International Journal of Intelligent Robotics and Applications最新文献

The development of a non-destructive and patient-friendly method for examining the intestines is crucial for early prevention and timely diagnosis of prevalent intestinal diseases that pose a threat to human health worldwide. Although the soft robot shows promise as an examination method due to its safe human-machine interaction and high maneuverability, achieving controlled and non-damaging movements within the flexible and delicate structure of the intestines remains a significant challenge. In this study, we propose and design a leech-inspired soft robot capable of operating in an intestine-like environment while ensuring lossless and controllable functionality. The soft robot consists of two dual-chambered adsorption actuators serving as “feet” and a retractable actuator as the body, enabling the robot to crawl by programmatically controlling the alternating movements of the adsorption actuators and the cooperation of the retractable actuator. Through numerical simulations, and movement tests in various scenarios such as planes, slopes, and intestine-like pipelines, we verified the adsorption characteristics and regulation mechanism of the adsorption actuator, as well as the movement performance of the robot. The results demonstrate that the adsorption actuator achieves a maximum adsorption force of 3.17 N, and the soft robot attains a maximum moving speed of 9.29 mm/s. This research offers a non-destructive and patient-friendly approach that holds promise for the detection and treatment of intestinal diseases in practical applications.

In recent years, Unmanned aerial vehicles (UAVs) have witnessed a surge in popularity and implementation for both civilian and military usage. UAVs can be utilized for a wide range of applications, including mapping, surveillance, and inspection. For many of these applications, a high level of autonomy is required. Autonomy refers to the ability to complete missions or tasks without human intervention. Autonomous navigation is an essential element of autonomy, especially in GPS-denied environments where GNSS-based navigation is not reliable. Due to size and weight limitations, many UAVs employ vision-based localization and navigation techniques for GPS-denied environments. Reinforcement Learning (RL) is also increasingly being implemented for robotic applications, including obstacle avoidance, battery management, and navigation. Existing reviews typically focus on either vision-based autonomous navigation of drones or RL navigation for drones in general, but none specifically concentrate on the use of vision-based methods and RL for drone navigation. Moreover, previous reviews have highlighted the use of reinforcement learning based on tasks such as takeoff, landing, and navigation, whereas this review categorizes the use of RL based on the navigation problem and image input types for the RL models as these define the needed hardware and processing capabilities of the system. We define the current challenges and limitations for vision based RL navigation to provide direction for future works. Finally we provide an analysis of the favorable conditions for each category and the possibility of combining multiple categories to overcome the disadvantages of each.

Researchers have recently focused on studying the flight dynamics and control of multicopters and fixed-wing aerial vehicles. However, investigating the transition phase between multicopter hover and fixed-wing cruise modes for a Dual-thrust Aerial Vehicle (DAV) is still challenging. In this paper, we develop two sets of nonlinear equations of motion for a DAV to create a multi-purpose dynamic model for designing control and transition mode scenarios. The first set considers the multicopter torque as the control input, while the second set considers the elevator torque as the control input. By analyzing three transition scenarios between multicopter hover and fixed-wing cruise flights, we observe that the best performance occurs for the third scenario in which the control system switches from multicopter control torque to elevator control torque when the multicopter thrust equals the wings’ lift. In this case, the vehicle will be protected from critical flight conditions like wing stalls while the transition will go smoothly with minimum height drop. The transition mode strategies are implemented using a model predictive controller in flight simulation. The numerical results show the dynamic behavior of the DAV in different transition scenarios from hover to cruise and vice versa, demonstrating successful altitude control and stable transitions in both phases.

The teleoperation robot system (TRS) stands as a prominent research frontier within robot control, amalgamating human decision-making capacity with robot operation, thus markedly enhancing safety and precision compared to autonomous operation. This paper selects TRS hardware and designs master–slave interaction software comprising six distinct modules tailored to diverse functionalities. It further derives forward and backward kinematic equations based on master–slave device linkage parameters, proposing a Cartesian workspace-based master–slave mapping algorithm. Additionally, a human–robot interaction (HRI) control framework emphasizing direct force feedback is devised to bolster system HRI performance and operator immersion. To ensure smooth, safe, and agile slave device movement, an innovative impedance controller-based TRS force feedback HRI control framework is introduced. The effectiveness of the TRS HRI control framework is validated via comprehensive experiments conducted across multiple scenarios, including remote robot axle-hole assembly, blackboard erasing, text writing, and auxiliary welding operations, on a constructed experimental platform for robot remote operation system HRIs.

This paper focuses on the implementation of the Extended Kalman Filter for indoor localization of a semi-autonomous Ambulance Robot system named Ambubot. The system is designed to reduce the response time for lay rescuers to locate an Automated External Defibrillator (AED) during sudden cardiac arrest events. To achieve this objective, the robot is equipped with an AED, and the Extended Kalman Filter is utilized for optimal indoor localization. The filter is implemented using data from the robot’s Inertial Measurement Unit, which comprises 9 Degrees of Freedom. The paper provides an explicit description of the performance of the Extended Kalman Filter in estimating the position of Ambubot, and demonstrates that the proposed approach is effective in accurately determining and estimating the robot’s position in unknown indoor environments. The results suggest that the proposed method is a promising solution for improving survival rates in cardiac arrest cases, and may have potential applications in other fields where accurate indoor localization is required.

With the breakthroughs in machine learning and computing infrastructures that have led to significant performance improvements in cognitive robotics, the challenge of continuous-trajectory task creation persists. This challenge stems from the need to account for inter-joint relationships, which define constraints between different robot joints due to the kinematic structure, and intra-joint relationships, which are constraints within a single joint like limits. Accounting for these coupled, nonlinear inter-joint and intra-joint relationships is crucial for trajectory planning. However, various constraints in the physical capability of robots, environmental changes, and long-time reliance on sequential dependencies between these inter-joint and intra-joint relationships make the work of modifying robot trajectories exceptionally hard. Many robot environments function under structured static work-cell completing extended series of subtasks. The conventional descriptors for robot trajectory rely on symbolic rules with human intelligence, which involves skilled individuals and possesses significant limitations, such as being time-consuming and exhibiting low flexibility even for minor changes, due to the static nature of task descriptions alone. The suggested technique employs a probabilistic network and data-efficient modelling termed generative adversarial networks, which learns the underlying constraints, probability distributions and arbitrations, along with generating trajectory instances at each time of sampling. Integrating prior knowledge into the symbolic trajectory learner as a dataset facilitates the learning process. The model assessment was carried out by utilising a custom-built dataset in a simulation based environment. This research also proposed two GAN inversion methods to compute the generated trajectory and its closest instance in the dataset. Furthermore, GAN Inversion method I and II calculated the robot path accuracy in extrinsic generative models yielded path position accuracy of 9.2 cm and 4.9 cm respectively. In addition to that, the study contributes a probabilistic model for interpolating between various trajectories to generate new trajectories.

Many robotic systems face substantial challenges when trying to grasp and manipulate objects. Thought of initially as humanoid automata a century ago, this viewpoint is still influential in modern robot design. Many robotic grippers are inspired by the deftness of the human hand. The perceptual, processing, and control issues that conventional grippers have are also experienced by soft-fingered grippers. Precise finger placement, dictated by the shape and attitude of the object, is necessary to accomplish force closure when using soft fingertips to grasp. Soft robotic end-effectors have several advantages, such as a good interface with humans, the capacity to adapt to different environments, a number of degrees of freedom, and the ability to non-destructively grasp items of various shapes. Adding to earlier research that looked at the soft robot in a theoretical way, this study creates an optimized model based on the deformation in terms of bending of the channel cavity under applied pneumatic pressure. A correlation between pneumatic pressure and the pneumatic soft actuator's bending angle has been demonstrated. This research looks at how different design factors affect the bending of a multi-chambered soft actuator that is pneumatically networked. The finite element approach involves fine-tuned (optimised) actuator construction. Using FEM to evaluate aspects affecting actuator mechanical output, the ideal design parameters were discovered using DoE, resulting in a bending angle of ~ 104 degrees at 30 kPa. This study used ANOVA at a 5% significant level to identify which variables most affected the pneumatic actuator's deformation (bending angle). The significant R-square value of 96.42% supports the study's conclusions that the parameters utilised explain an immense percentage of bending angle deviations. Experimental verification of the optimized finite element model findings was conducted. The verification of the actuators' bending angles and output forces reveals that the discrepancy between the two sets of data stayed below 9%. Also, the average gripping success rate attained in the grasping evaluation, which involved four distinct types of items, was almost 97%.

This study explores the novel concept of Multi-Object Grasping (MOG) and develops an architecture based on autoencoders and transformers for accurate object prediction in MOG scenarios. The approach employs different deep learning methods and diverse training approaches using the ping pong ball dataset. The parameters obtained from this training enhance the model's performance on the actual system dataset, serving as the final test and validation of the model's reliability in real-world situations. Comparing the model's performance on both datasets facilitates validation and refinement, affirming its effectiveness in practical robotic applications. The study highlights that training various dataset features significantly improves prediction accuracy compared to the Naïve model using dense neural networks. Using five-time steps notably enhances prediction accuracy, especially with the GRU model in time-series data architecture, achieving a peak accuracy of 96%. While MOG has been extensively studied, this study introduces a novel architecture distinct from traditional visual methods. A framework is established that utilizes autoencoder and transformer technologies for managing tactile sensors, hand pose joint angles and force measurements. This approach demonstrates the potential for accurately predicting multiple objects in MOG scenarios.

Accurate localization is essential for enabling intelligent autonomous navigation in indoor environments. While global navigation satellite systems (GNSS) provide efficient outdoor solutions, applications in indoor environments require alternative approaches to determine the vehicle's global position. This study investigates a ROS-based multi-sensor integrated localization system utilizing wheel odometry, inertial measurement unit (IMU), and 2D light detection and ranging (LiDAR) based simultaneous localization and mapping (SLAM) for cost-effective and accurate indoor autonomous vehicle (AV) navigation. The paper analyzes the limitations of wheel odometry and IMU, highlighting their susceptibility to errors. To address these limitations, the proposed system leverages LiDAR SLAM for real-time map generation and pose correction. The Karto SLAM package from robot operating system (ROS) is chosen due to its superior performance according to the literature. Results indicate that the integration of these technologies reduces localization errors significantly, with the system achieving a high degree of accuracy in pose estimation under various test conditions. The experimental validation shows that the proposed system maintains consistent performance, proving its potential for widespread application in environments where GNSS is unavailable.

String insulators are components in high-voltage towers responsible for preventing energy dissipation through the tower structure; that is, they are responsible for isolating the high voltage in the electrical network cables. These string insulators must be clean for best performance and to avoid malfunctions. Verifying the necessity for cleaning/washing is most often performed by human visual observation, which can lead to interpretation errors, in addition to bringing risks to the physical integrity of humans in the vicinity of these electrical systems. Thus, this paper aims to develop an algorithm to detect and classify these insulators. The proposed algorithm uses artificial intelligence techniques and analyzes the image, inferring the state of cleanliness of the analyzed insulator. For the development of this algorithm, it was necessary to build a synthetic database using CAD software such as Inventor and Unity-3D due to image limitations available from dirty insulator strings. In this paper, two distinct neural networks are built using supervised learning techniques, where the first one is for detecting the chain of insulators, and the second is for detecting the type of dirt on the disk surface. In the first stage, techniques that use supervised learning are studied, more aimed explicitly at semantic segmentation networks, and in the second stage, classification deep neural networks were used to detect the type of impurities. In detecting insulator strings, an average dice coefficient of 0.95 was achieved for simulated images and 0.92 for natural images, with learning parameters based on a database with only simulated images. The average accuracy obtained in the dirt classification stage was 0.98.