ICRA 2021 论文列表

Numerous robotic and control applications have strict real-time requirements, which, when violated, result in reduced quality of service or, in case of safety critical applications, might even have catastrophic consequences. To ensure that certain real-time constraints are satisfied, roboticists have relied on real-time safe frameworks, environments and middleware. With the introduction of ROS 2, alongside kernel patches such as PREEMPT_RT, there is an abundance of solutions to pick from. This paper compares OROCOS and ROS 2 over PREEMPT_RT and vanilla Linux kernels in a variety of benchmarks and draws conclusions on their performance in real-time critical applications. The outcome of the benchmark shows comparable performances under normal conditions. However, when the system is under stress both frameworks suffer in different fashions. Furthermore, the results show an accumulating error which over time violates the real-time requirements in both frameworks. These findings are paramount in conducting real world application with real-time constraints.

Domain randomisation is a very popular method for visual sim-to-real transfer in robotics, due to its simplicity and ability to achieve transfer without any real-world images at all. Nonetheless, a number of design choices must be made to achieve optimal transfer. In this paper, we perform a comprehensive benchmarking study on these different choices, with two key experiments evaluated on a real-world object pose estimation task. First, we study the rendering quality, and nd that a small number of high-quality images is superior to a large number of low-quality images. Second, we study the type of randomisation, and nd that both distractors and textures are important for generalisation to novel environments.

We propose a memory-constrained partition-based method to extract symbolic representations of the belief state and its dynamics in order to solve planning problems in a partially observable Markov decision process (POMDP). Our K-means partitioning strategy uses a fixed number of symbols to represent the partitions of the belief space and ensures the parameterization of the belief dynamics does not grow exponentially as the system dimension increases. By casting our problem as a partitioning of the POMDP, we can then solve planning problems using traditional symbolic planning solvers (such as HTN or A* solvers). Our work is motivated by an autonomous underwater vehicle navigation problem where the vehicle is affected by uncertain flow conditions and receives severely limited position observations. Simulation experiments are provided to validate the performance of the proposed algorithms.

We consider the problem of people search by a mobile social robot in case of a situation that cannot be solved by the robot alone. Examples are physically opening a closed door or operating an elevator. Based on the Behavior Tree framework, we create a modular and easily extendable action sequence with the goal of finding a person to assist the robot. By decomposing the Behavior Tree as a Discrete Time Markov Chain, we obtain an estimate of the probability and rate of success of the options for action, especially where the robot should wait or search for people. In a real-world experiment, the presented method is compared with other common approaches in a total of 588 test runs over the course of one week, starting at two different locations in a university building. We show our method to be superior to other approaches in terms of success rate and duration until a finding person and returning to the start location.

We present an algorithm for learning behavior trees for robotic task planning, which alleviates the need for time-intensive or infeasible manual design of control architectures. Our method involves representing the search space of behavior trees as a formal grammar and searching over this grammar by means of a new generalization of Monte Carlo tree search (MCTS) for directed acyclic graphs (DAGs), named MCDAGS. Additionally, our method employs simulated annealing to expedite the aggregation of the most functional subtrees. We present simulated experiments for a marine target search and response scenario, and an abstract task selection problem. Our results demonstrate that the learned behavior trees compare favorably with a manually-designed tree, and outperform baseline learning methods. Overall, these results show that our method is a viable technique for the automatic design of behavior trees for robotic task planning.

For autonomous vehicles, effective behavior planning is crucial to ensure safety of the ego car. In many urban scenarios, it is hard to create sufficiently general heuristic rules, especially for challenging scenarios that some new human drivers find difficult. In this work, we propose a behavior planning structure based on reinforcement learning (RL) which is capable of performing autonomous vehicle behavior planning with a hierarchical structure in simulated urban environments. Application of the hierarchical structure [1] allows the various layers of the behavior planning system to be satisfied. Our algorithms can perform better than heuristic-rule-based methods for elective decisions such as when to turn left between vehicles approaching from the opposite direction or possible lane-change when approaching an intersection due to lane blockage or delay in front of the ego car. Such behavior is hard to evaluate as correct or incorrect, but some aggressive expert human drivers handle such scenarios effectively and quickly. On the other hand, compared to traditional RL methods, our algorithm is more sample-efficient, due to the use of a hybrid reward mechanism and heuristic exploration during the training process. The results also show that the proposed method converges to an optimal policy faster than traditional RL methods.

The commonly used metrics for motion prediction do not correlate well with a self-driving vehicle’s system-level performance. The most common metrics are average displacement error (ADE) and final displacement error (FDE), which omit many features, making them poor self-driving performance indicators. Since high-fidelity simulations and track testing can be resource-intensive, the use of prediction metrics better correlated with full-system behavior allows for swifter iteration cycles. In this paper, we offer a conceptual framework for prediction evaluation highly specific to self-driving. We propose two complementary metrics that quantify the effects of motion prediction on safety (related to recall) and comfort (related to precision). Using a simulator, we demonstrate that our safety metric has a significantly better signal-to-noise ratio than displacement error in identifying unsafe events.

Scenarios requiring humans to choose from multiple seemingly optimal actions are commonplace, however standard imitation learning often fails to capture this behavior. Instead, an over-reliance on replicating expert actions induces inflexible and unstable policies, leading to poor generalizability in an application. To address the problem, this paper presents the first imitation learning framework that incorporates Bayesian variational inference for learning flexible nonparametric multi-action policies, while simultaneously robustifying the policies against sources of error, by introducing and optimizing disturbances to create a richer demonstration dataset. This combinatorial approach forces the policy to adapt to challenging situations, enabling stable multi-action policies to be learned efficiently. The effectiveness of our proposed method is evaluated through simulations and real-robot experiments for a table-sweep task using the UR3 6-DOF robotic arm. Results show that, through improved flexibility and robustness, the learning performance and control safety are better than comparison methods.

Aerial vehicles are revolutionizing the way filmmakers can capture shots of actors by composing novel aerial and dynamic viewpoints. However, despite great advancements in autonomous flight technology, generating expressive camera behaviors is still a challenge and requires non-technical users to edit a large number of unintuitive control parameters. In this work, we develop a data-driven framework that enables editing of these complex camera positioning parameters in a semantic space (e.g. calm, enjoyable, establishing). First, we generate a database of video clips with a diverse range of shots in a photo-realistic simulator, and use hundreds of participants in a crowd-sourcing framework to obtain scores for a set of semantic descriptors for each clip. Next, we analyze correlations between descriptors and build a semantic control space based on cinematography guidelines and human perception studies. Finally, we learn a generative model that can map a set of desired semantic video descriptors into low-level camera trajectory parameters. We evaluate our system by demonstrating that our model successfully generates shots that are rated by participants as having the expected degrees of expression for each descriptor. We also show that our models generalize to different scenes in both simulation and real-world experiments. Data and video found at: https://sites.google.com/view/robotcam.

This paper presents a simulation study of the problem of balancing a planar double pendulum in which the lower body (the leg) has been modified to include a spring-loaded passive prismatic joint. Robots of this kind can travel by hopping, and can also stand and balance on a single point. The purpose of this study is to investigate the degree to which a balance controller can cope with the large and rapidly changing forces from the spring. It is shown that good performance can be achieved using an existing balance controller if the spring-loaded joint is instrumented so that its position and velocity can be taken into account when calculating the state variables needed by the balance controller.

This paper presents a balance control technique for a novel wheel-legged robot. We first derive a dynamic model of the robot and then apply a linear feedback controller based on output regulation and linear quadratic regulator (LQR) methods to maintain the standing of the robot on the ground without moving backward and forward mightily. To take into account nonlinearities of the model and obtain a large domain of stability, a nonlinear controller based on the interconnection and damping assignment - passivity-based control (IDA-PBC) method is exploited to control the robot in more general scenarios. Physical experiments are performed with various control tasks. Experimental results demonstrate that the proposed linear output regulator can maintain the standing of the robot, while the proposed nonlinear controller can balance the robot under an initial starting angle far away from the equilibrium point, or under a changing robot height.

Research on tracking control has been on-going for many years. The accuracy and the practicality of the tracking control method has always been one of the most important aspects when designing the control strategy. Autonomous Underwater Vehicles are becoming increasingly important in the applications of ocean surveillance and military, etc. Therefore, this paper aims to develop a control method for autonomous underwater vehicles based on bioinspired neural dynamics. The proposed method practically solves the speed jump and chattering issues that are respectively in conventional backstepping and sliding mode controls with the aid of the bioinspired neural dynamics. In addition, the proposed control method also takes the dynamic uncertainties for the autonomous underwater vehicle into the consideration. The combined tracking method has relatively good overall performance for autonomous underwater vehicle against model uncertainties and disturbances.

Purely vision-based localization and mapping is a cost-effective and thus attractive solution to localization and mapping on smart ground vehicles. However, the accuracy and especially robustness of vision-only solutions remain rivalled by more expensive, lidar-based multi-sensor alternatives. We show that a significant increase in robustness can be achieved if taking non-holonomic kinematic constraints on the vehicle motion into account. Rather than using approximate planar motion models or simple, pair-wise regularization terms, we demonstrate the use of B-splines for an exact imposition of smooth, non-holonomic trajectories inside the 6 DoF bundle adjustment. We introduce both hard and soft formulations and compare their computational efficiency and accuracy against traditional solutions. Through results on both simulated and real data, we demonstrate a significant improvement in robustness and accuracy in degrading visual conditions.

In this paper, a degeneracy avoidance method for a point and line based visual SLAM algorithm is proposed. Visual SLAM predominantly uses point features. However, point features lack robustness in low texture and illuminance variant environments. Therefore, line features are used to compensate the weaknesses of point features. In addition, point features are poor in representing discernable features for the naked eye, meaning mapped point features cannot be recognized. To overcome the limitations above, line features were actively employed in previous studies. However, since degeneracy arises in the process of using line features, this paper attempts to solve this problem. First, a simple method to identify degenerate lines is presented. In addition, a novel structural constraint is proposed to avoid the degeneracy problem. At last, a point and line based monocular SLAM system using a robust optical-flow based lien tracking method is implemented. The results are verified using experiments with the EuRoC dataset and compared with other state-of-the-art algorithms. It is proven that our method yields more accurate localization as well as mapping results.

Motion planning among dynamic obstacles is an essential capability towards navigation in the real-world. Sampling-based motion planning algorithms find solutions by approximating the robot’s configuration space through a graph representation, predicting or computing obstacles’ trajectories, and finding feasible paths via a pathfinding algorithm. In this work, we seek to improve the performance of these subproblems by identifying regions critical to dynamic environment navigation and leveraging them to construct sparse probabilistic roadmaps. Motion planning and pathfinding algorithms should allow robots to prevent encounters with obstacles, irrespective of their trajectories, by being conscious of spatial context cues such as the location of chokepoints (e.g., doorways). Thus, we propose a self-supervised methodology for learning to identify regions frequently used for obstacle avoidance from local environment features. As an application of this concept, we leverage a neural network to generate hierarchical probabilistic roadmaps termed Avoidance Critical Probabilistic Roadmaps (ACPRM). These roadmaps contain motion structures that enable efficient obstacle avoidance, reduce the search and planning space, and increase a roadmap’s reusability and coverage. ACPRMs are demonstrated to achieve up to five orders of magnitude improvement over grid-sampling in the multi-agent setting and up to ten orders of magnitude over a competitive baseline in the multi-query setting.

Trajectory planning in dynamic environments can be decomposed into two sub-problems: 1) planning a path to avoid static obstacles, 2) then planning a speed profile to avoid dynamic obstacles. This is also called path-speed decomposition. In this work, we present a novel approach to solve the first sub-problem, motion planning with static obstacles. From an optimization perspective, motion planning for autonomous vehicles can be viewed as non-convex constrained nonlinear optimization, which requires a good enough initial guess to start and is often sensitive to algorithm parameters. We formulate motion planning as convex spline optimization. The convexity of the formulated problem makes it able to be solved fast and reliably, while guaranteeing a global optimum. We then reorganize the constrained spline optimization into a recurrent formulation, which further reduces the computational time to be linear in the optimization horizon size. The proposed method can be applied to both trajectory generation and motion planning problems. Its effectiveness is demonstrated in challenging scenarios such as tight lane changes and sharp turns.

Camera-equipped UAVs, or drones, are increasingly employed in a wide range of applications. Thus, ensuring their safe flight in areas containing people is a top priority. In this paper, a deep neural network-based method is proposed for the task of visual human crowd detection from UAV footage, allowing a drone to rapidly extract semantic segmentation maps from captured video frames during flight. These maps can be exploited (e.g., by a path planner) to define no-fly zones over, or near human crowds and, hence, enhance UAV flight safety. To this end, a novel neural architecture for binary (crowd/non- crowd) semantic segmentation from single RGB images is proposed, based on Convolutional Neural Networks (CNNs). It consists of a semantic segmentation and an image-to-image translation (I2I) neural branch. The overall network is trained using a novel multi-task loss function that addresses both tasks by processing the output of the corresponding branch. During inference, information flows across branches through additional skip synapses to further assist the crowd detection task. In order to evaluate the proposed method, we introduce a real and a synthetic human crowd RGB image dataset. The proposed method outperforms previous aerial crowd detection methods by a large margin and without any post-processing. Moreover, it demonstrates increased generalization ability, while running at real-time and near-real-time speeds on a ground computer and on embedded AI hardware, respectively.

This paper presents a technique that allows a robot to escort a human to their destination. Unlike tracking where the robot follows the human from behind, the proposed technique locates the robot in front of the human by incorporating human intention in addition to conventional motion prediction. Human head pose is used as an effective past-proven implicit indicator of intention. A particle filter allows accurate estimation and prediction of the non-Gaussian human trajectory. The predicted pose from both the human motion and intention determines the robot control action which leads to efficient autonomous escorting. Experimental analysis shows that the incorporation of the proposed human intention model reduces human position prediction error by approximately 40% when turning. Experimental validation with an omnidirectional mobile robotic platform shows successful and effective escorting compared to the conventional techniques.

Professional race-car drivers can execute extreme overtaking maneuvers. However, existing algorithms for autonomous overtaking either rely on simplified assumptions about the vehicle dynamics or try to solve expensive trajectory-optimization problems online. When the vehicle approaches its physical limits, existing model-based controllers struggle to handle highly nonlinear dynamics, and cannot leverage the large volume of data generated by simulation or real-world driving. To circumvent these limitations, we propose a new learning-based method to tackle the autonomous overtaking problem. We evaluate our approach in the popular car racing game Gran Turismo Sport, which is known for its detailed modeling of various cars and tracks. By leveraging curriculum learning, our approach leads to faster convergence as well as increased performance compared to vanilla reinforcement learning. As a result, the trained controller outperforms the built-in model-based game AI and achieves comparable over-taking performance with an experienced human driver.