IROS 2021 论文列表

This study addresses the problem of path optimization for conducting a mapping mission using a multi-robot system with limited sensing capability, which aims to ensure efficient mapping with emphasis on the cooperative aspect of the mission. To achieve the cooperative mapping, a new path planning algorithm is proposed which can take advantage of the multi-robot system while dealing with the lack of observability due to the nature of bearing-only or range-only sensing. The Fisher information matrix is used to estimate the mapping uncertainty affected by the robots’ geometric configuration. Also, a simple method to predict the convergence rate of the uncertainty over a short time horizon is presented for efficient path planning. The performance of the proposed algorithm is shown through simulations and compared with other path planning algorithms.

When dealing with the haptic teleoperation of multi-limbed mobile manipulators, the problem of mitigating the destabilizing effects arising from the communication link between the haptic device and the remote robot has not been properly addressed. In this work, we propose a passive control architecture to haptically teleoperate a legged mobile manipulator, while remaining stable in the presence of time delays and frequency mismatches in the master and slave controllers. At the master side, a discrete-time energy modulation of the control input is proposed. At the slave side, passivity constraints are included in an optimization-based whole-body controller to satisfy the energy limitations. A hybrid teleoperation scheme allows the human operator to remotely operate the robot’s end- effector while in stance mode, and its base velocity in locomotion mode. The resulting control architecture is demonstrated on a quadrupedal robot with an artificial delay added to the network.

The DARPA subterranean challenge requires teams of robots to traverse difficult and diverse underground environments. Traversing small gaps is one of the challenging scenarios that robots encounter. Imperfect sensor information makes it difficult for classical navigation methods, where behaviours require significant manual fine tuning. In this paper we present a deep reinforcement learning method for autonomously navigating through small gaps, where contact between the robot and the gap may be required. We first learn a gap behaviour policy to get through small gaps (only centimeters wider than the robot). We then learn a goal-conditioned behaviour selection policy that determines when to activate the gap behaviour policy. We train our policies in simulation and demonstrate their effectiveness with a large tracked robot in simulation and on the real platform. In simulation experiments, our approach achieves 93% success rate when the gap behaviour is activated manually by an operator, and 63% with autonomous activation using the behaviour selection policy. In real robot experiments, our approach achieves a success rate of 73% with manual activation, and 40% with autonomous behaviour selection. While we show the feasibility of our approach in simulation, the difference in performance between simulated and real world scenarios highlight the difficulty of direct sim-to-real transfer for deep reinforcement learning policies. In both the simulated and real world environments alternative methods were unable to traverse the gap.

As robotic systems move from highly structured environments to open worlds, incorporating uncertainty from dynamics learning or state estimation into the control pipeline is essential for robust performance. In this paper we present a nonlinear particle model predictive control (PMPC) approach to control under uncertainty, which directly incorporates any particle-based uncertainty representation, such as those common in robotics. Our approach builds on scenario methods for MPC, but in contrast to existing approaches, which either constrain all or only the first timestep to share actions across scenarios, we investigate the impact of a partial consensus horizon. Implementing this optimization for nonlinear dynamics by leveraging sequential convex optimization, our approach yields an efficient framework that can be tuned to the particular information gain dynamics of a system to mitigate both over-conservatism and over-optimism. We investigate our approach for two robotic systems across three problem settings: time-varying, partially observed dynamics; sensing uncertainty; and model-based reinforcement learning, and show that our approach improves performance over baselines in all settings.

The fabrication method to utilize poly (vinyl alcohol) hydrogels with additional stiff parts in a single structure for hydrogel-based soft robots to realize an intensive motion with elastic energy is proposed in this paper. An inorganic material which is often seen in the hard tissues of our body; hydroxyapatite, was partially formed on a hydrogel with a simple procedure of alternatingly soaking a hydrogel with a covering device into two salt liquids. With this bioinspired structure, the whole hydrogel was able to change its physical properties while maintaining its whole structure as one. We succeeded in flicking motion of this structure only by bias displacement. This structure will help soft robots to move more intensively in a simple fabrication.

Data augmentation has greatly contributed to improving the performance in image recognition tasks, and a lot of related studies have been conducted. However, data augmentation on 3D point cloud data has not been much explored. 3D label has more sophisticated and rich structural information than the 2D label, so it enables more diverse and effective data augmentation. In this paper, we propose part-aware data augmentation (PA-AUG) that can better utilize rich information of 3D label to enhance the performance of 3D object detectors. PA-AUG divides objects into partitions and stochastically applies five augmentation methods to each local region. It is compatible with existing point cloud data augmentation methods and can be used universally regardless of the detector’s architecture. PA-AUG has improved the performance of state-of-the-art 3D object detector for all classes of the KITTI dataset and has the equivalent effect of increasing the train data by about 2.5×. We also show that PA-AUG not only increases performance for a given dataset but also is robust to corrupted data. The code is available at https://github.com/sky77764/pa-aug.pytorch

In this paper, we introduce a new type of compliant actuator named the Parallel Variable Stiffness Actuator (PVSA) which consists of a variable stiffness spring placed in parallel with a direct-drive motor. Parallel variable stiffness actuators provide (i) high-fidelity force control and (ii) controllable energy storage, as they inherit the benefits of direct-drive motors and variable stiffness springs. We present a compact design of the PVSA using a flat motor connected to an adjustable mechanical advantage torsional spring. We show that this PVSA is (1) not subject to the fundamental force control bandwidth limitation of series elastic and variable stiffness actuators, and most notably, (2) enables resonant energy accumulation despite the limited deformation of the spring and the constrained motion of the load attached to the actuator. The latter differentiates parallel variable stiffness actuators from fixed-stiffness parallel elastic actuators. PVSAs may be used with smaller direct-drive motors to match the peak power of larger motors without compromising force control fidelity. PVSAs may be used to implement resonant forcing under joint angle limitations in walking, jumping, running, swimming robots, or robotic exoskeletons used to augmented human motion in the aforementioned tasks.

Automated guided vehicles operation in human populated factory environments is a challenging task, especially when there is a demand to operate without following fixed paths defined by guide wires, magnetic tape, magnets, or transponders embedded in the floor. The paper at hand introduces a vision-based method enabling safe and autonomous operation of pallet moving vehicles that accommodate pallet detection, pose estimation, docking control and pallet pick up in such industrial environments. A dedicated perception topology relying on monocular vision and laser-based measurements has been applied and installed on-board a novel robotic pallet truck. Pallet detection and pose estimation are performed in two steps. Firstly, a deep neural network is used for the fast isolation of pallets' regions of interest and, secondly, model-based geometrical pattern matching on point cloud data is applied to extract the pallet pose. Robot alignment with candidate pallet is performed with a dedicated visual servoing controller. The developed method has been extensively evaluated both in simulated and real industrial environments with the pallet truck and proved to have real-time performance achieving increased accuracy in navigation, pallet detection and pick-up.

Production plants are being re-designed to implement human-centered solutions. Especially considering high added-value operations, robots are required to optimize their behavior to achieve a task quality at least comparable to the one obtained by the skilled operators. A manual programming and tuning of the manipulator is not an efficient solution, requiring to adopt towards automated strategies. Adding external sensors (e.g., cameras) increases the robotic cell complexity and it doesn’t solve the issue since it is usually difficult to build explicit reward functions measuring the robot performance, while it is easier for the user to define a qualitative comparison between two experiments. According to these needs, in this paper, the recently-developed preferences-based optimization approach GLISp is employed and adapted to tune the novel developed path-based velocity planner. The implemented solution defines an intuitive human-centered procedure, capable of transferring (through pairwise preferences between experiments) the task knowledge from the operator to the manipulator. A Franka EMIKA panda robot has been employed as a test platform to perform a robotic sealing task (i.e., material deposition task), validating the proposed methodology. The proposed approach has been compared with a programming by demonstration approach, and with the manual tuning of the path-based velocity planner. Achieved results demonstrate the improved deposition quality obtained with the proposed optimized path-based velocity planner methodology in a limited number of experimental trials (20).

A human patient simulator (HPS) can achieve effective visual-, auditory-, text-, and alarm-based feedback methods in care or nursing education. Among these, the method of visual feedback is important to design an HPS that can express emotions or feelings of pain like an actual human does because this method allows an immediate reaction between robots and humans. This study aims to develop an avatar-based visual feedback method for a care training assistant robot that can express pain states in joint care education. First, this study introduces its own pain facial expression database from Ritsumeikan University (RU-PITENS) for an avatar with pain expression. The RU-PITENS database contains pain images of 41 Japanese people in their 20s, 30s, 40s, and 60s, and an experiment of pain stimulus is conducted based on transcutaneous electrical nerve stimulation, which is low-cost and easy to use in daily life. Based on the pain images in the RU-PITENS database, we generated an avatar with pain expression to achieve the goal of our study. Since the RUPITENS database does not contain the quantitative pain level, the Siamese network was used to calculate the pain intensity. In addition, the care training assistant robot (CaTARo) developed in our previous study reproduces symptoms of musculoskeletal diseases, and the pain of CaTARo was measured using fuzzy logic theory. As a result, a visual feedback system was constructed to express five types of pain (no pain at all, very faint, weak, moderate, and strong pain) with avatars according to the intensity of the pain output of CaTARo in care training environments.

Product packing is a typical application in ware-house automation that aims to pick objects from unstructured piles and place them into bins with optimized placing policy. However, it still remains a significant challenge to finish the product packing tasks in general logistics scenarios where the objects are variable-sized and the configurations are complex. In this work, we present the PackerBot, a complete robotic pipeline for performing variable-sized product packing in unstructured scenes. First, by leveraging the imperfect experience of human packer, we propose a heuristic DRL framework for learning optimal online 3D bin packing policy. Then we integrate it with a 6-DoF suction-based picking module and a product size estimation module, leading to a complete product packing system, namely the PackerBot. Extensive experimental results show that our method achieves the state-of-the-art performance in both simulated and real-world tests. The video demonstration is available at: https://vsislab.github.io/packerbot.

3D single object tracking is a key issue for robotics. In this paper, we propose a transformer module called Point-Track-Transformer (PTT) for point cloud-based 3D single object tracking. PTT module contains three blocks for feature embedding, position encoding, and self-attention feature computation. Feature embedding aims to place features closer in the embedding space if they have similar semantic information. Position encoding is used to encode coordinates of point clouds into high dimension distinguishable features. Self-attention generates refined attention features by computing attention weights. Besides, we embed the PTT module into the open-source state-of-the-art method P2B to construct PTT-Net. Experiments on the KITTI dataset reveal that our PTT-Net surpasses the state-of-the-art by a noticeable margin \left( {\sim 10\% } \right)\left( {\sim 10\% } \right). Additionally, PTT-Net could achieve real-time performance (~40FPS) on NVIDIA 1080Ti GPU. Our code is open-sourced for the robotics community at https://github.com/shanjiayao/PTT.

In this paper, we focus on the problem of learning online an optimal policy for Active Visual Search (AVS) of objects in unknown indoor environments. We propose POMP++, a planning strategy that introduces a novel formulation on top of the classic Partially Observable Monte Carlo Planning (POMCP) framework, to allow training-free online policy learning in unknown environments. We present a new belief reinvigoration strategy that enables the use of POMCP with a dynamically growing state space to address the online generation of the floor map. We evaluate our method on two public benchmark datasets, AVD that is acquired by real robotic platforms and Habitat ObjectNav that is rendered from real 3D scene scans, achieving the best success rate with an improvement of >10% over the state-of-the-art methods.

Reinforcement Learning (RL) has made remarkable achievements, but it still suffers from inadequate exploration strategies, sparse reward signals, and deceptive reward functions. To alleviate these problems, a Population-guided Novelty Search (PNS) parallel learning method is proposed in this paper. In PNS, the population is divided into multiple sub-populations, each of which has one chief agent and several exploring agents. The chief agent evaluates the policies learned by exploring agents and shares the optimal policy with all sub-populations. The exploring agents learn their policies in collaboration with the guidance of the optimal policy and, simultaneously, upload their policies to the chief agent. To balance exploration and exploitation, the Novelty Search (NS) is employed in every chief agent to encourage policies with high novelty while maximizing per-episode performance. We apply PNS to the twin delayed deep deterministic (TD3) policy gradient algorithm. The effectiveness of PNS to promote exploration and improve performance in continuous control domains is demonstrated in the experimental section. Notably, PNS-TD3 achieves rewards that far exceed the SOTA methods in environments with sparse or delayed reward signals. We also demonstrate that PNS enables robotic agents to learn control policies directly from pixels for sparse-reward manipulation in both simulated and real-world settings.

3D object detection is a key component of many robotic applications such as self-driving vehicles. While many approaches rely on expensive 3D sensors such as LiDAR to produce accurate 3D estimates, methods that exploit stereo cameras have recently shown promising results at a lower cost. Existing approaches tackle this problem in two steps: first depth estimation from stereo images is performed to produce a pseudo LiDAR point cloud, which is then used as input to a 3D object detector. However, this approach is suboptimal due to the representation mismatch, as the two tasks are optimized in two different metric spaces. In this paper we propose a model that unifies these two tasks and performs them in the same metric space. Specifically, we directly construct a pseudo LiDAR feature volume (PLUME) in 3D space, which is then used to solve both depth estimation and object detection tasks. Our approach achieves state-of-the-art performance with much faster inference times when compared to existing methods on the challenging KITTI benchmark [1].

Achieving a proper balance between planning quality, safety and efficiency is a major challenge for autonomous driving. Optimisation-based motion planners are capable of producing safe, smooth and comfortable plans, but often at the cost of runtime efficiency. On the other hand, naïvely deploying trajectories produced by efficient-to-run deep imitation learning approaches might risk compromising safety. In this paper, we present PILOT a planning framework that comprises an imitation neural network followed by an efficient optimiser that actively rectifies the network’s plan, guaranteeing fulfilment of safety and comfort requirements. The objective of the efficient optimiser is the same as the objective of an expensive-to-run optimisation-based planning system that the neural network is trained offline to imitate. This efficient optimiser provides a key layer of online protection from learning failures or deficiency in out-of-distribution situations that might compromise safety or comfort. Using a state-of-the-art, runtime-intensive optimisation-based method as the expert, we demonstrate in simulated autonomous driving experiments in CARLA that PILOT achieves a seven-fold reduction in runtime when compared to the expert it imitates without sacrificing planning quality.

Path planning and smooth trajectory generation are critical capabilities for efficient navigation of mobile robots operating in challenging and cluttered environments. For real time and autonomous operations of mobile robots, intelligent algorithms, efficient and light-weight compute, and smooth trajectory are key components. In this work, we propose an intelligent, probabilistic Gaussian mixture model driven Bi-RRT (PG-RRT) algorithm which generates nodes in the most probable regions for faster convergence. The proposed algorithm is tested in various simulated environments including highly cluttered obstacles. The experimental results of PG-RRT are compared with state-of-the-art path planning algorithms. The results show significant improvement in the number of iterations (up to 26X) and runtime (up to 17.5X) demonstrating the superiority of the proposed PG-RRT algorithm.

Inferring a complete 3D geometry given an in-complete point cloud is essential in many vision and robotics applications. Previous work mainly relies on a global feature extracted by a Multi-layer Perceptron (MLP) for predicting the shape geometry. This suffers from a loss of structural details, as its point generator fails to capture the detailed topology and structure of point clouds using only the global features. The irregular nature of point clouds makes this task more challenging. This paper presents a novel method for shape completion to address this problem. The Transformer structure is currently a standard approach for natural language processing tasks and its inherent nature of permutation invariance makes it well suited for learning point clouds. Furthermore, the Transformer’s attention mechanism can effectively capture the local context within a point cloud and efficiently exploit its incomplete local structure details. A morphing-atlas-based point generation network further fully utilizes the extracted point Transformer feature to predict the missing region using charts defined on the shape. Shape completion is achieved via the concatenation of all predicting charts on the surface. Extensive experiments on the Completion3D and KITTI data sets demonstrate that the proposed PCTMA-Net outperforms the state-of-the-art shape completion approaches and has a 10% relative improvement over the next best-performing method.

Localization is an essential module that supports many intelligent functions of a mobile robot such as transportation or inspection. However, justifying that a localization module is sufficiently accurate for supporting all downstream tasks is one of the most difficult questions to answer in practice. To overcome this problem, we move away from the traditional calculation of pose errors and propose a new approach that instead evaluates the potential map inconsistency introduced by those pose errors.For this purpose, we propose a new metric, which we call Overlap Displacement Error (ODE). This metric measures the relative displacements between multiple overlapping sensor frustums with respect to the ground truth. All you need to compute this metric are a query trajectory, a ground truth trajectory and the sensor frustum used for mapping. Having the sensor frustum and the map representation as part of the metric, the ODE is customized to the hardware configuration and the mapping strategy. This design allows the analysis of pose accuracy in a space that matters to map creation, and also allows the identification of problems sitting in the interplay between localization and mapping. We demonstrate the potential of this new analysis tool on synthetic and the real-world sequences.