to satisfy many common task space motion constraints in motion planning, without explicitly modeling them. In particular, any motion constraint that forms a subgroup of the group of rigid body motions can be incorporated in our planning scheme, without explicit modeling. We present simulation and experimental results on Baxter robot for different tasks with task space constraints that demonstrates the usefulness of our approach.

Space manipulator systems in orbit are subject to link flexibilities since they are designed to be lightweight and long reaching. Often, their joints are driven by harmonic gear-motor units, which introduce joint flexibility. Both of these types of flexibility may cause structural vibrations. To improve endpoint tracking, advanced control strategies that benefit from the knowledge of system parameters, including those describing link and joint flexibilities, are required. In this paper, first, the equations of motion of space manipulator systems whose manipulators are subject to both link and joint flexibilities are derived. Then, a parameter estimation method is developed, based on the energy balance during the motion of a flexible space manipulator. The method estimates all system parameters including those that describe both link and joint flexibilities and can reconstruct the system full dynamics required for the application of advanced control strategies. The method, developed for spatial systems, is illustrated by a planar example.

Towards long range and high endurance sailing, energy is of utmost importance. Moreover, benefiting from the dominance of the sailboat itself, it is energy-saving and environment-friendly. Thus, the sailboat with energy planning problem is meaningful. However, until now, the sailboat energy optimization problem has rarely been considered. In this paper, we focus on the energy consumption optimization of an autonomous sailboat. It has been formulated as a Nonlinear Programming problem (NLP). We deal with it with a hybrid control scheme, in which pseudo-spectral (PS) optimal control method is used in heading control, and a model-free framework guided by Extreme Seeking Control (ESC) is used in sail control. The optimal path is generated with the optimal input motor torques in time series. As a result, both simulation and experiments have validated motion planning and energy planning performance. Notably, about 7% of energy is saved on average. Our proposed method can make sailboats sailing longer and sustainable.

Visual motion estimation is an integral and well-studied challenge in autonomous navigation. Recent work has focused on addressing multimotion estimation, which is especially challenging in highly dynamic environments. Such environments not only comprise multiple, complex motions but also tend to exhibit significant occlusion.Previous work in object tracking focuses on maintaining the integrity of object tracks but usually relies on specific appearance-based descriptors or constrained motion models. These approaches are very effective in specific applications but do not generalize to the full multimotion estimation problem.This paper presents a pipeline for estimating multiple motions, including the camera egomotion, in the presence of occlusions. This approach uses an expressive motion prior to estimate the SE(3) trajectory of every motion in the scene, even during temporary occlusions, and identify the reappearance of motions through motion closure. The performance of this occlusion-robust multimotion visual odometry (MVO) pipeline is evaluated on real-world data and the Oxford Multimotion Dataset.

Industrial robots contain minimal sensing capability beyond recognition of their internal state. It is critical that an external vision system should cover the designated robot work space with awareness of blind spots and occlusions. This work presents two mechanisms to handle occlusions in an external multi-robot vision system: occlusion-aware optimal sensor positioning, and event-driven occlusion detection. When deploying sensors to the system, various scenarios are considered during optimization to reduce potential occlusions and increase sensor coverage. These methods are tested on a working cell with three industrial robot arms. The experimental results demonstrate the effectiveness of the proposed scenario-based multi-objective optimization for sensor positioning. Once the sensors are deployed, occlusion detection is actively triggered prior to robot path planning.

In the last decade, soft robots have been at the forefront of a robotic revolution. Due to the flexibility of the soft materials employed, soft robots are equipped with a capability to execute new tasks in new application areas -beyond what can be achieved using classical rigid-link robots. Despite these promising properties, many soft robots nowadays lack the capability to exert sufficient force to perform various real-life tasks. This has led to the development of stiffness-controllable inflatable robots instilled with the ability to modify their stiffness during motion. This new capability, however, poses an even greater challenge for robot control. In this paper, we propose a model-based kinematic control strategy to guide the tip of an inflatable robot arm in its environment. The bending of the robot is modelled using an Euler-Bernoulli beam theory which takes into account the variation of the robot's structural stiffness. The parameters of the model are estimated online using an observer based on the Extended Kalman Filter (EKF). The parameters' estimates are used to approximate the Jacobian matrix online and used to control the robot's tip considering also variations in the robot's stiffness. Simulation results and experiments using a fabric-based planar 3-degree-of-freedom (DOF) inflatable manipulators demonstrate the promising performance of the proposed control algorithm.

Collaborative robotics allows merging the best capabilities of humans and robots to perform complex tasks. This allows the user to interact with remote and directly inaccessible environments such as the micro-scale world. This interaction is made possible by the bidirectional exchange of information (displacement - force) between the user and the environment through a haptic interface. The effectiveness of the human/robot interaction is highly dependent on how the human feels the forces. This is a key point to enable humans to make the right decisions in a collaborative task. This paper discusses the design of a dynamic observer to estimate the forces applied by a human operator on a class of parallel pantograph-type haptic interfaces used to control small-scale robotic systems. The objective is to reject disturbances in order to improve the human force perception capability over a wide frequency range. A dynamic pantograph model is proposed and experimentally validated. The observer is designed on the basis of the proposed dynamic model and its efficiency in estimating the applied human force is demonstrated for the first time with pantograph-type interfaces. Experimental validation first shows the effectiveness of the perturbation observer for external human force estimation with a response time of less than 0.2 s and a mean error of less than 7 mN and then the effectiveness of the controller in improving the quality of human sensation of forces down to 10 mN.

In this paper, we propose objective functions to estimate the principal contact between a unknown manipulated target object and its unknown surroundings from the motion of the object. We derived the objective functions based on the fact that contact involves a pair of geometrical primitives (a point of vertex, a line of edge, and a plane of face) for the singular condition of the calculation for their intersection or their spanned space from the point of view of geometrical algebra. The minimization of the proposed objective functions, which are differential quadratic forms of the Kronecker product of geometrical parameters, efficiently provided us the contact geometries that constrained the object movements. Additionally, the proposed objective functions are fundamentals for identifying contact during compliant manipulation, and we showed that the objective functions provide a clue for contact identification via experiments.