Search results

Because of the various advantages of reinforcement learning (RL) mentioned above, this study uses RL to train unmanned aerial vehicles to perform two tasks: target search and cooperative obstacle avoidance.

This study draws inspiration from the recurrent state-space model and recurrent models (RPM) to propose a simpler yet highly effective model called the unmanned aerial vehicles prediction model (UAVPM). The main objective is to assist in training the UAV representation model with a recurrent neural network, using the soft actor-critic algorithm.

This study proposes a generalized actor-critic framework consisting of three modules: representation, policy and value. This architecture serves as the foundation for training UAVPM. This study proposes the UAVPM, which is designed to aid in training the recurrent representation using the transition model, reward recovery model and observation recovery model. Unlike traditional approaches reliant solely on reward signals, RPM incorporates temporal information. In addition, it allows the inclusion of extra knowledge or information from virtual training environments. This study designs UAV target search and UAV cooperative obstacle avoidance tasks. The algorithm outperforms baselines in these two environments.

It is important to note that UAVPM does not play a role in the inference phase. This means that the representation model and policy remain independent of UAVPM. Consequently, this study can introduce additional “cheating” information from virtual training environments to guide the UAV representation without concerns about its real-world existence. By leveraging historical information more effectively, this study enhances UAVs’ decision-making abilities, thus improving the performance of both tasks at hand.

Limited by the types of sensors, the state information available for musculoskeletal robots with highly redundant, nonlinear muscles is often incomplete, which makes the control face a bottleneck problem. The aim of this paper is to design a method to improve the motion performance of musculoskeletal robots in partially observable scenarios, and to leverage the ontology knowledge to enhance the algorithm’s adaptability to musculoskeletal robots that have undergone changes.

A memory and attention-based reinforcement learning method is proposed for musculoskeletal robots with prior knowledge of muscle synergies. First, to deal with partially observed states available to musculoskeletal robots, a memory and attention-based network architecture is proposed for inferring more sufficient and intrinsic states. Second, inspired by muscle synergy hypothesis in neuroscience, prior knowledge of a musculoskeletal robot’s muscle synergies is embedded in network structure and reward shaping.

Based on systematic validation, it is found that the proposed method demonstrates superiority over the traditional twin delayed deep deterministic policy gradients (TD3) algorithm. A musculoskeletal robot with highly redundant, nonlinear muscles is adopted to implement goal-directed tasks. In the case of 21-dimensional states, the learning efficiency and accuracy are significantly improved compared with the traditional TD3 algorithm; in the case of 13-dimensional states without velocities and information from the end effector, the traditional TD3 is unable to complete the reaching tasks, while the proposed method breaks through this bottleneck problem.

In this paper, a novel memory and attention-based reinforcement learning method with prior knowledge of muscle synergies is proposed for musculoskeletal robots to deal with partially observable scenarios. Compared with the existing methods, the proposed method effectively improves the performance. Furthermore, this paper promotes the fusion of neuroscience and robotics.

This study aims to investigate the morphing concepts able to manipulate the dynamics of the downstream unsteadiness in the separated shear layers and, in the wake, be able to modify the upstream shock–boundary layer interaction (SBLI) around an A320 morphing prototype to control these instabilities, with emphasis to the attenuation or even suppression of the transonic buffet. The modification of the aerodynamic performances according to a large parametric study carried out at Reynolds number of 4.5 × 10⁶, Mach number of 0.78 and various angles of attack in the range of (0, 2.4)° according to two morphing concepts (travelling waves and trailing edge vibration) are discussed, and the final benefits in aerodynamic performance increase are evaluated.

This article examines through high fidelity (Hi-Fi) numerical simulation the effects of the trailing edge (TE) actuation and of travelling waves along a specific area of the suction side starting from practically the most downstream position of the shock wave motion according to the buffet and extending up to nearly the TE. The present paper studies through spectral analysis the coherent structures development in the near wake and the comparison of the aerodynamic forces to the non-actuated case. Thus, the physical mechanisms of the morphing leading to the increase of the lift-to-drag ratio and the drag and noise sources reduction are identified.

This study investigates the influence of shear-layer and near-wake vortices on the SBLI around an A320 aerofoil and attenuation of the related instabilities thanks to novel morphing: travelling waves generated along the suction side and trailing-edge vibration. A drag reduction of 14% and a lift-to-drag increase in the order of 8% are obtained. The morphing has shown a lift increase in the range of (1.8, 2.5)% for angle of attack of 1.8° and 2.4°, where a significant lift increase of 7.7% is obtained for the angle of incidence of 0° with a drag reduction of 3.66% yielding an aerodynamic efficiency of 11.8%.

This paper presents results of morphing A320 aerofoil, with a chord of 70cm and subjected to two actuation kinds, original in the state of the art at M = 0.78 and Re = 4.5 million. These Hi-Fi simulations are rather rare; a majority of existing ones concern smaller dimensions. This study showed for the first time a modified buffet mode, displaying periodic high-lift “plateaus” interspersed by shorter lift-decrease intervals. Through trailing-edge vibration, this pattern is modified towards a sinusoidal-like buffet, with a considerable amplitude decrease. Lock-in of buffet frequency to the actuation is obtained, leading to this amplitude reduction and a drastic aerodynamic performance increase.

The purpose of this study is to propose a precise and standardized strategy for numerically simulating vehicle aerodynamics.

Error sources in computational fluid dynamics were analyzed. Additionally, controllable experiential and discretization errors, which significantly influence the calculated results, are expounded upon. Considering the airflow mechanism around a vehicle, the computational efficiency and accuracy of each solution strategy were compared and analyzed through numerous computational cases. Finally, the most suitable numerical strategy, including the turbulence model, simplified vehicle model, calculation domain, boundary conditions, grids and discretization scheme, was identified. Two simplified vehicle models were introduced, and relevant wind tunnel tests were performed to validate the selected strategy.

Errors in vehicle computational aerodynamics mainly stem from the unreasonable simplification of the vehicle model, calculation domain, definite solution conditions, grid strategy and discretization schemes. Using the proposed standardized numerical strategy, the simulated steady and transient aerodynamic characteristics agreed well with the experimental results.

Building upon the modified Low-Reynolds Number k-e model and Scale Adaptive Simulation model, to the best of the authors’ knowledge, a precise and standardized numerical simulation strategy for vehicle aerodynamics is proposed for the first time, which can be integrated into vehicle research and design.

Community detection of dynamic networks provides more effective information than static network community detection in the real world. The mainstream method for community detection in dynamic networks is evolutionary clustering, which uses temporal smoothness of community structures to connect snapshots of networks in adjacent time intervals. However, the error accumulation issues limit the effectiveness of evolutionary clustering. While the multi-objective evolutionary approach can solve the issue of fixed settings of the two objective function weight parameters in the evolutionary clustering framework, the traditional multi-objective evolutionary approach lacks self-adaptability.

This paper proposes a community detection algorithm that integrates evolutionary clustering and decomposition-based multi-objective optimization methods. In this approach, a benchmark correction procedure is added to the evolutionary clustering framework to prevent the division results from drifting.

Experimental results demonstrate the superior accuracy of this method compared to similar algorithms in both real and synthetic dynamic datasets.

To enhance the clustering results, adaptive variances and crossover probabilities are designed based on the relative change amounts of the subproblems decomposed by MOEA/D (A Multiobjective Optimization Evolutionary Algorithm based on Decomposition) to dynamically adjust the focus of different evolutionary stages.

Occupant behavior can lead to considerable uncertainties in thermal comfort and air quality within buildings. To tackle this challenge, the use of probabilistic controls to simulate occupant behavior has emerged as a potential solution. This study seeks to analyze the performance of free-running households by examining adaptive thermal comfort and CO2 concentration, both crucial variables in indoor air quality. The investigation of indoor environment dynamics caused by the occupants' behavior, especially after the COVID-19 pandemic, became increasingly important. Specifically, it investigates 13 distinct window and shading control strategies in courtyard houses to identify the factors that prompt occupants to interact with shading and windows and determine which control approach effectively minimizes the performance gap.

This paper compares commonly used deterministic and probabilistic control functions and their effects on occupant comfort and indoor air quality in four zones surrounding a courtyard. The zones are differentiated by windows facing the courtyard. The study utilizes the energy management system (EMS) functionality of EnergyPlus within an algorithmic interface called Ladybug Tools. By modifying geometrical dimensions, orientation, window-to-wall ratio (WWR) and window operable fraction, a total of 465 cases are analyzed to identify effective control scenarios. According to the literature, these factors were selected because of their potential significant impact on occupants’ thermal comfort and indoor air quality, in addition to the natural ventilation flow rate. Additionally, the Random Forest algorithm is employed to estimate the individual impact of each control scenario on indoor thermal comfort and air quality metrics, including operative temperature and CO2 concentration.

The findings of the study confirmed that both deterministic and probabilistic window control algorithms were effective in reducing thermal discomfort hours, with reductions of 56.7 and 41.1%, respectively. Deterministic shading controls resulted in a reduction of 18.5%. Implementing the window control strategies led to a significant decrease of 87.8% in indoor CO2 concentration. The sensitivity analysis revealed that outdoor temperature exhibited the strongest positive correlation with indoor operative temperature while showing a negative correlation with indoor CO2 concentration. Furthermore, zone orientation and length were identified as the most influential design variables in achieving the desired performance outcomes.

It’s important to acknowledge the limitations of this study. Firstly, the potential impact of air circulation through the central zone was not considered. Secondly, the investigated control scenarios may have different impacts on air-conditioned buildings, especially when considering energy consumption. Thirdly, the study heavily relied on simulation tools and algorithms, which may limit its real-world applicability. The accuracy of the simulations depends on the quality of the input data and the assumptions made in the models. Fourthly, the case study is hypothetical in nature to be able to compare different control scenarios and their implications. Lastly, the comparative analysis was limited to a specific climate, which may restrict the generalizability of the findings in different climates.

Occupant behavior represents a significant source of uncertainty, particularly during the early stages of design. This study aims to offer a comparative analysis of various deterministic and probabilistic control scenarios that are based on occupant behavior. The study evaluates the effectiveness and validity of these proposed control scenarios, providing valuable insights for design decision-making.

The purpose of this paper is to analyze numerically the blowup in finite time of the solutions to a one-dimensional, bidirectional, nonlinear wave model equation for the propagation of small-amplitude waves in shallow water, as a function of the relaxation time, linear and nonlinear drift, power of the nonlinear advection flux, viscosity coefficient, viscous attenuation, and amplitude, smoothness and width of three types of initial conditions.

An implicit, first-order accurate in time, finite difference method valid for semipositive relaxation times has been used to solve the equation in a truncated domain for three different initial conditions, a first-order time derivative initially equal to zero and several constant wave speeds.

The numerical experiments show a very rapid transient from the initial conditions to the formation of a leading propagating wave, whose duration depends strongly on the shape, amplitude and width of the initial data as well as on the coefficients of the bidirectional equation. The blowup times for the triangular conditions have been found to be larger than those for the Gaussian ones, and the latter are larger than those for rectangular conditions, thus indicating that the blowup time decreases as the smoothness of the initial conditions decreases. The blowup time has also been found to decrease as the relaxation time, degree of nonlinearity, linear drift coefficient and amplitude of the initial conditions are increased, and as the width of the initial condition is decreased, but it increases as the viscosity coefficient is increased. No blowup has been observed for relaxation times smaller than one-hundredth, viscosity coefficients larger than ten-thousandths, quadratic and cubic nonlinearities, and initial Gaussian, triangular and rectangular conditions of unity amplitude.

The blowup of a one-dimensional, bidirectional equation that is a model for the propagation of waves in shallow water, longitudinal displacement in homogeneous viscoelastic bars, nerve conduction, nonlinear acoustics and heat transfer in very small devices and/or at very high transfer rates has been determined numerically as a function of the linear and nonlinear drift coefficients, power of the nonlinear drift, viscosity coefficient, viscous attenuation, and amplitude, smoothness and width of the initial conditions for nonzero relaxation times.

To improve the position tracking efficiency of the upper-limb rehabilitation robot for stroke hemiplegia patients, the optimization Learning rate of the membership function based on the fuzzy impedance controller of the rehabilitation robot is propose.

First, the impaired limb’s damping and stiffness parameters for evaluating its physical recovery condition are online estimated by using weighted least squares method based on recursive algorithm. Second, the fuzzy impedance control with the rule has been designed with the optimal impedance parameters. Finally, the membership function learning rate online optimization strategy based on Takagi-Sugeno (TS) fuzzy impedance model was proposed to improve the position tracking speed of fuzzy impedance control.

This method provides a solution for improving the membership function learning rate of the fuzzy impedance controller of the upper limb rehabilitation robot. Compared with traditional TS fuzzy impedance controller in position control, the improved TS fuzzy impedance controller has reduced the overshoot stability time by 0.025 s, and the position error caused by simulating the thrust interference of the impaired limb has been reduced by 8.4%. This fact is verified by simulation and test.

The TS fuzzy impedance controller based on membership function online optimization learning strategy can effectively optimize control parameters and improve the position tracking speed of upper limb rehabilitation robots. This controller improves the auxiliary rehabilitation efficiency of the upper limb rehabilitation robot and ensures the stability of auxiliary rehabilitation training.

Existing Simultaneous Localization and Mapping (SLAM) algorithms have been relatively well developed. However, when in complex dynamic environments, the movement of the dynamic points on the dynamic objects in the image in the mapping can have an impact on the observation of the system, and thus there will be biases and errors in the position estimation and the creation of map points. The aim of this paper is to achieve more accurate accuracy in SLAM algorithms compared to traditional methods through semantic approaches.

In this paper, the semantic segmentation of dynamic objects is realized based on U-Net semantic segmentation network, followed by motion consistency detection through motion detection method to determine whether the segmented objects are moving in the current scene or not, and combined with the motion compensation method to eliminate dynamic points and compensate for the current local image, so as to make the system robust.

Experiments comparing the effect of detecting dynamic points and removing outliers are conducted on a dynamic data set of Technische Universität München, and the results show that the absolute trajectory accuracy of this paper's method is significantly improved compared with ORB-SLAM3 and DS-SLAM.

In this paper, in the semantic segmentation network part, the segmentation mask is combined with the method of dynamic point detection, elimination and compensation, which reduces the influence of dynamic objects, thus effectively improving the accuracy of localization in dynamic environments.

The purpose of this study is to find the suitable trajectory path of the Numerical model of the Quadcopter. Quadcopters are widely used in various applications due to their compact size and ease of assembly. Because they are quite unstable, autonomous control systems would be used to overcome this problem. Modelling autonomous control is predominant as the research scope faces challenges because of its highly non-linear, multivariable system with 6 degree of freedom.

Quadcopters with antonym systems can operate in an unknown environment by overcoming unexpected disturbances. The first objective when designing such a system is to design an accurate mathematical model to describe the dynamics of the system. Newton’s law of motion was used to build the mathematical model of the system.

Establishment of the mathematical model and the physics behind a four propeller drone for the frame TAROT 650 carbon was done. Simulink model was developed based on the mathematical model for simulating the complete dynamics of the drone as well as location and gusts were included to check the stability.

The control response of the system was simulated numerically results are discussed. The trajectory path was found. The phases with their own parameters can be used to implement the mathematical model for another type of quadcopter model and achieve quick development.