Search results

Slim and dexterous manipulators with long reaches can perform various exploration and inspection tasks in confined spaces. This paper aims to present the development of such a dexterous continuum manipulator for potential applications in the aviation industry.

Benefiting from a newly conceived dual continuum mechanism and the improved actuation scheme, this paper proposes a design of a slim and dexterous continuum manipulator. Kinematics modeling, simulation-based dimension synthesis, structural constructions and system descriptions are elaborated.

Experimental validations show that the constructed prototype possesses the desired dexterity to navigate through confined spaces with its kinematics calibrated and actuation compensation implemented. The continuum manipulator with different deployed tools (e.g. graspers and welding guns) would be able to perform inspections and other tasks at remote locations in constrained environments.

The current construction of the continuum manipulator possesses quite some friction inside its structure. The bending discrepancy caused by friction could accumulate to an obvious level. It is desired to further reduce the friction, even though the actuation compensation had been implemented.

The constructed continuum manipulator could perform inspection and other tasks in confined spaces, acting as an active multi-functional endoscopic platform. Such a device could greatly facilitate routine tasks in the aviation industry, such as guided assembling, inspection and maintenance.

The originality and values of this paper mainly lay on the design, modeling, construction and experimental validations of the slim and dexterous continuum manipulator for the desired mobility and functionality in confined spaces.

We describe an intelligent video categorization engine (IVCE) that uses the learning capability of artificial neural networks (ANNs) to classify suitably preprocessed video segments into a predefined number of semantically meaningful events (categories).

We provide a survey of existing techniques that have been proposed, either directly or indirectly, towards achieving intelligent video categorization. We also compare the performance of two popular ANNs: Kohonen's self‐organizing map (SOM) and fuzzy adaptive resonance theory (Fuzzy ART). In particular, the ANNs are trained offline to form the necessary knowledge base prior to online categorization.

Experimental results show that accurate categorization can be achieved near instantaneously.

The main limitation of this research is the need for a finite set of predefined categories. Further research should focus on generalization of such techniques.

Machine understanding of video footage has tremendous potential for three reasons. First, it enables interactive broadcast of video. Second, it allows unequal error protection for different video shots/segments during transmission to make better use of limited channel resources. Third, it provides intuitive indexing and retrieval for video‐on‐demand applications.

The paper aims to advance methodologies to optimize fuzzy logic controller parameters via neural network and use the neuro‐fuzzy scheme to control two cooperating robots.

The paper presents a special neural network architecture that can be converted to fuzzy logic controller. Concepts of model predictive control (MPC) have been used to generate optimal signal to be used to train the neural network via backpropagation. Subsequently, a trained neural network is used to obtain fuzzy logic controller parameters.

The proposed neuro‐fuzzy scheme is able to precisely learn the control relation between input‐output training data generated by the learning algorithm. From the experiments performed on the industrial grade robots at Robotics and Manufacturing Automation (RAMA) Laboratory, it was found that the neuro‐fuzzy controller was able to learn fuzzy logic rules and parameters accurately.

The backpropagation method, used in this research, is extremely dependent on initial choice of parameters, and offers no mechanism to restrict the parameters within specified range during training. Use of alternative learning mechanisms, such as reinforcement learning, needs to be investigated.

The neuro‐fuzzy scheme presented can be used to develop controller for plants for which it is difficult to obtain analytical model or sufficient information about input‐output heuristic relation is not available.

The paper presents the neural network architecture and introduces a learning mechanism to train this architecture online.

The purpose of this paper is to propose an on-line iterative compensation method combining with a feed-forward compensation method to enhance the assembly accuracy of a metrology-integrated robot system (MIRS).

By the integration of a six degrees of freedom (6DoF) measurement system (T-Mac), the robot’ movement can be tracked with real-time measurement. With the on-line measured data, the proposed iterative compensation for absolute positioning and the feed-forward compensation for relative linear motion are integrated into the assembly process to improve the assembly accuracy.

It is found that the MIRS exhibits good performance in both accuracy and efficiency with the application of the proposed compensation method. With the proposed assembly process, a component can be automatically aligned to the target in seconds, and the assembly error can be decreased to 0.021 mm for position and 0.008° for orientation on average.

This paper presents a 6DoF MIRS for high-precision assembly. Based on the system, a novel on-line compensation method is proposed to enhance the assembly accuracy. In this paper, the assembly accuracy and the corresponding distance parameter are given by a series of experiments as reference for assembly applications.

Rapid satellite capture by a free-floating space robot is a challenge problem because of no-fixed base and time-delay issues. This paper aims to present a modified target capturing control scheme for improving the control performance.

For handling such control problem including time delay, the modified scheme is achieved by adding a delay calibration algorithm into the visual servoing loop. To identify end-effector motions in real time, a motion predictor is developed by partly linearizing the space robot kinematics equation. By this approach, only ground-fixed robot kinematics are involved in the predicting computation excluding the complex space robot kinematics calculations. With the newly developed predictor, a delay compensator is designed to take error control into account. For determining the compensation parameters, the asymptotic stability condition of the proposed compensation algorithm is also presented.

The proposed method is conducted by a credible three-dimensional ground experimental system, and the experimental results illustrate the effectiveness of the proposed method.

Because the delayed camera signals are compensated with only ground-fixed robot kinematics, this proposed satellite capturing scheme is particularly suitable for commercial on-orbit services with cheaper on-board computers.

This paper is original as an attempt trying to compensate the time delay by taking both space robot motion predictions and compensation error control into consideration and is valuable for rapid and accurate satellite capture tasks.

The purpose of this paper is to analyze the sensitivity of the motion error parameters in microassembly process, thereby improving the assembly accuracy. The motion errors of the precision motion stages directly affect the final assembly quality after the machine visual alignment.

This paper presents the error parameters of the in-house microassembly system with coaxial alignment function, builds the error transfer model by the multi-body system theory, analyzes the error sensitivity on the sensitive direction using the Sobol method, which was based on variance, and then gets the ones which made a great degree of influence. Before the sensitivity analyzing, parts of the error sources have been measured to obtain their distribution ranges.

The results of the sensitivity analysis by the Sobol method, which was based on variance, are coincident with the theoretical analysis. Besides, the results provide a reference for the error compensation in control process, for the selection of the precision motion stages and for the installation index of the motion stages of the assembly system with coaxial alignment.

This kind of error sensitivity analysis method is of great significance for improving the assembly accuracy after visual system positioning, and increasing efficiency from the initial motion stage selection to final error compensation for designers. It is suitable for general precision motion systems be of multi-degree of freedom, for the method of modeling, measuring and analyzing used in this paper are all universal and applicative.

In view of the “staircase or stepped effect” in the present commercial rapid prototyping (RP) systems, a new approach is proposed, and is currently under development to produce true stepless parts. The new RP system is a combination of five‐axis motion control technology and precision milling. The system consists of three linear and two rotary axes, which enable the milling tool to be orientated for tangent cutting. Because of the complexity of the five‐axis configuration, ensuring high accuracy and precision is both important and difficult. This paper presents an open‐loop motion compensation algorithm, which enables the machine tool to move more accurately to a given position and orientation than its physical setup allows. Experimental procedures to measure the geometrical errors in the system setup are also presented in this paper.

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.

This study aims to evolve an enhanced butterfly optimization algorithm (BOA) with respect to convergence and accuracy performance for numerous benchmark functions, rigorous constrained engineering design problems and an inverse synthetic aperture radar (ISAR) image motion compensation.

Adaptive BOA (ABOA) is thus developed by incorporating spatial dispersal strategy to the global search and inserting the fittest solution to the local search, and hence its exploration and exploitation abilities are improved.

The accuracy and convergence performance of ABOA are well verified via exhaustive comparisons with BOA and its existing variants such as improved BOA (IBOA), modified BOA (MBOA) and BOA with Levy flight (BOAL) in terms of various precise metrics through 15 classical and 12 conference on evolutionary computation (CEC)-2017 benchmark functions. ABOA has outstanding accuracy and stability performance better than BOA, IBOA, MBOA and BOAL for most of the benchmarks. The design optimization performance of ABOA is also evaluated for three constrained engineering problems such as welded beam design, spring design and gear train design and the results are compared with those of BOA, MBOA and BOA with chaos. ABOA, therefore, optimizes engineering designs with the most optimal variables. Furthermore, a validation is performed through translational motion compensation (TMC) of the ISAR image for an aircraft, which includes blurriness. In TMC, the motion parameters such as velocity and acceleration of target are optimally predicted by the optimization algorithms. The TMC results are elaborately compared with BOA, IBOA, MBOA and BOAL between each other in view of images, motion parameter and numerical image measuring metrics.

The outperforming results reflect the optimization and design successes of ABOA which is enhanced by establishing better global and local search abilities over BOA and its existing variants.