Wei Huo

In this paper, we propose a Bi-layer Predictionbased Reduction Branch (BP-RB) framework to speed up the process of finding a high-quality feasible solution for Mixed Integer Programming (MIP) problems. A graph convolutional network (GCN) is employed to predict binary variables' values. After that, a subset of binary variables is fixed to the predicted value by a greedy method conditioned on the predicted probabilities. By exploring the logical consequences, a learning-based problem reduction method is proposed, significantly reducing the variable and constraint sizes. With the reductive sub-MIP problem, the second layer GCN framework is employed to update the prediction for the remaining binary variables' values and to determine the selection of variables which are then used for branching to generate the Branch and Bound (B&B) tree. Numerical examples show that our BP-RB framework speeds up the primal heuristic and finds the feasible solution with high quality.

Internet of Things(IOT) technology provides sufficient information technology and necessary decision-making basis for intelligent transportation. Internet of vehicles(IOV) is based on the architecture of the perception layer, network layer, and application layer. This article focuses on the tail light status in the perception layer of automotive active safety systems and proposes a method of vehicle taillamp detection and recognition based on active contour extraction and color feature analysis to solve the problem of information perception against the front vehicle. First, the front vehicles are detected and extracted by convolutional neural network and active contour models. Second, it analyzes the characteristics of taillamp pair and then segment the correct taillamp pair by color space conversion and location correlation principle. On this basis, the histograms distribution of taillamp are analyzed in different states and complete the recognition of taillamp. The experimental results show that the method can accurately identify the taillamp signal of the front vehicle in the daytime and effectively reduce the rate of false judgement caused by light interference.

This paper investigates the robust relative pose control for spacecraft rendezvous and proximity operations subject to input saturation, kinematic couplings, parametric uncertainties, and unknown external disturbances. Relative rotational and relative translational nonlinear system models are first derived, and relative attitude and relative position controllers are then proposed, respectively. The kinematic couplings, parametric uncertainties, and unknown external disturbances in dynamical models are treated as compound disturbances, and nonlinear disturbance observers are developed and incorporated into the relative pose control design, which can avoid the assumption on the bounded derivatives of compound disturbances. Meanwhile, input saturation effect of the control torques and forces is compensated by synthesizing the outputs of the auxiliary systems into the controllers. Based on the proposed disturbance observers and auxiliary systems, saturated attitude synchronization and position tracking controllers are developed to reject the unknown compound disturbances and ensure the convergence of the relative pose and velocities. The stability of the closed-loop system is rigorously proved in the Lyapunov framework; relative pose and velocities ultimately converge to the small neighborhoods of the origin in spite of input saturation and model uncertainties. Simulation experiments validate the performance of the proposed robust saturated control strategy.