Robotics and Autonomous Systems最新文献

Flexible Manufacturing Systems (FMS) is known as one of the recurring themes that possess these promising characteristics with a synergistic combination of productivity-efficiency transport and flexibility through a number of machine tools alongside other material handling devices. In FMS, mobile robots are commonly deployed in material handling system for the purpose of increasing the efficiency and productivity of the manufacturing process. A reliable, efficient, and optimal scheduling is the most important in manufacturing system. The scheduling problems can become highly complex, especially in large-scale systems with numerous tasks and constraints. Thus, schedule optimization becomes crucial to enhance target performance by determining the best allocations and sequences of resources under specified constraints. Recently, Genetic Algorithm (GA) is a remarkably applicable search algorithm to solve scheduling problems to the way that near optimal could be found. While the performance of GA much depends on the selection of the main parameters, a standard GA may suffer from the issue of premature convergence due to the lack of control on its parameters especially crossover and mutation operators. As there is no specific method or way to tune these parameters, the algorithm is prone to converge on the local optimum, thereby leading to performance degradation. To overcome such flaw, this paper proposed an improved Genetic Algorithm using an adaptive Fuzzy Logic to control crossover and mutation operators (FGAOC) for the solution to the NP-hard problem of scheduling mobile robot within Job-Shop FMS environment. The proposed algorithm has been evaluated in several case studies such as small and large-scale problem, various numbers of mobile robots and the 40-test benchmark problem. The results have demonstrated that the proposed FGAOC has delivered a good performance in exploration-exploitation activities with better solution quality.

We address a motion planning and control problem for mobile robots to satisfy rich, time-varying tasks expressed as Signal Temporal Logic (STL) specifications. The specifications may include tasks with nested temporal operators or time-conflicting requirements (e.g., achieving periodic tasks or tasks defined within the same time interval). Moreover, the tasks can be defined in locations changing with time (i.e., dynamic targets), and their future motions are not known a priori. This unpredictability requires an online control approach which motivates us to investigate the use of control barrier functions (CBFs). The proposed CBFs take into account the actuation limits of the robots and a feasible sequence of STL tasks. They define time-varying feasible sets of states the system must always stay inside. We show the feasible sequence generation process that even includes the decomposition of periodic tasks and alternative scenarios due to disjunction operators. The sequence is used to define CBFs, ensuring STL satisfaction. We also show some theoretical results on the correctness of the proposed method. We illustrate the benefits of the proposed method and analyze its performance via simulations and experiments with aerial robots.

Exploration and safe navigation in previously unknown GPS-denied obstructed areas are major challenges for autonomous robots when deployed in subterranean environments. In response, this work establishes an Exploration-Planning framework developed for the real-world deployment of Micro Aerial Vehicles (MAVs) in subterranean exploration missions. The fundamental task for an autonomous MAV to navigate in an unknown area, is to decide where to look while navigating such that the MAV will acquire more information about the surrounding. The work presented in this article focuses around 3D exploration of large-scale caves or multi-branched tunnel like structures, while still prioritizing the Look-Ahead and Move-Forward approach for fast navigation in previously unknown areas. In order to achieve such exploration behaviour, the proposed work utilizes a two-layer navigation approach. The first layer deals with computing traversable frontiers to generate the look ahead poses in the constrained field of view, aligned with the MAV’s heading vector that leads to rapid continuous exploration. The proposed frontier distribution based switching goal selection approach allows the MAV to explore various terrains, while still regulating the MAV’s heading vector. The second layer of the proposed scheme deals with global cost based navigation of the MAV to the potential junction in a multi-branched tunnel system leading to a continuous exploration of partially seen areas. The proposed framework is a combination of a novel frontier goal selection approach, risk-aware expandable grid based path planning, nonlinear model predictive control and artificial potential fields based on local reactive navigation, obstacle avoidance, and control for the autonomous deployment of MAVs in extreme environments.

Research in autonomous systems is gaining popularity both in academia and industry. These systems offer comfort, new business opportunities such as self-driving taxis, more efficient resource utilization through car-sharing, and most importantly, enhanced road safety. Different forms of Vehicle-to-Everything (V2X) communication have been under development for many years to enhance safety. Advances in wireless technologies have enabled more data transmission with lower latency, creating more possibilities for safer driving. Collaborative perception is a critical technique to address occlusion and sensor failure issues in autonomous driving. To enhance safety and efficiency, recent works have focused on sharing extracted features instead of raw data or final outputs, leading to reduced message sizes compared to raw sensor data. Reducing message size is important to enable collaborative perception to coexist with other V2X applications on bandwidth-limited communication devices.

To address this issue and significantly reduce the size of messages sent while maintaining high accuracy, we propose our model: LaCPF (Late Collaborative Perception Fusion), which uses deep learning for late fusion. We demonstrate that we can achieve better results while using only half the message size over other methods. Our late fusion framework is also independent of the local perception model, which is essential, as not all vehicles on the road will employ the same methods. Therefore LaCPF can be scaled more quickly as it is model and sensor-agnostic.

In this paper, we propose a manipulation planning method for cable wiring in the assembly of electric appliances etc. We address a scenario where a robot grasps a connector attached to the end of a cable and has to bring the connector to a socket. To accomplish this automatically, we propose a novel manipulation planning method. The method extends the Encode-Manipulate-Decode network (EMD net), which can predict shape changes of deformable objects and generate robot motion sequences for producing desired shape transitions. This enables us to find connector trajectories that avoid collision between the cable and the surrounding environment. We conducted experiments with several different cable lengths. We also introduce some functions required for real-world wiring, such as online cable shape modification. Experimental results show that the proposed method can achieve stable manipulation of real cables.

3D human pose estimation from a single image is still a challenging problem despite the large amount of work that has been performed in this field. Generally, most methods directly use neural networks and ignore certain constraints ($e.g.$, reprojection constraints, joint angle, and bone length constraints). While a few methods consider these constraints but train the network separately, they cannot effectively solve the depth ambiguity problem. In this paper, we propose a GAN-based model for 3D human pose estimation, in which a reprojection network is employed to learn the mapping of the distribution from 3D poses to 2D poses, and a discriminator is employed for 2D–3D consistency discrimination. We adopt a novel strategy to synchronously train the generator, the reprojection network and the discriminator. Furthermore, inspired by the typical kinematic chain space (KCS) matrix, we introduce a weighted KCS matrix and take it as one of the discriminator’s inputs to impose joint angle and bone length constraints. The experimental results on Human3.6M show that our method significantly outperforms state-of-the-art methods in most cases.

Aiming at the problem that Ant Colony Optimization (ACO) is subject primarily to the parameters, we propose a hybrid algorithm SOA-ACO-2Opt to optimize the ACO parameter combination through Seagull Optimization Algorithm (SOA) to strengthen ACO’s search capability. To obtain a uniform initial distribution of the ACO parameter combination, we incorporated the Kent Chaos Map (KCM) to randomly initialize the seagull’s position, reducing the tendency of SOA to fall into the local optimum. To avoid slow calculation speed and premature convergence of ACO, we improved the adaptive multi-population mechanism to reduce repeated redundant calculations and used the $ϵ-greedy$ and default strategy, respectively, to update the ants’ position. 2Opt is applied to find shorter paths in each iteration. In addition, when AUV navigates on the planned path, it may encounter obstacles. Therefore, this paper proposes an autonomous obstacle avoidance algorithm based on forward-looking sonar to ensure safety during tasks. SOA-ACO-2Opt is verified against twelve different problems extracted from TSPLIB and compared with some state-of-the-art algorithms. Furthermore, sea trials were carried out for several representative marine engineering applications of TSP and obstacle avoidance. Experimental results show that this work can significantly improve AUV’s work efficiency and intelligence and protect the AUV’s safety.

Recent advancements in sensors and deep learning techniques have improved reliability of robotic perceptual systems, but current systems are not robust enough for real-world challenges such as occlusions and sensing uncertainties in cluttered scenes. To overcome these issues, active or interactive perception actions are often necessary, such as sensor repositioning or object manipulation to reveal more information about the scene. Existing perception systems lack a comprehensive approach that incorporates both active and interactive action spaces, thereby limiting the robot’s perception capabilities. Moreover, these systems focus on exploring a single object or scene, without utilizing object information to guide the exploration of multiple objects. In this work, we propose an object-aware hybrid perception system that selects the next best action by considering both active and interactive action spaces and enhances the selection process with an object-aware approach to guide the cognitive robot operating in tabletop scenarios. Novel volumetric utility metrics are used to evaluate actions that include positioning sensors from a heterogeneous set or manipulating objects to gain a better perspective of the scene. The proposed system maintains the volumetric information of the scene that includes semantic information about objects, enabling it to exploit object information, associate occlusion with corresponding objects, and make informed decisions about object manipulation. We evaluate the performance of our system both in simulated and real-world experiments using a Baxter robotic platform equipped with two arms, RGB and depth cameras. Our experimental results show that the proposed system outperforms the compared state-of-the-art methods in the given scenarios, achieving an 11.2% performance increase.

Trajectory planning is a crucial aspect of macro-micro robotic systems (MMRSs), especially when the system has high degrees of freedom (DOFs). In the field of robotic polishing, the MMRS is usually composed of an industrial robot and an end-effector, which is responsible for polishing force control. Therefore, the compliance of the macro-robot can be minimized by trajectory planning to reduce its impact on the micro-robot. This study proposes a trajectory planning strategy based on Gauss pseudospectral method for a 9-DOF MMRS. Different from traditional sequential solution strategies, it can be used to obtain an approximate global optimal trajectory. Firstly, the velocity-level kinematics model of MMRS is built, which comprehensively considers the workpiece placement pose and task redundancy. Secondly, an optimal control model for trajectory planning is developed through an effective variable allocation. On the premise of considering traditional trajectory smoothness constraints, a constraint on manipulability is additionally analyzed to avoid reaching a singular configuration during compliance optimization. Thirdly, a Gauss pseudospectral framework based on the optimal control model is proposed, and the costate mapping theorem is proved. The latter provides a theoretical basis for the efficiency and accuracy of the proposed method. Finally, comparison results demonstrate the effectiveness of the proposed method.

Simultaneous localization and mapping (SLAM) aims to solve the problems of robot localization and mapping in unknown environments. Recent related research usually uses closed-loop correction or integrate GNSS (Global Navigation Satellite System) into the optimization framework to ensure the long-term system accuracy and stability at the cost of huge computational resources. To balance efficiency and accuracy, this paper presents a fast and stable GNSS-LiDAR-inertial state estimator: GNSS, LiDAR and IMU are fused to achieve state estimation from coarse to fine, thereby improving the system accuracy and stability; the overall framework based on iterated error-state Kalman filter makes our system faster than most multi-sensor fusion SLAM. We also design a fast GNSS online initialization method and a multi-layer outlier rejection mechanism for our system. In addition, we apply backward propagation for multi-sensor motion compensation to overcome the limitations of fast motion. Finally, comprehensive experiments demonstrate that our system achieves higher accuracy and computational efficiency than the state-of-the-art navigation systems on the latest challenging public datasets, and perform equally well in the real environment.