Journal of Field Robotics最新文献

While visual Simultaneous Localization and Mapping systems are well studied and achieve impressive results in indoor and urban settings, natural, outdoor, and open-field environments are much less explored and still present relevant research challenges. Visual navigation and local mapping have shown a relatively good performance in open-field environments. However, globally consistent mapping and long-term localization still depend on the robustness of loop detection and closure, for which the literature is scarce. In this work, we propose a novel method to pave the way towards robust loop detection in open fields, particularly in agricultural settings, based on local feature search and stereo geometric refinement, with a final stage of relative pose estimation. Our method consistently achieves good loop detections, with a median error of 15 cm. We aim to characterize open fields as a novel environment for loop detection, understanding the limitations and problems that arise when dealing with them. Code is available at: https://github.com/CIFASIS/StereoLoopDetector

In recent years, with the continuous development of human exploration of the natural world, there has been a growing demand across various fields for robots capable of free movement in diverse environments. In this study, we address the issue of compliant control for a hexapod robot in diverse environments and propose a novel control method based on an adaptive impedance model for position control. Our approach enables the hexapod robot to stabilize foot force on complex terrains while preserving balance and body height. Specifically, we analyze the algorithm's parameters and stability by establishing the hexapod robot's structural and impedance control models. To tackle this challenge, we introduce an adaptive impedance control algorithm that estimates environmental parameters using Lyapunov's asymptotic stability theorem and achieves tracking of actual foot-end forces to desired foot forces. Furthermore, to ensure body stability and height, we incorporate attitude feedback and body feedback. Experimental results from foot force control experiments conducted on a multilegged robot demonstrate that our proposed algorithm enhances the adaptability and robustness of the robot. This research holds significant implications for the stable control of hexapod robots in complex environments and has promising practical applications.

Replacing manned ships with unmanned surface vehicle (USV) for oil spill containment can reduce the consumption of manpower and resources. This article studies the formation control method of dual USVs during the process of capturing oil spill, based on the engineering background of towing oil boom by dual USVs oil spill recovery system. To calculate the drag force of the oil boom acting on the USV, the shape of the oil boom is simplified into a catenary, the oil boom is modeled, and the hydrodynamic numerical simulation is carried out. To address the issue of “winding” phenomenon and “towing separation,” the formation behavior is designed when double USVs are towing oil fences to capture oil spill. In response to the problem of low task execution efficiency caused by fixed behavior priority in traditional null-space-based (NSB) behavior fusion methods, a fuzzy-priority NSB (FNSB) behavior fusion formation method is proposed by combining fuzzy control with NSB behavior fusion method. In the FNSB behavior fusion formation control method, a smooth transition rule is introduced to make the behavior priority change, USVs can still maintain good formation performance, ensuring the smooth execution of oil spill recovery tasks. Simulation shows that FNSB behavior fusion formation method based on flexible transition rules can improve the rounding efficiency by 26.6% in the environment without obstacles and 37.2% in the environment with multiple obstacles compared with the NSB method. The effectiveness and practicality of this method have been verified through simulation experiments and field experiments on the “Dolphin” series of small USVs.

In ground mobile robots, effective path planning relies on their ability to assess the types and conditions of the surrounding terrains. Neural network-based methods, which primarily use visual images for terrain classification, are commonly employed for this purpose. However, the reliability of these models can vary due to inherent discrepancies between the training images and the actual environment, leading to erroneous classifications and operational failures. Retraining models with additional images from the actual operating environment may enhance performance, but obtaining these images is often impractical or impossible. Moreover, retraining requires substantial offline processing, which cannot be performed online by the robot within an embedded processor. To address this issue, this paper proposes a neural network-based terrain classification model, trained using an existing data set, with a novel uncertainty rejection filter (URF) for terrain-aware path planning of mobile robots operating in unknown environments. A robot, equipped with a pretrained model, initially collects a small number of images (10 in this work) from its current environment to set the target uncertainty ratio of the URF. The URF then dynamically adjusts its sensitivity parameters to identify uncertain regions and assign associated traversal costs. This process occurs entirely online, without the need for offline procedures. The presented method was evaluated through simulations and physical experiments, comparing the point-to-point trajectories of a mobile robot equipped with (1) the neural network-based terrain classification model combined with the presented adaptive URF, (2) the classification model without the URF, and (3) the classification model combined with a nonadaptive version of the URF. Path planning performance measured the Hausdorff distances between the desired and actual trajectories and revealed that the adaptive URF significantly improved performance in both simulations and physical experiments (conducted 10 times for each setting). Statistical analysis via t-tests confirmed the significance of these results.

Semantic segmentation, a task of assigning class labels to each pixel in an image, has found applications in various real-world scenarios, including autonomous driving and scene understanding. However, its widespread use is hindered by the high computational burden. In this paper, we propose an efficient semantic segmentation method based on Feature Cascade Fusion Network (FCFNet) to address this challenge. FCFNet utilizes a dual-path framework comprising the Spatial Information Path (SIP) and the Context Information Path (CIP). SIP is a shallow structure that captures the local dependencies of each pixel to improve the accuracy of detailed segmentation. CIP is the main branch with a deeper structure that captures sufficient contextual information from input features. Moreover, we design an Efficient Receptive Field Module (ERFM) to enlarge the receptive field in the SIP. Meanwhile, Attention Shuffled Refinement Module is used to refine feature maps from different stages. Finally, we present an Attention-Guided Fusion Module to fuse the low- and high-level feature maps effectively. Experimental results show that our proposed FCFNet achieves 70.7% mean intersection over union (mIoU) on the Cityscapes data set and 68.1% mIoU on the CamVid data set, respectively, with inference speeds of 110 and 100 frames per second (FPS), respectively. Additionally, we evaluated FCFNet on the Nvidia Jetson Xavier embedded device, which demonstrated competitive performance while significantly reducing power consumption.

This paper presents the design and optimization of a climbing robot. The design of a ladder-climbing robot is done with fundamental mathematical considerations. The designed robot is robust enough to manage all environmental calamities, and at the same time, it is optimized for lightweight to reduce the actuator's cost and ease of transportation. An analytical evaluation is carried out for both static and dynamic conditions to determine strength and motion characteristics. The multiobjective optimization of the design parameters of a ladder-climbing robot is done to obtain optimized values of design parameters. The formulation of an optimization problem that considers the minimization of weight and natural frequency is performed. Using an evolutionary genetic algorithm (GA) for the multicriteria optimization problem is solved, and a Pareto front solution is obtained. The optimal values of the parameters are decided based on the knee selection technique. As both objective functions are contradictory, the optimum results significantly improve the robot's performance. Controlling the proportional–integral–derivative (PID) parameters is crucial as the robot climbs with a two-point contact gait pattern. The controlling parameters impart stability to the robot. PID parameters like proportional, integral and derivative gain are tunned using the GA. Finally, the developed prototype is tested on the ladders of the tower, and satisfactory climbing motion is achieved.

Screw-driven robot possesses outstanding omnidirectional mobility. Due to the friction between wheels and the ground, the screw-driven robot can traverse rough terrain flexibly. This paper proposes a screw-driven robot, which is equipped with four screws wheels. The directions of thread between adjacent screws are opposite to each other. By controlling the rotating speed and direction of the wheels, the omnidirectional traction could be generated, so the robot possesses the omnidirectional mobility in the field. The robot can work in two locomotion modes, the screw-driven mode and the wheel-driven mode. In addition, the robot can be used for the forestry information-collecting tasks, which improve the efficiency of the forestry task. Two prototypes are fabricated. A series of experiments and field tests are conducted, and the results verify that the robot can traverse rough terrain with the omnidirectional mobility.

Terrain-aided navigation (TAN) is a viable method to achieve long-term underwater navigation for long-range autonomous underwater vehicles (AUVs). However, the high-accuracy positioning results of most TAN systems rely on precise a priori seabed terrain maps, which restricts their applicability to a few areas with accurate bathymetric measurements of the seabed terrain. This article introduces a TAN system based on the General Bathymetric Chart of the Oceans (GEBCO) data set for global marine applications. Specifically, to address the low accuracy and poor robustness of the TAN system with imprecise bathymetric measurement and low-resolution data from the GEBCO data set, this article proposes a multizonotope TAN method based on set-membership filter (SMF) theory. The SMF theory is employed to handle the unknown distribution of the measurement noise from the GEBCO data set, introducing a multizonotope measurement update model to achieve more precise navigational results while addressing the perceptual ambiguity caused by self-similar terrain. The smoothness of the terrain is incorporated as a parameter in the generation ranges of multizonotope, enabling adaptive adjustment based on terrain smoothness to reduce costs and enhance navigational performance. The accuracy and robustness of the proposed method are verified through all shipboard experiments, publicly available data sets, and AUV experiments. Compared with state-of-the-art TAN methods, the average and maximum positioning errors have decreased by 64.83% and 48.84%, respectively. Finally, based on the experimental results, a preliminary distribution of suitable areas in the oceans is provided.

This paper aims to develop a miniature mobile robot suitable to assist archeologists in their first exploration of unknown underground tombs. Due to the rather complex and irregular terrains in the tombs and inspired by the classic RHex design, we have developed a two-segment articulated robot (A-RHex) with two RHex design units. The robot is compact and lightweight, with dimensions of 25 cm long, 6.5 cm wide, 7 cm high, and weighs 283 g. To assist the robot in entering the tomb, we have also designed a deployment platform that can take the robots underground through a 10-cm exploration hole. We introduce the overall design, control, and communication methods of A-RHex, and theoretically analyze how the articulated design can improve the stability of the robot on slopes. Laboratory experiments and field testings at two real archeological excavation sites in China have validated A-RHex's mechanical design, control strategies, communications, and capabilities for pre-exploration of open and closed tombs. We believe that this kind of robot with high terrain adaptability and a small profile may become an important tool for field archeology in the future.

Given the demanding and unpredictable operational conditions, autonomous underwater vehicles (AUVs) often encounter different propulsion faults, leading to significant economic losses and mission impairments. To address this challenge, vibratory time-series features can be extracted for the precise propulsion fault diagnosis of AUVs. A squeeze-and-excitation (SE) attention residual network (SEResNet) is therefore put forward to enhance the feature extraction for AUV propulsion fault diagnosis. By leveraging the vibratory time-series data obtained from the AUV, an SE attention mechanism is embedded into a residual network. This integration facilitates the extraction of pertinent vibratory fault features, subsequently utilized for accurate diagnosis of any propulsion faults. The effectiveness of the proposed SEResNet was validated through its application to an actual experimental AUV, with comparison against the state-of-the-arts. The results reveal that the present SEResNet outperforms all other comparison methods in terms of diagnosis performance for AUV propulsion faults.