IROS 2021 Papers

Autonomous train navigation using only a low-cost MEMS IMU and a track map is considered in this paper. The approach is designed for urban rail or subway environments where GNSS measurements are unreliable or unavailable, and is intended as a baseline against which more complex sensor fusion approaches can be compared to ensure the consistency of the estimates. The estimator exploits the track motion constraint and information about position and velocity derived from centripetal acceleration and angular velocity measurements to improve the dead-reckoning solution and keep error and uncertainty bounded. In experimental validation over a 6 km run of a subway train during commuter service, the proposed approach had a maximum error of 6.0 m, validating the approach as an independent estimator.

We aim to assess the performance of LiDAR-to-map registration on compressive maps. Modern autonomous vehicles utilize pre-built HD (High-Definition) maps to perform sensor-to-map registration, which recovers pose estimation failures and reduces drift in a large-scale environment. However, sensor-to-map registration is usually realized by registering the sensor to a dense 3D model, which occupies massive storage space in the HD map and requires much data processing overhead. Although smaller 3D models are preferable, the optimal compressive map format for preservation of the best registration performance remains unclear.In this paper, we propose a novel and challenging benchmark to evaluate existing LiDAR-to-map registration methods from three perspectives: map compressibility, robustness, and precision. We compared various map formats, including raw points, hierarchical GMMs, and feature points, and show their performance trade-offs between compressibility and robustness on real-world LiDAR datasets: KITTI Odometry Dataset and Argoverse Tracking Dataset. Our benchmark reveals that state-of-the-art deep feature point based methods outperform traditional methods significantly when the map size budget is high. However, when map size budget is low, deep methods are outperformed by the methods using simpler models in Argoverse Tracking Dataset due to poor spatial coverage. In addition, we observe that the recently published TEASER++ significantly outperforms RANSAC for the feature point methods. Our analysis provides a valuable reference for the community to design budgeted real-world systems and find potential research opportunities. We will release the benchmark for public use.

In the paper, we show how scalable, low-cost trunk-like robotic arms can be constructed using only basic 3D-printing equipment and simple electronics. The design is based on uniform, stackable joint modules with three degrees of freedom each. Moreover, we present an approach for controlling these robots with recurrent spiking neural networks. At first, a spiking forward model learns motor-pose correlations from movement observations. After training, intentions can be projected back through unrolled spike trains of the forward model essentially routing the intention-driven motor gradients towards the respective joints, which unfolds goal-direction navigation. We demonstrate that spiking neural networks can thus effectively control trunk-like robotic arms with up to 75 articulated degrees of freedom with near millimeter accuracy.

A prominent challenge in the field of robotics is manipulation of flexible objects. One major factor that makes this task difficult is the complex dynamics emerging from its high-dimensional structure. This argues against the use of popular optimization-based approaches, which scale poorly with system dimension (the "curse of dimensionality"). Nevertheless, almost indifferent to this complexity, humans handle it on a daily basis, without any apparent difficulty.Inspired by human motor control, we propose that encoding movements based on dynamic primitives can simplify the task of manipulating flexible objects and provides a way around the curse of dimensionality. Using an extreme example — manipulating a whip — we tested in simulation whether targets at various locations could be reached with a whip by using a controller based on dynamic primitives. Regardless of the target location, this approach successfully managed the complexity of a 54 degree-of-freedom system (yielding a 108-dimensional state-space representation) and identified an upper-limb movement that achieved the task. This approach did not require a detailed model of the whip, which thereby significantly simplified the computational complexity of the control task. We believe that this approach may facilitate robotic manipulation of flexible materials, and in general afford a simplified way to control dynamically complex objects.

A major challenge in electroencephalogram (EEG) signal classification is that the EEG signals recorded from different subjects are drawn from different distributions. When the unlabeled EEG data of the new subject arrive, called target domain, classifying them with a classifier trained on prerecorded EEG data of other subjects, called source domain, will greatly decrease the classification accuracy. Being able to use the classifiers trained on data of source domain to accurately classify the data of target domain could reduce the time of the calibration phase in the actual application of the brain-computer interface. This study considers an offline cross-subject classification scenario. We propose a novel manifold trial selection method, which reduces the distribution distance between the source and target domains by manifold transformation and domain adaptation. The proposed method provides a trial selection strategy to suppress negative transfer by removing some abnormal samples. The proposed method is applied to the motor imagery-based brain–computer interface and compared with several existing algorithms. Experimental results show that the proposed method outperforms the state-of-the-art methods.

The ability to predict the future movements of other vehicles is a subconscious and effortless skill for humans and key to safe autonomous driving. Therefore, trajectory prediction for autonomous cars has gained a lot of attention in recent years. It is, however, still a hard task to achieve human-level performance. Interdependencies between vehicle behaviors and the multimodal nature of future intentions in a dynamic and complex driving environment render trajectory prediction a challenging problem. In this work, we propose a new, datadriven approach for predicting the motion of vehicles in a road environment. The model allows for inferring future intentions from the past interaction among vehicles in highway driving scenarios. Using our neighborhood-based data representation, the proposed system jointly exploits correlations in the spatial and temporal domain using convolutional neural networks. Our system considers multiple possible maneuver intentions and their corresponding motion and predicts the trajectory for five seconds into the future. We implemented our approach and evaluated it on two highway datasets taken in different countries and are able to achieve a competitive prediction performance.

The paper proposes a method for automatic multi-view dataset construction based on formula-driven supervised learning (FDSL). Although data collection and human annotation of 3D objects are labor-intensive, we automatically generate their training data and labels in the proposed multi-view dataset. To create a large-scale multi-view dataset, we employ fractal geometry, which is considered the background information of many objects in the real world. We project in a circle from the rendered 3D fractal models to construct the Multi-view Fractal DataBase (MV-FractalDB), which is then used to make a pre-trained CNN model. According to the experimental results, the MV-FractalDB pre-trained model surpasses the accuracies with self-supervised methods (e.g., SimCLR and MoCo) and is close to supervised methods (e.g., ImageNet) in terms of performance rates on multi-view image datasets. We demonstrate the potential of FDSL for multi-view image recognition.

We present multi-robot iSAM2 (MR-iSAM2), an efficient incremental smoothing and mapping (iSAM) algorithm to solve multi-robot simultaneous localization and mapping (SLAM) inference problems. MR-iSAM2 is based on a novel data structure multi-root Bayes tree (MRBT), which packs multiple Bayes trees with the same undirected clique structure. In multi-robot scenarios, the MRBT enables new measurements from different robots to be updated in different root branches, while all updates are performed around the single root of the Bayes tree in the original iSAM2 algorithm. As a result, the MRBT better reveals the underlying sparsity and information flow in multi-robot SLAM inference problems than the Bayes tree. Based on this insight, we further develop MR-iSAM2 to incrementally update and maintain the sparsity structure of the MRBT and enable efficient information propagation among the roots for inter-robot inference. We analyze the properties of the MR-iSAM2 algorithm, and show with both synthetic and real world datasets that it significantly outperforms iSAM2 in efficiency when solving multi-robot SLAM problems.

Visual Odometry (VO) is used in many applications including robotics and autonomous systems. However, traditional approaches based on feature matching are computationally expensive and do not directly address failure cases, instead relying on heuristic methods to detect failure. In this work, we propose a deep learning-based VO model to efficiently estimate 6-DoF poses, as well as a confidence model for these estimates. We utilise a CNN - RNN hybrid model to learn feature representations from image sequences. We then employ a Mixture Density Network (MDN) which estimates camera motion as a mixture of Gaussians, based on the extracted spatio-temporal representations. Our model uses pose labels as a source of supervision, but derives uncertainties in an unsupervised manner. We evaluate the proposed model on the KITTI and nuScenes datasets and report extensive quantitative and qualitative results to analyse the performance of both pose and uncertainty estimation. Our experiments show that the proposed model exceeds state-of-the-art performance in addition to detecting failure cases using the predicted pose uncertainty.

Surgical robots are usually controlled using a priori models based on the robots’ geometric parameters, which are calibrated before the surgical procedure. One of the challenges in using robots in real surgical settings is that those parameters can change over time, consequently deteriorating control accuracy. In this context, our group has been investigating online calibration strategies without added sensors. In one step toward that goal, we have developed an algorithm to estimate the pose of the instruments’ shafts in endoscopic images. In this study, we build upon that earlier work and propose a new framework to more precisely estimate the pose of a rigid surgical instrument. Our strategy is based on a novel pose estimation model called MBAPose and the use of synthetic training data. Our experiments demonstrated an improvement of 21 % for translation error and 26 % for orientation error on synthetic test data with respect to our previous work. Results with real test data provide a baseline for further research.

Multi-robot systems can benefit from reinforcement learning (RL) algorithms that learn behaviours in a small number of trials, a property known as sample efficiency. This research thus investigates the use of learned world models to improve sample efficiency. We present a novel multi-agent model-based RL algorithm: Multi-Agent Model-Based Policy Optimization (MAMBPO), utilizing the Centralized Learning for Decentralized Execution (CLDE) framework. CLDE algorithms allow a group of agents to act in a fully decentralized manner after training. This is a desirable property for many systems comprising of multiple robots. MAMBPO uses a learned world model to improve sample efficiency compared to model-free Multi-Agent Soft Actor-Critic (MASAC). We demonstrate this on two simulated multi-robot tasks, where MAMBPO achieves a similar performance to MASAC, but requires far fewer samples to do so. Through this, we take an important step towards making real-life learning for multi-robot systems possible.

State estimation with sensors is essential for mobile robots. Due to different performance of sensors in different environments, how to fuse measurements of various sensors is a problem. In this paper, we propose a tightly coupled multi-sensor fusion framework, Lvio-Fusion, which fuses stereo camera, Lidar, IMU, and GPS based on the graph optimization. Especially for urban traffic scenes, we introduce a segmented global pose graph optimization with GPS and loop-closure, which can eliminate accumulated drifts. Additionally, we creatively use a actor-critic method in reinforcement learning to adaptively adjust sensors’ weight. After training, actor-critic agent can provide the system better and dynamic sensors’ weight. We evaluate the performance of our system on public datasets and compare it with other state-of-the-art methods, which shows that the proposed method achieves high estimation accuracy and robustness to various environments. And our implementations are open source and highly scalable.

Recently, needs for unmanned aerial vehicles (UAVs) that are attachable to the wall have been highlighted. As one of the ways to address the need, researches on various tilting multirotors that can increase maneuverability has been employed. Unfortunately, existing studies on the tilting multirotors require considerable amounts of prior information on the complex dynamic model. Meanwhile, reinforcement learning on quadrotors has been studied to mitigate this issue. Yet, these are only been applied to standard quadrotors, whose systems are less complex than those of tilting multirotors. In this paper, a novel reinforcement learning-based method is proposed to control a tilting multirotor on real-world applications, which is the first attempt to apply reinforcement learning to a tilting multirotor. To do so, we propose a novel reward function for a neural network model that takes power efficiency into account. The model is initially trained over a simulated environment and then fine-tuned using real-world data in order to overcome the sim-to-real gap issue. Furthermore, a novel, efficient state representation with respect to the goal frame that helps the network learn optimal policy better is proposed. As verified on real-world experiments, our proposed method shows robust controllability by overcoming the complex dynamics of tilting multirotors.

Reinforcement learning algorithms have proven to be capable of solving complicated robotics tasks in an end-to-end fashion without any need for hand-crafted features or policies. Especially in the context of robotics, in which the cost of real-world data is usually extremely high, Reinforcement Learning solutions achieving high sample efficiency are needed. In this paper, we propose a framework combining the learning of a low-dimensional state representation, from high-dimensional observations coming from the robot’s raw sensory readings, with the learning of the optimal policy, given the learned state representation. We evaluate our framework in the context of mobile robot navigation in the case of continuous state and action spaces. Moreover, we study the problem of transferring what learned in the simulated virtual environment to the real robot without further retraining using real-world data in the presence of visual and depth distractors, such as lighting changes and moving obstacles. A video of our experiments can be found at: https://youtu.be/rUdGPKr2Wuo.

Multi-agent path finding (MAPF) determines an ensemble of collision-free paths for multiple agents between their respective start and goal locations. Among the available MAPF planners for workspace modeled as a graph, A*-based approaches have been widely investigated due to their guarantees on completeness and solution optimality, and have demonstrated their efficiency in many scenarios. However, almost all of these A*-based methods assume that each agent executes an action concurrently in that all agents start and stop together. This article presents a natural generalization of MAPF with asynchronous actions (MAPF-AA) where agents do not necessarily start and stop concurrently. The main contribution of the work is a proposed approach called Loosely Synchronized Search (LSS) that extends A*-based MAPF planners to handle asynchronous actions. We show LSS is complete and finds an optimal solution if one exists. We also combine LSS with other existing MAPF methods that aims to trade-off optimality for computational efficiency. Numerical results are presented to corroborate the performance of LSS and the applicability of the proposed method is verified in the Robotarium, a remotely accessible swarm robotics research platform.

Vision-based 3D object detection is a research focus in the field of autonomous driving system. While recently proposed pseudo-LiDAR is a promising solution, its performance is severely restricted by the image-based depth estimator, leading to a considerable performance gap against the LiDAR-based counterparts. In this paper, substantial advances are developed along an orthogonal direction to the previous efforts in the pseudo-LiDAR pipeline. Concretely, we propose a plug- and-play module, called Online Semantic Embedding (OSE), aligning image semantics with the pseudo-LiDAR detection in an end-to-end manner. On the KITTI object detection benchmark, existing stereo-based baselines integrated with our approach show impressive improvements without bells and whistles. Furthermore, we emphasize that OSE works in retrieving the performance under geometric imperfection conditions.

A coordinate system is proposed that replaces the usual three-dimensional Cartesian x, y, z position coordinates, for use in robotic localization applications. Range, azimuth, and elevation measurement models become greatly simplified, and, unlike spherical coordinates, the proposed coordinates do not suffer from the same kinematic singularities and angle wraparound. When compared to Cartesian coordinates, the proposed coordinate system results in a significantly enhanced ability to represent the true distribution of robot positions, ultimately leading to large improvements in state estimation consistency.

Real-time visual localization of needles is necessary for various surgical applications, including surgical automation and visual feedback. In this study we investigate localization and autonomous robotic control of needles in the context of our magneto-suturing system. Our system holds the potential for surgical manipulation with the benefit of minimal invasiveness and reduced patient side effects. However, the nonlinear magnetic fields produce unintuitive forces and demand delicate position-based control that exceeds the capabilities of direct human manipulation. This makes automatic needle localization a necessity. Our localization method combines neural network-based segmentation and classical techniques, and we are able to consistently locate our needle with 0.73mm RMS error in clean environments and 2.72mm RMS error in challenging environments with blood and occlusion. The average localization RMS error is 2.16 mm for all environments we used in the experiments. We combine this localization method with our closed-loop feedback control system to demonstrate the further applicability of localization to autonomous control. Our needle is able to follow a running suture path in (1) no blood, no tissue; (2) heavy blood, no tissue; (3) no blood, with tissue; and (4) heavy blood, with tissue environments. The tip position tracking error ranges from 2.6mm to 3.7mm RMS, opening the door towards autonomous suturing tasks.

In this paper, we propose a novel local to global plane regularity aggregation framework for dense surfel mapping, aiming for real-time reconstruction of high-quality 3D global models in both indoor and urban environments. Different from prior works that directly localize surfels globally, we investigate three interplanar geometric relations: {coplanarity, parallelism, orthogonality} from local to global scales as additional structural regularities in reconstruction, promoting the performance in plane-dominated scenes remarkably. Given a monocular RGB-D video as input, our framework extracts and utilizes the interplanar relations in three stages: local surfel creation, local to global relation propagation, and global plane-guided re-localization. In the first stage, surfels are created and refined within the current frame by aggregating temporal and spatial cues. The interplanar relations are adopted to regulate the normal and position of each surfel. Then in the second stage, we simultaneously establish correspondences between the created surfels and global model and propagate the interplanar relations from local to global. Finally, the positions of surfels are further relocated and optimized in a larger scale, based on the global interplanar relation priors aggregated across all local frames. Extensive experiments on datasets of different scales demonstrate that our framework achieves superior performance in terms of consistency and accuracy of the reconstructed global model. Meanwhile, the capability of our framework in the real-time 3D reconstruction on CPU opens the door to practical application.