International Journal of Intelligent Robotics and Applications最新文献

Marine seismic surveys rely on the precise localization of seismic airguns to ensure high-quality data acquisition. The current state-of-the-art systems for airgun localization, such as Ultra-Short Baseline (USBL), Long Baseline (LBL), and GPS-aided and hybrid systems, provide reliable positioning but are often costly, operationally complex or limited to surface tracking. To address these limitations, this research presents a novel, cost-effective, and robust underwater localization system designed explicitly for real-time trajectory tracking of seismic airguns. The proposed system integrates a low-cost Inertial Motion Unit (IMU), an open-source, modular, and low-power acoustic modem, and a depth sensor, employing an Extended Kalman Filter (EKF) using Robot Operating System (ROS) for the development of the sensor fusion and localization algorithm. The system achieves a position accuracy within 0.3-2 m, meeting the tolerances required for seismic surveys. Compared to USBL + INS and GPS-aided systems, the proposed system provides comparable precision while significantly reducing deployment complexity and operational costs. Unlike LBL systems, it does not rely on pre-installed seabed transponders, enhancing adaptability to different operational environments. Furthermore, its capacity to operate underwater without surface GPS dependency overcomes the limitations of existing systems in deep-sea or complex acoustic environments. The system is expected to enhance marine seismic data quality by enabling real-time positioning tracking, reduce marine seismic exploration times, and mitigate potential environmental impacts on marine ecosystems. This innovation bridges the gap between precision, affordability, and environmental sustainability in marine seismic exploration, making it a promising alternative for integrating trajectory position estimates of airguns into seismic survey workflows, helping to enhance the efficiency and effectiveness of marine seismic surveys worldwide.

Supplementary information: The online version contains supplementary material available at 10.1007/s41315-025-00429-3.

Autonomous vehicles (AVs) and unmanned aerial vehicles (UAVs) have brought about transformative changes in transportation and aviation. However, making these systems fully autonomous and able to navigate safely in unpredictable real-world situations remains a big challenge. Fuzzy logic and related mathematical algorithms have emerged as practical tools to tackle uncertainty and complex decision-making in these systems. This paper reviews how fuzzy logic and mathematical approaches are applied in areas like navigation, control, avoiding obstacles, planning routes, and decision-making for AVs and UAVs. It delves into the key methods, designs, pros, and cons of using fuzzy logic in autonomous vehicles. The paper also compares fuzzy logic with other AI techniques. The review shows that fuzzy logic manages the uncertainties and imprecision involved in how autonomous vehicles perceive and navigate dynamic environments. Fuzzy controllers often perform better than traditional methods in vehicle control and UAV direction control. High-level decisions and route planning in AVs have also benefited from fuzzy inference systems. Still, challenges like computational efficiency, adaptability, and integrating fuzzy logic with other AI components remain. The paper concludes with suggestions for future research to make autonomous vehicles and drones smarter and safer using fuzzy logic. This review is a useful guide for anyone developing intelligent autonomous systems.

As a result of the occupational migration, there is an acute shortage of trained laborers for performing tasks such as harvesting, pruning, pesticide spraying etc. Robotic tree climbers can solve the problem of shortage of laborers. Tree climbing robots belong to vertical climbing robots which moves against the gravitational pull of the earth. But the non-uniformity of tree trunk, inclination of tree irregularity and discontinuity, branches in chaotic arrangements etc. make it extremely challenging to develop a robot that climbs a tree and perform tasks. This review paper presents an introduction to the imperative for climbing robots and challenges in design of tree climbing robots. A study on existing tree climbing robots- climbing strategies, generic mechanical designs, requirements for autonomous climbers are also included. This survey found that there is a need of a novel robotic climber design which can offer better stability and faster operation with higher payload capability.

With the increasing adoption of mobile robots for transporting components across several locations in industries, congestion problems appear if the movement of these robots is not correctly planned. This paper introduces a fleet management system where a central agent coordinates, plans, and supervises the fleet, mitigating the risk of deadlocks and addressing issues related to delays, deviations between the planned paths and reality, and delays in communication. The system uses the TEA* graph-based path planning algorithm to plan the paths of each agent. In conjunction with the TEA* algorithm, the concepts of supervision and graph-based environment representation are introduced. The system is based on ROS framework and allows each robot to maintain its autonomy, particularly in control and localization, while aligning its path with the plan from the central agent. The effectiveness of the proposed fleet manager is demonstrated in a real scenario where robots operate on a shop floor, showing its successful implementation.

Artificial Intelligence is revolutionizing industries by enhancing efficiency through real-time Object Detection (OD) applications. Utilizing advanced computer vision techniques, OD systems automate processes, analyze complex visual data, and facilitate data-driven decisions, thus increasing productivity. Domain Adaptation for OD has recently gained prominence for its ability to recognize target objects without annotations. Innovative approaches that merge traditional cross-disciplinary domain modeling with cutting-edge deep learning have become essential in addressing complex AI challenges in real-time scenarios. Unlike traditional methods, this study proposes a novel, effective Cross-Pollination of Knowledge (CPK) strategy for domain adaptation inspired by botanical processes. The CPK approach involves merging target samples with source samples at the input stage. By incorporating a random and unique selection of a few target samples, the merging process enhances object detection results efficiently in domain adaptation, supporting detectors in aligning and generalizing features with the source domain. Additionally, this work presents the new Planeat digit recognition dataset, which includes 231 images. To ensure robust comparison, we employ a self-supervised Domain Adaptation (UDA) method that simultaneously trains target and source domains using unsupervised techniques. UDA method leverages target data to identify high-confidence regions, which are then cropped and augmented, adapting UDA for effective OD. The proposed CPK approach significantly outperforms existing UDA techniques, improving mean Average Precision (mAP) by 10.9% through rigorous testing on five diverse datasets across different conditions- cross-weather, cross-camera, and synthetic-to-real. Our code is publicly available https://github.com/anwaar0/CPK-Object-Detection

Investigation and subsequent abstraction of biological systems have become a popular approach to innovate new mechanisms. Robot gripper design is an active research area that seeks continuous improvements to cater to its wide spectrum of applications. Tetraonchus momenteron is an infamous parasite which has a robust gripping system (haptor) inserted into the fish-gills. We have redesigned this pull-based gripper for practical application by adjusting the link lengths using many-objective optimization. An existing pusher type of gripper mechanism is also selected to elucidate the similarities and dissimilarities between both the mechanisms. Both push and pull based grippers have been analyzed based on quasi static kinematics and are formulated as many objective optimization problems subjected to geometric-cum-force constraints. The goal of optimization is to find the best link lengths and joint angles for each configuration based on desired criteria. Four possible variants of piezoelectric actuator models are integrated into the optimization formulation to actuate the gripper models. These integrations have led to four different optimization problems for each case. Evolutionary many-objective optimization method is used to obtain non-dominated solutions. The optimized designs are compared based on objective function values to evaluate the performance of each gripper. The results of many-objective optimization are normalized and represented using parallel coordinate plots to aid the decision maker. Our comparative study reveals that while the new nature inspired pull based design has shown promising results in many cases, each gripper configuration has its own merits and requires decision maker’s intervention to choose a particular model based on specific requirement.

The legs must move in a pattern to ensure that a four-legged robot walks organically and uses less energy. This is still a challenging issue today since four-legged creatures with incredibly complicated structures and precise motions are beyond the reach of current technology. This paper proposes a gait generation model for a spider robot that examines the guarantee between stability and speed. First, the robot spider’s movement rules are initiatively determined via four gait parameters—vertical step length, horizontal step length, leg lift, and knee bend. Meanwhile, the 3rd-order interpolation function determines the trajectory of the hips and feet at each leg. By applying analytical methods to solve the inverse kinematics issue, the orbits of the hips and feet at the four legs of the spider robot will be used to deduce twelve joint angle orbits at those locations. Then, a multi-objective function is proposed regarding both speed and stability based on the gait characteristics (gait parameters, CoP/ZMP trajectory) of the spider robot as to train the gait generation model by addressing the forward kinematics issue analytically. Finally, the multi-object MO-Jaya optimization technique is used to find four optimal gait parameters so that the spider robot performs a stable walking gait at the fastest speed. This proposal is implemented for the experiment B3-SBOT spider robot, simulation/experiment outcomes show that B3-SBOT moves at its fastest feasible speed while walking stably.

Drones are technically denoted as unmanned aerial vehicles (UAVs) or unmanned aerial systems (UAS). These are remotely or autonomously piloted aircraft and can fly without an onboard human pilot while being controlled from the ground. The need for UAVs is commonly found in wireless communication beyond human trajectory. Recently, drones have been used for a wide variety of tasks, including military operations, mining, agriculture, and general-purpose applications such as surveillance, exploration, rescue, arms and product delivery, mapping and surveying, entertainment, photography, etc. These UAVs are equipped with a range of Micro-electromechanical Systems (MEMS), including sensors, cameras, controllers, etc. In this manuscript, a novel taxonomy of flying drones with their explicitly defined applications, encompassing UAVs to smart sensors, has been reviewed. The paper also consists of a quick overview of various designing, controlling, and intelligent controller and navigation approaches of UAVs with possible applications. The goal of this study is to offer a review of the best techniques that the researcher can employ or suggest for various UAV-based applications. The paper highlights several challenges that are associated with the modeling, building, controlling, and navigating of UAVs. The article also attempts to suggest some lesser-explored areas that have the potential for future research work, such as battery refilling management, anti-theft devices, sensors to avoid collision, etc.

In recent years, mobile robots have been widely used in industrial automation, logistics, military, medical, and service industries. However, traditional mobile robots face difficulties in complex path planning. To solve this problem, a robot 2D and 3D path planning method has been proposed. Based on improved sparrow search algorithm (SSA), this method introduces a parameter adaptive update strategy to balance search and development. To enhance the algorithm’s path search capability in 3D environments, this study innovatively introduces ant colony algorithm to generate initial paths based on SSA. It also introduces local search mechanism to help the algorithm select the optimal path. Experiments have shown that the improved algorithm can quickly plan the shortest path, with a length reduction of 3.37% and 6.82%. It has reduced the average length by 6.09% and 5.54%, respectively. To verify its search performance, the improved algorithm significantly improved its convergence accuracy, and improved its mean fitness by about 20.5%, 24%, and 27.5%, respectively. The proposed technology has good application effects and will provide important technical references for mobile robots’ path planning in load environments.

Aiming at the problem of defining the shortest distance between poses in traditional virtual fixtures, this paper proposes a proxy-based guidance virtual fixtures with orientation constraints, which can assist operators in the path following task. By designing the dynamics of the proxy and then connecting it to the robot using a spring-damping model, the end-effector position can be constrained within the pipeline and cone, and the stiffness coefficients of the virtual forces can be adjusted linearly or nonlinearly. To overcome the cumulative errors, we then propose a discretized improved algorithm, which applies the virtual fixtures on a discretized reference curve. Furthermore, the orientation constraints are defined, and hence the end-effector orientation can also be constrained within a specific range to comply with ergonomics. An experiment was conducted using a Franka Emika Panda robot and involving 9 subjects. The results show that compared to the gravity compensation mode, the pipeline virtual fixtures and the virtual fixtures with the pipeline and cone, as proposed in this paper, reduced task completion time by 38% and 44.7%, respectively. Additionally, they reduced the total mental burden obtained from NASA-TLX by 29.96% and 47.42%, respectively.