Smooth state switching and accurate speed tracking are important for the stability reactivity of bipedal robots when running. However, previous studies have rarely been able to synthesize these two capabilities online. In this paper, we present an online running-gait generator that allows smooth tracking. Considering a fluctuating height nature computational expediency, robot is represented by simplified variable-height inverted-pendulum (VHIP) model. order achieve at beginning end running,...

In this paper, the effect of surface textures braking disc on performance is experimentally investigated under conditions different working gaps and applied currents. For purpose, a new configuration magnetorheological fluid brake (MRB) with adjustable gap developed to improve manufacturing accuracy cost, reduce problem replacing disc. addition, discs three types texture are designed machined. Based test bed for proposed MRB, series experiments carried out manufactured prototype results...

In order to optimize the three-dimensional (3D) reconstruction and obtain more precise actual distances of object, a 3D system combining binocular depth cameras is proposed in this paper. The whole consists two identical color cameras, TOF camera, an image processing host, mobile robot control robot. Because structural constraints, resolution camera very low, which difficultly meets requirement trajectory planning. stereo can be high, but effect matching not ideal for low-texture scenes....

Fire on sealing ring is the main fire form of large external floating-roof tanks. Once occurs large-scale tank ring, existing fixed extinguishing equipment cannot put out quickly and efficiently, which would result in giant losses. The signal can be detected, source located automatically extinguished precisely through combination applicable imagine monitoring system intelligent fireextinguishing fire-fighting gun control system. capacity self-explosion prevention control, scale oil will...

Knee-stretched walking is considered to be a human-like and energy-efficient gait. The strategy of extending legs obtain vertical center mass trajectory commonly used avoid the problem singularities in knee-stretched gait generation. However, generation utilizing this with toe-off heel-strike has kinematics conflicts at transition moments between single support double phases. In article, for position-controlled humanoid robot been proposed. position constraints have based on model predictive...

To automatically detect the front obstacles of smart car, a traffic avoidance system based on binocular ranging is presented in this paper. The whole consists two cameras, an image processing system, intelligent car control host, and two-drive tracked car. Specifically, cameras are calibrated to get distortion parameters, internal relative rotation matrix translation parameters. Then stereo matching algorithm applied obtain disparity map scene, three-dimensional scene reconstructed real...

A fine velocity planning model can make robots move efficiently and stably. The smoothness of the velocity, acceleration jerk protect arm from moving problems, such as motion vibration transient overshoot. In order to have a movement for multiple degrees freedom (DOF) robots, people generally optimize track on connection point two tracks. If tracks are fixed, an adaptive is needed optimized trajectory planning. This paper presents algorithm manage speeds 6-DOF industrial with fixed according...

Having human-like legged structure, biped robots are regarded as the most promising that can take place of human to work in environments designed for beings. To this end, it is premise achieve stable walking on uneven ground. This paper presents a control strategy robot based spring-loaded inverted pendulum (SLIP) model, which controls joint torque by obtaining errors length and angular position virtual spring legs. Additionally, proposes new method adjusting legs enable unstructured...

Recent years deep learning based visual object trackers have achieved state-of-the-art performance on multiple benchmarks. However, most of these lack an effective mechanism to avoid the wrong template update or re-detect when unreliable tracking result appears. In this paper, a novel framework consisting network for locating target and policy decision making is proposed. Firstly, during off-line training phase, variant gradient algorithm adopted, which makes model converge better faster....

Good balancing is very important to improve the moving performance of grinding machine. General technology grinders was analyzed. The structure features a new type hybrid CNC machine were introduced. necessity obtaining good balance grinder demonstrated. Based on simulation and calculation distribution law potential energy over grinder, solution for partial worked out. An optimal parameter value designed unit reached through optimization with effect as target function. max actuating force...

For humanoid robots, online motion generation on non-flat terrain remains an ongoing research challenge. Computational complexity is one of the primary restrictions that preclude planners from generating adaptive behaviors online. In this paper, we investigate problem and decompose it into two sequential components: Efficient Behavior Generator (EBG) a Nonlinear Centroidal Model Predictive Controller (NC-MPC). The EBG responsible for optimizing physically feasible whole-body template...

A 3-TPS(RRR) hybrid machine tool that could process vertically as well horizontally was designed in Northeastern University. parallel locking mechanism, equipped between the moving platform and fixed base, is used to constrain partial degrees of freedom (DOF) platform. It leads has three translational DOF a rotate DOF. The kinematic analysis discussed this paper. Several important issues were addressed, namely, position, velocity, acceleration, Jacobian matrix. simulation also carried out,...

In this paper the control object is non-holonomic mobile robot and problem focus on point stabilization which one of most important issues motion in view both theory application. The integrity kinematic model position error formula established under polar coordinates for problem. Then step-by-step studied where method design given. By successively changing state variables, will get expecting be convergence. On basis law simulation carried by MATLAB. results prove feasibility controller.

A common approach to learning rank is minimize the pair-wise loss. However, established analysis shows that loss does not necessarily lead an optimal list-wise ranking measures, e.g., average precision (AP) or area under precision-recall curve (AUPRC). It becomes more difficult in online setting, where data arrives sequentially and scanned only once. This paper proposes learning-to-rank algorithm by minimizing error, which achieves a vanishing gap between measures. Experiments also testify...

Rate distortion optimization (RDO) is the basis for algorithm in video coding [1], such as mode decision, rate control and etc. Minimizing cost usually employed to determine optimal parameters quantization level, mode, However, calculations solution decision from massive possible candidates suffer dramatically high computation complexity. To resolve this problem, paper proposes a fast TU level model with higher accuracy by fully imitating behavior pattern hid entropy.