ABSTRACT This paper investigates the penetration performance of soil‐burrowing probes with different tip designs during shallow‐depth penetration in various media, including terrestrial soils (Hostun sand) and well‐characterized planetary soil simulants (LHS‐1 Lunar regolith simulant and MGS‐1 Martian regolith simulant). The analysis evaluates performance based on the pressure required to successfully penetrate the soil, comparing a conical tip design (i.e., the traditional tip design of penetrometers) with a plant root‐inspired design. For each soil type, three different levels of soil compaction were considered to verify how initial soil porosity affects penetration performance. The study involves both experimental and numerical analyses. Experimentally, penetration tests were conducted in chambers filled with Hostun sand, LHS‐1, and MGS‐1. Numerically, a three‐dimensional Discrete Element Model was developed to simulate probe penetration in soil packings with geomechanical properties of Hostun sand, LHS‐1, and MGS‐1, respectively. In accordance with the experimental findings, the modeling results show significant advantages of the plant root‐inspired tip design over the conical tip. Indeed, the plant root‐inspired design encountered lower soil resistance pressure during penetration across all soil types and compaction levels: the arithmetic mean values of the pressure reductions associated with the use of the bioinspired tip design resulted 25.5% experimentally and 25.4% numerically compared with the non‐bioinspired tip.

ABSTRACTSerial robotic manipulators find applications across diverse sectors such as industries, military, medical, space exploration, and underwater operations, where precision, efficiency, and safety are paramount. However, achieving effective control performance necessitates the effective resolution of the inverse kinematics (IK) problem inherent to manipulators. This challenge involves determining the requisite joint configurations to attain a desired endpoint position and orientation within the manipulator's workspace. The complexity arises due to the nonlinear equations and geometric relationships between cartesian space and joint space. Although various methods have been employed to address the IK problem in serial manipulators, there is a noticeable scarcity of publications comprehensively reviewing these techniques. Therefore, this paper undertakes an extensive review, thoroughly examining diverse IK methods applied in serial manipulators. This investigation of IK compiles literature spanning the last decade (2014–2023). Articles were collected from several web‐based repositories such as WoS, IEEE Xplore, ACM Digital Library, Google Scholar, and ScienceDirect, and a preferred reporting items for systematic reviews and meta‐analyses (PRISMA) protocol was adopted to acquire appropriate literature. The protocol involves examining the literature methodologies, major findings, and qualitative analysis. Thus, nearly 321 articles were screened and utilized in this review. Further, the IK methods are segregated into model‐driven methods (analytical, geometric, algebraic, numerical, and other model‐driven methods) and model‐free/data‐driven methods (metaheuristic (particle swarm optimization, genetic algorithm, and other metaheuristic methods), machine learning (artificial neural network, neuro‐fuzzy, and reinforcement learning), and other data‐driven methods). This paper offers a detailed exploration of limitations, challenges, performance comparisons, and future research directions of IK techniques. By conducting this study, we aim to assist prospective stakeholders in expanding their understanding of the remarkable potential of IK methods for serial manipulators.

ABSTRACT Agriculture is undergoing a significant transformation process with the help of robots. Weeding robots have made their way into the market, and they play a crucial role in automating the weeding process in the field. This study introduces a generalized concept of autonomy levels for currently available weeding robots in the field as well as a comprehensive rating system that allows for a comparison of different weeding robots, irrespective of their developmental stages. We examine the different abilities of market‐available robots, tractor implements, and smart weeding systems when it comes to navigating and recognizing crops and weeds in the field. A technological rating system is employed to rate the robots based on their advantages and critical aspects. To achieve this, we introduce a comparison system based on a measurable ability scale for the three important robot skills navigation, recognition, and target specificity. To demonstrate its applicability, we apply this system of robotic capability clustering to different available weeding system: the market‐available self‐propelled Farmdroid FD 20, the Farming GT, the experimental self‐propelled research platform Bonn Bot, the market‐available smart tractor implement Ecorobotix Ara, and the Bosch BASF Smart Sprayer. We discuss the outlook of interaction models from remote sensing and robots starting from swarm robot aspects to the spot farming, advantages, and limitations of GNSS‐ and vision‐based robots, as well as current challenges for the use of robots in the field and try to answer the question how robots can support farmers in existing workflows.

The cover image is based on the article FTR-bench: Benchmarking deep reinforcement learning for flipper-track robot control by Huimin Lu et al., 10.1002/rob.22528.

ABSTRACT Most existing vision‐based simultaneous localization and mapping systems and their variants still assume that the observation is absolutely static and cannot work well in dynamic environments. In this paper, we propose a direct geometrically constrained SLAM method based on target detection and depth image segmentation, named YGDD‐SLAM. The YGDD‐SLAM system can work robustly, accurately, and continuously in highly dynamic environments. The method first acquires static and potential dynamic feature points in the current frame through a target detection network. Then, dynamic targets are identified by combining the geometric change relationship between static and potential dynamic feature points between adjacent frames. To improve the accuracy of the dynamic judgment, the motion probability of the potential dynamic target in the past few frames is also used for judgment. Subsequently, the dynamic object regions at the pixel level are segmented out based on the double‐peak feature of the gray‐scale histogram of the dynamic target region in the depth image, which ultimately achieves the accurate deletion of all dynamic features points. Meanwhile, we validate YGDD‐SLAM on TUM data set and Bonn data set and prove that it significantly improves the localization accuracy and system stability in different types of dynamic environments.

ABSTRACT This paper presents the design and development of a bionic blind‐guiding robot based on a wheeled‐legged foldable structure, aimed at assisting visually impaired individuals in navigating complex environments. The robot integrates the advantages of both legged and wheeled robots, combining high mobility and environmental adaptability with improved movement efficiency and stability. The mechanical design incorporates a foldable structure inspired by guide dogs, enabling the robot to navigate uneven terrain, climb stairs, and curl up under furniture. A comprehensive kinematic and dynamic model is established to facilitate precise control of the robot's motion. The control system employs a Virtual Model Control approach, with force distribution optimized using Quadratic Programming to ensure stable locomotion. The robot features multiple motion modes, including legged, wheeled, and wheeled‐legged cooperative walking, allowing for flexible adaptation to different terrains. Human–robot interaction is enhanced through a guide saddle connection, providing tactile feedback and control options for the user. Experimental results demonstrate the robot's ability to maintain stability under varying traction forces and its effectiveness in guiding blind individuals through indoor and outdoor environments. The proposed design offers a promising solution for improving the mobility and independence of visually impaired individuals.

ABSTRACT Magnetic resonance imaging (MRI)‐guided focused ultrasound (MRg‐FUS) is an effective and noninvasive procedure for treating diseases such as neurological disorders. Phase adjustment on ultrasound transducers can only achieve a limited focal‐spot steering range. When treating large abdominopelvic targets, mechanical adjustment on the transducers' position and orientation is the prerequisite for enlarging the steering range. Therefore, we previously designed an MRI‐guided robot to manipulate the transducers to offer sufficient focal‐spot movement range. This could provide more modulation solutions to constructive ultrasound interference. However, full‐wave ultrasound propagation inside a patient's heterogeneous abdominal media is complex and nonlinear, posing significant challenges in ultrasound modulation and beam motion control. Here, we propose a novel learning‐based phase‐aberration correction and model‐free control framework for robot‐assisted MRg‐FUS treatments. The correction policy guarantees rapid aberration compensation within 5.0 ms. Submillimeter refocusing accuracy is achieved in both the liver (0.32 mm) and pancreas (0.51 mm), meeting clinical requirements for focal targeting. Our controller can accommodate nonlinear phase actuation with fast convergence (< 5.7 ms) and ensure accurate positional tracking with a mean error of 0.26 mm, without prior knowledge of inhomogeneous media. Compared with the conventional model‐based method, it contributes to 61.77%–70.39% mean error reduction without requiring model parameter tuning.

ABSTRACT Simultaneous localization and mapping (SLAM) is essential for mobile robotics, enabling autonomous navigation in dynamic, unstructured outdoor environments without relying on external positioning systems. These environments pose significant challenges due to variable lighting, weather conditions, and complex terrain. Visual‐Inertial SLAM has emerged as a promising solution for robust localization under such conditions. This paper benchmarks several open‐source visual‐Inertial SLAM systems, including traditional methods (ORB‐SLAM3, VINS‐Fusion, OpenVINS, Kimera, and SVO Pro) and learning‐based approaches (HFNet‐SLAM, AirSLAM), to evaluate their performance in unstructured natural outdoor settings. We focus on the impact of loop closing on localization accuracy and computational demands, providing a comprehensive analysis of these systems' effectiveness in real‐world environments and especially their application to embedded systems in outdoor robotics. Our contributions further include an assessment of varying frame rates on localization accuracy and computational load. The findings highlight the importance of loop closing in improving localization accuracy while managing computational resources efficiently, offering valuable insights for optimizing Visual‐Inertial SLAM systems for practical outdoor applications in mobile robotics. The data set and the benchmark code are available under https://github.com/iis-esslingen/vi-slam_lc_benchmark .

ABSTRACT Localization in mobile robotics is essential for achieving autonomy. Effective localization systems integrate data from multiple sensors to enhance state estimation and achieve accurate positioning. Accurate real‐time localization is crucial for robot control and trajectory following. Key challenges include initializing the inertial measurement unit (IMU) biases and the direction of gravity, as well as determining the metric scale with a monocular camera. Traditional visual–inertial (VI) initialization techniques rely on precise vision‐only motion assessments to address these issues. Multi‐sensor fusion faces challenges, such as precise calibration, initialization of sensor groups, and handling measurement errors with varying rates and delays. This paper introduces an Adaptive Tuna Schooling Optimization (ATSO) method to adjust localization strategies based on environmental conditions dynamically. The environmental factors affecting the localization process are considered in the optimization algorithm, and the position is optimally selected accordingly. Using Q‐learning with the Q‐DNN performs the decision‐making process based on past experiences. The dynamic adaptation of the weight parameter allows the algorithm to converge toward optimal solutions, reducing computational complexity. Experimental results demonstrate that the proposed approach improves localization performance, even in challenging conditions.

ABSTRACT Both visible and infrared images are important sources of intelligence information on the battlefield, and air‐to‐ground reconnaissance by UAV is an important means to obtain intelligence. However, there are great challenges in ground target detection and tracking, especially in complex battlefield environments. Aiming at the problem of insufficient accuracy of target detection by a single type of sensor in the battlefield environment at this stage, a target detection method by fusion of visible and infrared images is proposed in this paper, which is called ReconnaissanceFusion‐YOLO (RF‐YOLO), and with the help of infrared imagery, it can effectively improve the accuracy of target detection in the case of insufficient light. The performance of target detection in the battlefield is significantly improved by introducing two key innovative modules: dual feature fusion (DFF) module and feature fusion corrector (FFC) module. The DFF module enhances multi‐channel feature fusion through a novel concatenation and channel‐wise attention mechanism, while the FFC module performs feature correction between parallel streams using spatial and channel‐wise weights, addressing noise and uncertainty in different modalities. These modules are integrated on top of a dual‐stream YOLO architecture, allowing for effective fusion of visible and infrared information. RF‐YOLO was trained and evaluated using the FLIR data set, containing 5142 pairs of strictly aligned visible and infrared images. Results demonstrate that RF‐YOLO significantly outperforms benchmark networks in terms of robustness requirements. Specifically, the large model of RF‐YOLO achieves an mAP of 0.831, which is a significant improvement compared to the YOLOv5l inf benchmark's 0.739. This represents an improvement of over 12% in detection accuracy. Additionally, RF‐YOLO offers a Nano version that balances accuracy and speed. The Nano version achieves an mAP of 0.765, while maintaining a model size of only 11.5 MB, making it suitable for deployment on UAV edge computing devices with limited resources. To validate the practical applicability of our approach, this paper successfully implements target detection and tracking on a real UAV's edge computing device using the ROS system and SiameseRPN, combined with the proposed RF‐YOLO. Real‐world flight tests were conducted on an internal playground, demonstrating the effectiveness of our method in actual UAV applications. The system achieved a processing rate of approximately 10 fps at 640 × 640 resolution on an NVIDIA TX2 edge computing device, showcasing its real‐time performance capability in practical scenarios. This study contributes to enhancing UAV‐based battlefield reconnaissance capabilities by improving the accuracy and robustness of target detection and tracking in complex environments. The proposed RF‐YOLO method, along with its successful implementation on a real UAV platform, provides a promising solution for advanced military intelligence gathering and decision‐making support.