Path planning is a key research issue in the field of unmanned aerial vehicle (UAV) applications. In practical applications, multi-objective path planning is usually required for multi-UAVs, so this paper proposes the improved balanced artificial bee colony (IB-ABC) algorithm to optimize multi-objective path planning. The algorithm adopts the ABC algorithm that emphasizes the global search capability, which is based on iterative feedback information. It uses single-element points, multi-element points, and iteration constraints to optimize the strategies of employed bees, follower bees, and scout bees, respectively. In terms of time and priority, simulation experiments prove that the IB-ABC algorithm can balance local and global search capabilities, accelerate the speed of convergence, and realize multi-UAV path planning in complex mountain environments.

Machines and mechanisms realize processes, from the shaping process of a milling machine and the motion process of an automotive system to the trajectory realization of a robot. The dynamics of a machine generated by a properly chosen set of constraints in combination with an appropriate drive system is designed to meet the prescribed requirements of the process, which is done by projecting the machine equations of motion on the process dynamics. We get a set of nonlinear relations, which represent the machine motion in terms of the required process motion. A well-known example is the projection of arbitrary many robot degrees of freedom (DOFs) on a given path resulting in a set of nonlinear equations with one DOF only, the path coordinate s. Application of this idea can be used to construct a mobility space $(\ddot{s}, \dot{s}, s)$ for any combination of coordinates and constraints. The paper presents a corresponding approach for n-link robots by applying multibody system theory. Method might be interesting for layout of machines and mechanisms. Practical aspects are discussed, and an example is given.

Effective searching for target objects in indoor scenes is essential for household robots to perform daily tasks. With the establishment of a precise map, the robot can navigate to a fixed static target. However, it is difficult for mobile robots to find movable objects like cups. To address this problem, we establish an object search framework that combines navigation map, semantic map, and scene graph. The robot updates the scene graph to achieve a long-term target search. Considering the different start positions of the robots, we weigh the distance the robot walks and the probability of finding objects to achieve global path planning. The robot can continuously update the scene graph in a dynamic environment to memorize the position relation of objects in the scene. This method has been realized in both simulation and real-world environments. The experimental results show the feasibility and effectiveness of this method.

The goal of few-shot semantic segmentation is to learn a segmentation model that can segment novel classes in queries when only a few annotated support examples are available. Due to large intra-class variations, the building of accurate semantic correlation remains a challenging job. Current methods typically use 4D kernels to learn the semantic correlation of feature maps. However, they still face the challenge of reducing the consumption of computation and memory while keeping the availability of correlations mined by their methods. In this paper, we propose the adaptively mining correlation network (AMCNet) to alleviate the aforementioned issues. The key points of AMCNet are the proposed adaptive separable 4D kernel and the learnable pyramid correlation module, which form the basic block for correlation encoder and provide a learnable concatenation operation over pyramid correlation tensors, respectively. Experiments on the PASCAL VOC 2012 dataset show that our AMCNet surpasses the state-of-the-art method by $0.7\%$ and $2.2\%$ on 1-shot and 5-shot segmentation scenarios, respectively.

A model-free adaptive robust control based on time delay estimation (TDE) is proposed for robot in the presence of disturbance and input saturation. TDE is utilized to estimate the complicated nonlinear terms of the robot including unknown dynamics and disturbance, and a TDE error observer is developed to estimate the inevitable TDE error. When the input torque of the robot exceeds the upper or lower limit of the input saturation, an auxiliary system and a saturation deviation boundary adaptive law are employed to mitigate the negative impact of input saturation on the position tracking. Finally, the robust control law is obtained by backstepping. The stability of the closed-loop system is proved by Lyapunov functions, and the validity of the proposed method is demonstrated by comparative simulations and experiments. Compared with the model-based controllers and other model-free controllers, the proposed method does not necessitate the accurate dynamic model of the complicated system and with lower computation. Moreover, it can guarantee the desired position tracking performance of the robot even subject to disturbance and input saturation simultaneously.

Real-time gait trajectory planning is challenging for legged robots walking on unknown terrain. In this paper, to realize a more efficient and faster motion control of a quadrupedal robot, we propose an optimized gait planning generator (GPG) based on the decision tree (DT) and random forest (RF) model of the robot leg workspace. First, the framework of this embedded GPG and some of the modules associated with it are illustrated. Aiming at the leg workspace model described by DT and RF used in GPG, this paper introduces in detail how to collect the original data needed for training the model and puts forward an Interpolation Labeling with Dilation and Erosion (ILDE) data processing algorithm. After the DT and RF models are trained, we preliminarily evaluate their performance. We then present how these models can be used to predict the location relation between a spatial point and the leg workspace based on its distributional features. The DT model takes only 0.00011 s to process a sample, while the RF model can give the prediction probability. As a complement, the PID inverse kinematic model used in GPG is also mentioned. Finally, the optimized GPG is tested during a real-time single-leg trajectory planning experiment and an unknown terrain recognition simulation of a virtual quadrupedal robot. According to the test results, the GPG shows a remarkable rapidity for processing large-scale data in the gait trajectory planning tasks, and the results can prove it has an application value for quadruped robot control.

The study presents an adaptive robust control method for the Pendubot subjects to matched and mismatched uncertainty. First, the control task is formatted as a reduced-dimension equality constraint of the system states. To handle the matched and mismatched uncertainties, an orthogonal decomposition method is employed to make the mismatched part disappear after decomposition. Based on the above, an adaptive robust control law based on constraint-following is devised. By the Lyapunov approach, it is rigorously proven that the proposed approach ensures the uniform boundedness and uniform ultimate boundedness of the closed-loop control system and thus renders approximate constraint-following, regardless of uncertainty. Simulation and experimental results are provided and discussed, demonstrating the good performance of the proposed approach.

Sleep apnea is one of the most common sleep disorders. The consequences of undiagnosed sleep apnea can be very serious, increasing the risk of high blood pressure, heart disease, stroke, and Alzheimer’s disease over a long period of time. However, many people are often unaware of their condition. The gold standard for diagnosing sleep apnea is nighttime polysomnography monitoring in a specialized sleep laboratory. However, these diagnoses are expensive and the number of beds is limited, and there is insufficient monitoring in terms of time dimension. Existing methods for automated detection use no more than three physiological signals, but all other signals are also associated with the patient’s sleep. In addition, the limited amount of medical real annotation data, especially abnormal samples, lead to weak model generalization capability. The gap between model generalization capability and medical field needs still exists. In this paper, we propose a method for integrating medical interpretation rules into a long short-term memory neural network based on self-attention with multichannel respiratory signals as input. We obtain attention weights through a token-level attention mechanism and then extract key rules of medical interpretation to assist the weights, improving model generalization and reducing the dependence on data volume. Compared with the best prediction performance of existing methods, the average improvements of our method in accuracy, precision, and f1-score are 3.26%, 7.03%, and 1.78%, respectively. The algorithm tested the performance of our model on the Sleep Heart Health Study data set and found that the model outperformed existing methods and could help physicians make decisions in their practices.

Collective decision-making by a swarm of robots is of paramount importance. In particular, the problem of collective perception wherein a swarm of robots aims to achieve consensus on the prevalent feature in the environment. Recently, this problem has been formulated as a discrete collective estimation scenario to estimate their proportion rather than deciding about the prevalent one. Nevertheless, the performance of the existing strategies to resolve this scenario is either poor or depends on higher communication bandwidth. In this work, we propose a novel decision-making strategy based on maximum likelihood estimate sharing (MLES) to resolve the discrete collective estimation scenario. Experimentally, we compare the tradeoff speed versus accuracy of MLES with state-of-the-art methods in the literature, such as direct comparison (DC) and distributed Bayesian belief sharing (DBBS). Interestingly, MLES achieves an accurate consensus nearly 20% faster than DBBS, its communication bandwidth requirement is the same as DC but six times less than DBBS, and its computational complexity is $O(1)$. Furthermore, we investigate how noisy sensors affect the effectiveness of the strategies under consideration, with MLES showing better sustainability.

Dielectric elastomers (DEs) find applications in many areas, particularly in the field of soft robotics. When modeling and simulating DE-based actuators and sensors, a substantial portion of the literature assumes the selected DE material to behave in some perfectly hyperelastic manner, and the vast majority have assumed invariant permittivity. However, studies on simple planar DEs have revealed instabilities and hastened breakdowns when a variable permittivity is allowed. This is partly due to the intertwined electromechanical properties of DEs rooted on their labyrinthine polymeric microstructures. This work focuses on studying the effects of a varying (with stretch) permittivity on the out-of-plane deformation of a circular DE, using a model derived from principles of strain-induced polymer birefringence. In addition, we utilize the Edward–Vilgis model, which attempts to account for effects related to crosslinking, and length extension, slippage, and entanglement of polymer chains. Our approach reveals the presence of “stagnation” regions in the electromechanical behavior of the DE actuator material. These stagnation regions are characterized by both electrical and mechanical critical electrostrictive coefficient ratios. Mechanically, certain values of the electrostrictive coefficient ratio predict cases where deformation does not occur in response to a change in voltage. Electrically, certain cases are predicted where changes in capacitance cannot be measured in response to changes in deformation. Thus, some combined conditions of loading and material properties could limit the effectiveness of DE membranes in either actuation or sensing. Therefore, our results reveal mechanisms that could be useful to designers of actuators and sensors and unveil an opportunity for exploring new theoretical materials with potential novel applications. Furthermore, since there are known analogous formulations between electrical and optical properties, criticality principles studied in this article could be extended to optomechanical coupling.

The assist-as-needed (AAN) controller is effective in robot-assisted rehabilitation. However, variations of the engagement of subjects with fixed controller often lead to unsatisfying results. Therefore, adaptive AAN that adjusts control parameters based on individualized engagement is essential to enhance the training effect further. Nevertheless, current approaches mainly focus on the within-trial real-time engagement estimation, and the presence of measurement noise may cause improper evaluation of engagement. In addition, most studies on human-in-loop optimization strategies modulate the controller by greedy strategies, which are prone to fall into local optima. These shortcomings in previous studies could significantly limit the efficacy of AAN. This paper proposes an adaptive AAN to promote engagement by providing subjects with a subject-adaptive assistance level based on trial-wise engagement estimation and performance. Firstly, the engagement is estimated from energy information, which assesses the work done by the subject during a full trial to reduce the influence of measurement outliers. Secondly, the AAN controller is adapted by Bayesian optimization (BO) to maximize the subject’s performance according to historical trial-wise performance. The BO algorithm is good at dealing with noisy signals within limited steps. Experiments with ten healthy subjects resulted in a decrease of 34.59$\%$ in their average trajectory error, accompanied by a reduction of 9.71$\%$ in their energy consumption, thus verifying the superiority of the proposed method to prior work. These results suggest that the proposed method could potentially improve the effect of upper limb rehabilitation.

The control of industrial robot manipulators presents a difficult problem for control engineers due to the complexity of their nonlinear dynamics models. Nonlinear controls based on feedback linearization are developed to meet control requirements. Model-based nonlinear control is highly sensitive to parameter errors and leads to problems of robustness for tracking trajectories at high speeds, and there is the additional problem of a heavy computational burden to consider in the design of nonlinear controllers. In this paper, a mechatronic design approach is proposed, which aims to facilitate controller design by redesigning the mechanical structure. The problem is approached in two steps: first, the dynamic decoupling conditions of manipulators are described and discussed, involving redistribution of the moving mass, which leads to the decoupling of motion equations. A classical linear control law is then used to track the desired efficient bang-bang profile trajectory. Then, in the presence of parameter uncertainty and external disturbances, the nonlinear controls with simple structures are adopted to stabilize the decoupled system asymptotically. An analysis of the results from a simulation of this approach demonstrates its effectiveness in controller design. The proposed improvement in control performance is illustrated via two spatial manipulators.

This paper addresses the flight path planning problem for multiple missiles engaging stationary targets in high-risk areas. Targets protected by air defence are preferably engaged by a fleet or swarm of missiles, not individual missiles. The concept of a swarm attack is that a large number of approaching missiles overwhelm air defence. The deployment of missiles is often part of a broader mission including further participants. Flight path planning is then an integral element of mission planning, requiring strict timing coordination of all members involved. The flight times of the missiles are dictated by the master planning. We present algorithms for offline planning and online re-planning of flight paths for a fleet of missiles with flight time constraints. The algorithms are based on an advanced bidirectional RRT* algorithm that generates risk-minimizing flight paths with predefined flight times. Online planning generates the flight paths of the fleet sequentially, maintaining a safety distance between the missiles to prevent mutual collision. Offline planning uses a global optimization approach to determine an optimal selection of flight paths from a large set of potential paths. The selection is performed by a branch and bound algorithm that determines optimal cliques in the path compatibility graph. The optimization is embedded in an iterative algorithm that allows to successively improve the mission success.

Variational Bayesian inference is an important machine learning tool that finds application from statistics to robotics. The goal is to find an approximate probability density function (PDF) from a chosen family that is in some sense “closest” to the full Bayesian posterior. Closeness is typically defined through the selection of an appropriate loss functional such as the Kullback-Leibler (KL) divergence. In this paper, we explore a new formulation of variational inference by exploiting the fact that (most) PDFs are members of a Bayesian Hilbert space under careful definitions of vector addition, scalar multiplication, and an inner product. We show that, under the right conditions, variational inference based on KL divergence can amount to iterative projection, in the Euclidean sense, of the Bayesian posterior onto a subspace corresponding to the selected approximation family. We work through the details of this general framework for the specific case of the Gaussian approximation family and show the equivalence to another Gaussian variational inference approach. We furthermore discuss the implications for systems that exhibit sparsity, which is handled naturally in Bayesian space, and give an example of a high-dimensional robotic state estimation problem that can be handled as a result. We provide some preliminary examples of how the approach could be applied to non-Gaussian inference and discuss the limitations of the approach in detail to encourage follow-on work along these lines.

This paper discusses the design and the experimental tests on Porcospino, a bio-inspired single-track mobile robot for inspection of unstructured environments characterized by narrow spaces. It is an evolution of SnakeTrack, a single-track robot with steering capabilities; differently from SnakeTrack, the track modules of Porcospino are characterized by elastic spines, which improve traction on uneven and irregular terrains. The main body is a vertebral column, comprising a series of vertebrae connected by compliant joints and two end modules. Each end module carries two sprockets, sharing a common actuator, to drive the single peripherical track. Moreover, each end module hosts an actuator for track steering. The remaining mobilities of the vertebral column allow it to cope passively with the terrain profile, to enhance traction. The control unit, batteries, drivers, and environmental sensors are placed along the vertebral column. Both the end modules are equipped with a camera for intermittent vision, which is possible thanks to openings realized on the track modules. The experimental campaign on the first Porcospino prototype is discussed, highlighting the differences with its earlier version.

With the dangerous and troublesome nature of hollow defects inside building structures, hollowness inspection has always been a challenge in the field of construction quality assessment. Several methods have been proposed for inspecting hollowness inside concrete structures. These methods have shown great advantages compared to manual inspection but still lack autonomy and have several limitations. In this paper, we propose a range-point migration-based non-contact hollowness inspection system with sensor fusion of ultra-wide-band radar and laser-based depth camera to extract both outer surface and inner hollowness information accurately and efficiently. The simulation result evaluates the performance of the system based on the original range-point migration algorithm, and our proposed one and the result of our system show great competitiveness. Several simulation experiments of structures that are very common in reality are carried out to draw more convincing conclusions about the system. At the same time, a set of laboratory-made concrete components were used as experimental objects for the robotic system. Although still accompanied by some problems, these experiments demonstrate the availability of an automated hollow-core detection system.

Flexibility is one of the most significant advantages of legged robots in unstructured environments. However, quadruped robots cannot interact with environments to complete some manipulation tasks. One effective way is to load a manipulation arm. In this paper, we exhibit a quadruped locomotion manipulation system (LMS) named HITPhanT. This system comprises a quadruped locomotion platform and a six-degree-of-freedom manipulation arm. Besides, when the LMS moves to a designated position for operation, it is necessary to constrain the foot contact points to avoid sliding. Therefore, the foot contact point is regarded as a spherical hinge. So the locomotion platform can be considered as a parallel mechanism. A hybrid kinematics model is established by considering the serial robotic arms connecting this parallel mechanism. Besides, the trajectory planning method, which improves the system’s manipulability in evaluating the system balance, is also proposed. Finally, corresponding experiments verify the overall system’s stabilization and algorithm’s effectiveness.

Bending and elongation have been some of the most studied motions in soft actuators due to the variety of their applications. For that matter, multi-DOF actuators have been developed with the purpose to generate different movements in a single actuator, mainly bending.

However, these actuators are still limited in mobility range, and some of them do not perform continuous curvatures. This paper presents the design, characterisation and implementations of a multi-DOF soft pneumatic module. The internal structure of the proposed module is composed of four channels, which generate bending in several directions. The finite element method analysis demonstrates that the actuator performs continuous curvatures for different pressure values. We present a repeatable and easy manufacturing process using the casting technique, considering the material Ecoflex 00-50; as well as the kinematic model of the actuator, taking into consideration two bending Degrees of Freedom (DOFs). Furthermore, we performed bending characterisation for all possible combinations of the four channels via computer vision, demonstrating a wide mobility range and performing continuous curvatures. Additionally, we evaluated the kinematic model with characterisation data, obtaining the angular and cartesian relationship between the pressure and continuous curvatures. On the other hand, the authors propose the design of a modular soft manipulator based on two multi-DOF modules. The kinematic model is reported. In addition, we implement a motion sequence in the manipulator to pick and place tasks.

In laparoscopic surgery, surgeons should develop several manual laparoscopic skills before carrying out real operative procedures using a low-cost box trainer. The Fundamentals of Laparoscopic Surgery (FLS) program was developed as a program to assess fundamental knowledge and surgical skills, required for basic laparoscopic surgery. The peg transfer task is a hands-on exam in the FLS program that assists a trainee to understand the relative minimum amount of grasping force necessary to move the pegs from one place to another place without dropping them. In this paper, an autonomous, sequential assessment algorithm based on deep learning, a multi-object detection method, and, several sequential If-Then conditional statements have been developed to monitor each step of a surgeon’s performance. Images from three different cameras are used to assess whether the surgeon executes the peg transfer task correctly and to display a notification on any errors on the monitor immediately. This algorithm improves the performance of a laparoscopic box-trainer system using top, side, and front cameras and removes the need for any human monitoring during a peg transfer task. The developed algorithm can detect each object and its status during a peg transfer task and notifies the resident about the correct or failed outcome. In addition, this system can correctly determine the peg transfer execution time, and the move, carry, and dropped states for each object by the top, side, and front-mounted cameras. Based on the experimental results, the proposed surgical skill assessment system can identify each object at a high score of fidelity, and the train-validation total loss for the single-shot detector (SSD) ResNet50 v1 was about 0.05. Also, the mean average precision (mAP) and Intersection over Union (IoU) of this detection system were 0.741, and 0.75, respectively. This project is a collaborative research effort between the Department of Electrical and Computer Engineering and the Department of Surgery, at Western Michigan University.

During the operation of automatic navigation rice transplanter, the accuracy of path tracking is influenced by whether the transplanter can enter the stable state of linear path tracking quickly, thus affecting the operation quality and efficiency. To reduce the time to enter the path tracking stable state and improve the tracking accuracy and stability for the rice transplanter, path tracking control method based on variable universe fuzzy control (VUFC) and improved beetle antenna search (BAS) is proposed in this paper. VUFC is applied to achieve adaptive adjustment of the fuzzy universe by dynamically adjusting the quantization and scaling factors according to the variations of errors by the contraction–expansion factor. To solve the problem of setting the contraction–expansion factor in VUFC and real-time performance, an offline parameter optimization method is presented to calculate the optimal contraction–expansion factor by an iterative optimization algorithm in a path tracking simulation model, where the iterative optimization algorithm is the BAS algorithm improved by the isolated niching technique and adaptive step size strategy in this paper. To verify the effectiveness of the proposed path tracking control method, simulation and field linear path tracking experiments were carried out. Experimental results indicate that the proposed method reduces the time of entering the stable state of linear path tracking and improves the accuracy and stability of path tracking compared with the pure pursuit control method.