ICRA 2021 论文列表

We present a new algorithm for deploying passenger robots in marsupial robot systems. A marsupial robot system consists of a carrier robot (e.g., a ground vehicle), which is highly capable and has a long mission duration, and at least one passenger robot (e.g., a short-duration aerial vehicle) transported by the carrier. We optimize the performance of passenger robot deployment by proposing an algorithm that reasons over uncertainty by exploiting information about the prior probability distribution of features of interest in the environment. Our algorithm is formulated as a solution to a sequential stochastic assignment problem (SSAP). The key feature of the algorithm is a recurrence relationship that defines a set of observation thresholds that are used to decide when to deploy passenger robots. Our algorithm computes the optimal policy in O(NR) time, where N is the number of deployment decision points and R is the number of passenger robots to be deployed. We conducted drone deployment exploration experiments on real-world data from the DARPA Subterranean challenge to test the SSAP algorithm. Our results show that our deployment algorithm outperforms other competing algorithms, such as the classic secretary approach and baseline partitioning methods, and is comparable to an offline oracle algorithm.

Shared autonomous mobility-on-demand systems hold great promise for improving the efficiency of urban transportation, but are challenging to implement due to the huge scheduling search space and highly dynamic nature of requests. This paper presents a novel optimal schedule pool (OSP) assignment approach to optimally dispatch high-capacity ride-sharing vehicles in real time, including: (1) an incremental search algorithm that can efficiently compute the exact lowest-cost schedule of a ride-sharing trip with a reduced search space; (2) an iterative online re-optimization strategy to dynamically alter the assignment policy for new incoming requests, in order to maximize the service rate. Experimental results based on New York City taxi data show that our proposed approach outperforms the state-of-the-art in terms of service rate and system scalability.

This work considers the optimal non-revisiting coverage tasks with a single non-redundant manipulator for the case when the object can be positioned at a predefined set of locations within the workcell. The scenario is often encountered in typical industrial settings, for instance when the object presents itself along a conveyor belt and its surface can not be serviced at a single location - the object being large or complex for that endeavour. Given the non-bijective nature of manipulator kinematics between task and joint space, without explicit consideration of joint-space continuity during its construction, a continuous coverage path designed in task-space may easily be truncated into intermittent segments where the manipulator needs to adopt a different configuration to continue the task, resulting in manipulator motions where the end-effector will need to lift off the surface, an altogether undesirable characteristic affecting the quality of the final product for smooth operations on objects such as polishing, painting or deburring. In this work, a novel algorithm to optimally partition the task-space whilst considering the various finite locations where the object may be stationed is proposed that ensures joint-space coverage continuity with minimal lift-offs. Results from the algorithm being challenged to achieve coverage of a number of objects, both in simulation and in real tests with an industrial manipulator, prove the effectiveness of the proposed planner when compared with classical coverage strategies faced with the same problem.

Robot arm placements are oftentimes a limitation in surgical preoperative procedures, relying on trained staff to evaluate and decide on the optimal positions for the arms. Given new and different patient anatomies, it can be challenging to make an informed choice, leading to more frequently colliding arms or limited manipulator workspaces. In this paper, we develop a method to generate the optimal manipulator base positions for the multi-port da Vinci surgical system that minimizes self-collision and environment-collision, and maximizes the surgeon’s reachability inside the patient. Scoring functions are defined for each criterion so that they may be optimized over. Since for multi-manipulator setups, a large number of free parameters are available to adjust the base positioning of each arm, a challenge becomes how one can expediently assess possible setups. We thus also propose methods that perform fast queries of each measure with the use of a proxy collision-checker. We then develop an optimization method to determine the optimal position using the scoring functions. We evaluate the optimality of the base positions for the robot arms on canonical trajectories, and show that the solution yielded by the optimization program can satisfy each criterion. The metrics and optimization strategy are generalizable to other surgical robotic platforms so that patient-side manipulator positioning may be optimized and solved.

Estimating the centroidal dynamics of legged robots is crucial in the context of multi-contact locomotion of legged robots. In this paper, we formulate the estimation of centroidal dynamics as a maximum a posteriori problem and we use a differential dynamic programming approach for solving it. The soundness of the proposed approach is first validated on a simulated humanoid robot, where ground truth data is available, enabling error analysis, and then compared to other alternatives of the state of the art, namely an extend Kalman filter and a recursive complementary filter. The results demonstrate that, compared to other approaches, the proposed method reduces the estimation error on the centroidal state in addition to ensuring the dynamics consistency of the state trajectory. Finally, the effectiveness of the proposed method is illustrated on real measurements, obtained from walking experiments with the HRP-2 humanoid robot.

In this paper, we propose an operational space control formulation for a planar N-link underactuated manipulator (PAN–1)1 with a passive first joint subject to actuator constraints (N ⩾ 3), covering both stabilization and tracking tasks. Such underactuated manipulators have an inherent first-order nonholonomic constraint, allowing us to project their dynamics to a space consistent with the nonholonomic constraint. Based on the constrained dynamics, we can design operational space controllers with respect to tasks assuming that all joints of the manipulator are active. Due to underactuation, we design a Quadratic Programming (QP) based controller to minimize the error between the desired torque commands and available motor torques in the null space of the constraint, as well as involve the constraint of motor outputs. The proposed control framework was demonstrated by stabilization and tracking tasks in simulations with both planar PA2 and PA3 manipulators. Furthermore, we verified the controller experimentally using a planar PA2 robot.

The actuator bandwidth limitation deteriorates the stability and performance of torque-based robot controllers. Operational space control is especially prone to this problem, since the limited bandwidth of a single actuator can reduce the performance of all related tasks simultaneously. In this article, an intuitive way to penalize low performance actuators is proposed to improve the performance of the operational space controller. The basic concept is to add joint torques only to high performance actuators, when the control bandwidth cannot reach the target level using all actuators. If that is not enough, additional torques are commanded to even higher performance actuators. This procedure can be executed recursively, meaning the controller can achieve almost maximum performance under the actuator bandwidth limitation. The proposed method was experimentally verified using the robot manipulator Franka Emika Panda.

Current robots are either expensive or make significant compromises on sensory richness, computational power, and communication capabilities. We propose to leverage smartphones to equip robots with extensive sensor suites, powerful computational abilities, state-of-the-art communication channels, and access to a thriving software ecosystem. We design a small electric vehicle that costs $50 and serves as a robot body for standard Android smartphones. We develop a software stack that allows smartphones to use this body for mobile operation and demonstrate that the system is sufficiently powerful to support advanced robotics workloads such as person following and real-time autonomous navigation in unstructured environments. Controlled experiments demonstrate that the presented approach is robust across different smartphones and robot bodies.

Intention prediction is a crucial task for Autonomous Driving (AD). Due to the variety of size and layout of intersections, it is challenging to predict intention of human driver at different intersections, especially unseen and irregular intersections. In this paper, we formulate the prediction of intention at intersections as an open-set prediction problem that requires context specific matching of the target vehicle state and the diverse intersection configurations that are in principle unbounded. We capture map-centric features that correspond to intersection structures under a spatial-temporal graph representation, and use two MAAMs (mutually auxiliary attention module) that cover respectively lane-level and exit-level intentions to predict a target that best matches intersection elements in map-centric feature space. Under our model, attention scores estimate the probability distribution of the open-set intentions that are contextually defined by the structure of the current intersection. The proposed model is trained and evaluated on simulated dataset. Furthermore, the model, trained on simulated dataset and without any fine tuning, is directly validated on in-house real-world dataset collected at 98 real-world intersections and exhibits satisfactory performance, demonstrating the practical viability of our approach.

During the off-road path following of a wheeled mobile robot in presence of poor grip conditions, the longitudinal velocity should be limited in order to maintain safe navigation with limited tracking errors, while at the same time being high enough to minimize travel time. Thus, this paper presents a new approach of online speed fluctuation, capable of limiting the lateral error below a given threshold, while maximizing the longitudinal velocity. This is accomplished using a neural network trained with a reinforcement learning method. This speed modulation is done side-by-side with an existing model-based predictive steering control, using a state estimator and dynamic observers. Simulated and experimental results show a decrease in tracking error, while maintaining a consistent travel time when compared to a classical constant speed method and to a kinematic speed fluctuation method.

This work presents a two part framework for online planning and execution of dynamic aerial motions on a quadruped robot. Motions are planned via a centroidal momentum-based nonlinear optimization that is general enough to produce rich sets of novel dynamic motions based solely on the user-specified contact schedule and desired launch velocity of the robot. Since this nonlinear optimization is not tractable for real-time receding horizon control, motions are planned once via nonlinear optimization in preparation of an aerial motion and then tracked continuously using a variational-based optimal controller that offers robustness to the uncertainties that exist in the real hardware such as modeling error or disturbances. Motion planning typically takes between 0.05-0.15 s, while the optimal controller finds stabilizing feedback inputs at 500 Hz. Experimental results on the MIT Mini Cheetah demonstrate that the framework can reliably produce successful aerial motions such as jumps onto and off of platforms, spins, flips, barrel rolls, and running jumps over obstacles.

In this paper, we develop an online learning-based visual tracking framework that can optimize the target model and estimate the scale variation for object tracking. We propose a recommender-based tracker, which is capable of selecting the representative convolutional neural network (CNN) layers and feature maps autonomously. In addition, the proposed recommender computes the weights of these layers and feature maps. A discriminative target percept of each recommended layer is reconstructed by the weighted sum of the recommended feature maps. Then the target model of the correlation filter is updated by the weighted sum of the target percepts. Thus, a sub-network is extracted from the pre-trained CNN backbone for the tracking process of a specific target. To deal with scale changes, we propose a spatiotemporal-based min-channel method to estimate the target size variation directly from CNN features. Experimental results on 50 benchmark datasets and video data from a rescue drone demonstrate that the proposed tracker is quite competitive with the state-of-the-art CNN-based trackers in terms of accuracy, scale adaptation, and robustness for UAV-related applications.

This paper presents an online informative path planning approach for active information gathering on three-dimensional surfaces using aerial robots. Most existing works on surface inspection focus on planning a path offline that can provide full coverage of the surface, which inherently assumes the surface information is uniformly distributed hence ignoring potential spatial correlations of the information field. In this paper, we utilize manifold Gaussian processes (mGPs) with geodesic kernel functions for mapping surface information fields and plan informative paths online in a receding horizon manner. Our approach actively plans information-gathering paths based on recent observations that respect dynamic constraints of the vehicle and a total flight time budget. We provide planning results for simulated temperature modeling for simple and complex 3D surface geometries (a cylinder and an aircraft model). We demonstrate that our informative planning method outperforms traditional approaches such as 3D coverage planning and random exploration, both in reconstruction error and information-theoretic metrics. We also show that by taking spatial correlations of the information field into planning using mGPs, the information gathering efficiency is significantly improved.

We present a novel approach to the formation controlling of aerial robot swarms that demonstrates the flocking behavior. The proposed method stems from the Unmanned Aerial Vehicle (UAV) dynamics; thus, it prevents any unattainable control inputs from being produced and subsequently leads to feasible trajectories. By modeling the inter-agent relationships using a pairwise energy function, we show that interacting robot swarms constitute a Markov Random Field. Our algorithm builds on the Mean-Field Approximation and incorporates the collective behavioral rules: cohesion, separation, and velocity alignment. We follow a distributed control scheme and show that our method can control a swarm of UAVs to a formation and velocity consensus with real-time collision avoidance. We validate the proposed method with physical and high-fidelity simulation experiments.

Legged robot locomotion requires the planning of stable reference trajectories, especially while traversing uneven terrain. The proposed trajectory optimization framework is capable of generating dynamically stable base and footstep trajectories for multiple steps. The locomotion task can be defined with contact locations, base motion or both, making the algorithm suitable for multiple scenarios (e.g., presence of moving obstacles). The planner uses a simplified momentum-based task space model for the robot dynamics, allowing computation times that are fast enough for online replanning. This fast planning capability also enables the quadruped to accommodate for drift and environmental changes. The algorithm is tested on simulation and a real robot across multiple scenarios, which includes uneven terrain, stairs and moving obstacles. The results show that the planner is capable of generating stable trajectories in the real robot even when a box of 15 cm height is placed in front of its path at the last moment.

In this paper, we propose a novel online algorithm for motion similarity measurements during human-robot interaction (HRI). Specifically, we formulate a Segment-based Online Dynamic Time Warping (SODTW) algorithm that can be used for understanding of repeated and cyclic human motions, in the context of rehabilitation or social interaction. The algorithm can estimate both the human-robot motion similarity and the time delay to initiate motion and combine these values as a metric to adaptively select appropriate robot imitation repertoires. We validated the algorithm offline by post-processing experimental data collected from a cohort of 55 subjects during imitation episodes with our social robot Zeno. Furthermore, we implemented the algorithm online on Zeno and collected further experimental results with 13 human subjects. These results show that the algorithm can reveal important features of human movement including the quality of motion and human reaction time to robot stimuli. Moreover, the robot can adapt to appropriate human motion speeds based on similarity measurements calculated using this algorithm, enabling future adaptive rehabilitation interventions for conditions such as Autism Spectrum Disorders (ASD).

In this paper, we present a highly efficient Divergent Component of Motion (DCM) reference trajectory generator capable of adapting online to large perturbations acting on the center-of-mass (push recovery) and on the swing foot (stumble recovery). For push recovery, we propose an analytic solution for a footstep adjustment strategy based on the DCM dynamics. The proposed algorithm considers double support phases explicitly and is active throughout the motion, i.e., during both single and double support phases. For stumble recovery, we introduce a continuous DCM trajectory adaptation based on the instantaneous tracking error of the swing foot. Our method is highly efficient, computing a push recovery solution within 10 microseconds on the robot hardware. Furthermore, it achieves robust locomotion for large external perturbations, which we demonstrate in simulations and experiments with the humanoid robot TORO.

In this paper, we consider the dynamic multi-robot distribution problem where a heterogeneous group of networked robots is tasked to spread out and simultaneously move towards multiple moving task areas while maintaining connectivity. The heterogeneity of the system is characterized by various categories of units and each robot carries different numbers of units per category representing heterogeneous capabilities. Every task area with different importance demands a total number of units contributed by all of the robots within its area. Moreover, we assume the importance and the total number of units requested from each task area is initially unknown. The robots need first to explore, i.e., reach those areas, and then be allocated to the tasks so to fulfill the requirements. The multi-robot distribution problem is formulated as designing controllers to distribute the robots that maximize the overall task fulfillment while minimizing the traveling costs in presence of connectivity constraints. We propose a novel connectivity-aware multi-robot redistribution approach that accounts for dynamic task allocation and connectivity maintenance for a heterogeneous robot team. Such an approach could generate sub-optimal robot controllers so that the amount of total unfulfilled requirements of the tasks weighted by their importance is minimized and robots stay connected at all times. Simulation and numerical results are provided to demonstrate the effectiveness of the proposed approaches.

Ultimately, feedback control is about making adjustments using current state information in order to meet an objective in the future. In the control of bipedal locomotion, linear velocity of the center of mass has been widely accepted as the primary variable around which feedback control objectives are formulated. In this paper, we argue that it is easier to predict the one-step ahead evolution of angular momentum about the contact point than it is to make a similar prediction for linear velocity, and hence it provides a superior quantity for feedback control. So as not to confuse the benefits of predicting angular momentum with any other control design decisions, we reformulate the standard LIP model in terms of angular momentum and show how to regulate swing foot touchdown position at the end of the current step so as to meet an angular momentum objective at the end of the next step. We implement the resulting feedback controller on the 20 degree-of-freedom bipedal robot, Cassie Blue, where each leg accounts for nearly one-third of the robot’s total mass of 32 Kg. Under this controller, the robot achieves fast walking, rapid turning while walking, large disturbance rejection, and locomotion on rough terrain.