Hardware architecture design of block-matching and 3D-filtering denoising algorithm

Block-matching and 3D-filtering(BM3D) is a state of the art denoising algorithm for image/video,which takes full advantages of the spatial correlation and the temporal correlation of the video. The algorithm performance comes at the price of more similar blocks finding and filtering which bring high computation and memory access. Area, memory bandwidth and computation are the major bottlenecks to design a feasible architecture because of large frame size and search range. In this paper, we introduce a novel structure to increase data reuse rate and reduce the internal static-random-access-memory(SRAM) memory. Our target is to design a phase alternating line(PAL) or real-time processing chip of BM3 D. We propose an application specific integrated circuit(ASIC) architecture of BM3 D for a 720 × 576 BT656 PAL format. The feature of the chip is with 100 MHz system frequency and a 166-MHz 32-bit double data rate(DDR). When noise is σ = 25, we successfully realize real-time denoising and achieve about 10 d B peak signal to noise ratio(PSNR) advance just by one iteration of the BM3 D algorithm.     

Study on Oxidation Activity of CuCeZrO<Subscript><Emphasis Type="Italic">x</Emphasis></Subscript> Doped with K for Diesel Engine Particles in NO/O<Subscript>2</Subscript>

CuCeZrO_x and KCuCeZrO_x catalysts were synthesized and coated on the blank diesel particulate filter (DPF) substrate and a particulate matter (PM) loading apparatus was used for soot loading. The catalytic performances of soot oxidation were evaluated by temperature programmed combustion (TPC) test and characterization tests were conducted to investigate the physicochemical properties of the catalysts. The reaction mechanism in the oxidation process was analyzed with diffuse reflectance infrared Fourier transform spectroscopy. The results demonstrated that CuCeZrO_x catalyst exhibited high activities of soot oxidation at low temperature and the best results have been attained with Cu_0.9Ce_0.05Zr_0.05O_x over which the maximum soot oxidation rate decreased to 410 °C. Characterization tests have shown that catalysts containing 90% Cu have uniformly distributed grains and small particle sizes, which provide excellent oxidation activity by providing more active sites and forming a good bond between the catalyst and the soot. The low-temperature oxidation activity of soot could be further optimized due to the excellent elevated NO’s conversion rate by partially substituting Cu with K. The maximum particle oxidation rate can be easily realized at such a low temperature as 347°C.     

Vehicle trajectory challenge in predictive active steering rollover prevention

High center of mass vehicles are likely to rollover in extreme maneuvers. Available works present control strategies to prevent rollover. In these works, however, other important parameters such as path trajectory tracking are not a main concern. In this paper conflicts between rollover prevention and trajectory tracking is investigated. Model predictive control (MPC) is adopted to predict and avoid rollover while tracking desired trajectory. For this regard a model based future error estimation is introduced. The control framework predicts both rollover and trajectory error simultaneously. It avoids rollover while tries to track the trajectory. Simulation results for two controllers with and without trajectory tracking are presented. The results indicate that the controllers effectively limit rollover as a hard constraint while the trajectory tracking controller also minimizes and recovers the path error.     

Design and validation of an electro-hydraulic brake system using hardware-in-the-loop real-time simulation

This paper presents a novel electric booster (E-booster) that exibits superior performance advantages over traditional vacuum boosters. The proposed E-booster, consisting of an electric motor and a ball screw assembly, is designed for electro-hydraulic brake (EHB) systems to meet relevant requirements for electric vehicles and active safety technologies. A mathematical model for an EHB system is generated to determine the desired values of the parameters for the E-booster prototype using numerical simulation in MATLAB. Simulation results of the EHB system with the virtual E-booster demonstrate the feasibility and effectiveness of the innovative technique. Built upon the results derived from the numerical simualtions, an integrated algorithm based on the Kalman filter and a sliding mode control technique is designed to control the E-booster motor and to implement the brake booster function. A hardware-in-the-loop (HIL) real-time simulation system equipped with the E-booster prototype is developed. HIL real-time simulations are conducted to evaluate the proposed algorithm. The HIL real-time simulation results demonstrate that the proposed algorithm generates booster brake forces fast, and forces the ball nut to track the push rod well to ensure comfortable brake pedal feel.     

Shape optimization of a torsion beam axle for improving vehicle handling performance

In this study, shape optimization was conducted for a vehicle’s rear suspension torsion beam to improve its dynamic handling performance. To determine the design variables affecting the vehicle roll characteristics, a sensitivity analysis was conducted using the result of a Taguchi experiment with 6 factors in 8 runs. The upper and lower-flange lengths and web thickness of the torsion beam section, as well as the vertical height difference between the inner and outer of torsion beams, were determined as design variables through sensitivity analysis of the opposite wheel travel test for optimization of the torsion beam axle. The Box–Behnken experimental design with 4 factors and 27 runs was performed using the selected design variables and by performing opposite wheel travel analysis according to the experimental design, and the response surface functions of the roll stiffness, roll steer coefficient, roll center height, and mass of the torsion beam were generated. Using these response functions, shape optimization was conducted for the torsion beam of the rear suspension system. Dynamic performance analysis was performed by applying the optimized H-shaped torsion beam to the rear suspension of the vehicle dynamics model, and it was validated that the dynamic response performance of the optimized vehicle was improved.     

Cyclist target and test setup for evaluation of cyclist-autonomous emergency braking

From 2018, Autonomous Emergency Braking (AEB) systems dedicated to avoid or mitigate passenger car-tocyclist collisions will be considered in the safety assessment by Euro NCAP. To test such systems, appropriate equipment has been developed in a project called CATS “Cyclist-AEB Testing System”, that has run between April 2014 and August 2016. Moreover, a proposal for the most relevant test scenarios was set up. The objective of the project was to provide proof to Euro NCAP of the relevance of the proposed test scenarios and of the feasibility of practical implementation of the scenarios and test setup. The process regarding the selection, verification and validation of test scenarios is described. The cooperation between 17 industrial partners (car manufacturers and automotive suppliers) in the CATS project has stimulated the harmonization and acceptance of the protocol, target and test setup. The process and intermediate results including the used methodology, have been reviewed by the German Federal Highway Research Institute (BASt) and have been shared on a regular basis during the project with stakeholders in Europe, Japan and the USA. Euro NCAP already indicated to consider the results of the CATS project as the main input to draft the test protocol, including scenarios and target for Cyclist-AEB systems in 2018 and 2020.     

MPC-BASED steering control for backward-driving vehicle using stereo vision

We propose a steering control algorithm for autonomous backward driving in a narrow corridor. Passable spaces are detected using a stereo camera, and the steering angle is controlled by a model predictive controller (MPC). For passable space detection, an UV-disparity map is calculated from the original disparity map. Information regarding passable spaces collected by the stereo camera is used in steering control. Backward driving requires the driver’s preemptive actions, which can be learned by experience because of the non-intuitive responses (the initial motion of the vehicle is opposite to the driver’s steering angle input). This occurs because a backward-driving vehicle is a non-minimum phase system. One of the most popular steering control algorithms is Stanley method, which is based on the feedback of lateral displacement error and heading angle error. The method is very intuitive and works well for forward driving, but it exhibits significant undershoot for backward driving cases. Furthermore, the method does not explicitly consider any constraints on control inputs and states. We designed a steering controller based on the MPC technique that requires future information but can handle constraints explicitly. Because we have near-future information from the stereo camera under limited passable spaces, MPC can be effectively implemented. We performed several simulations and experiments to show the performance and superiority of the suggested method over a simple feedback-based control algorithm.     

Direct tire force generation algorithm based on non-iterative nonlinear inverse tire model

The function of vehicle dynamics control system is adjusting the yaw moment, the longitudinal force and lateral force of a vehicle body through several chassis systems, such as brakes, steering and suspension. Individual systems such as ESC, AFS and 4WD can be used to achieve desired performance by controlling actuator variables. However, integrated chassis control systems that have multiple objectives may not simply achieve the desired performance by controlling the actuators directly. Usually those systems determine the required tire forces in an upper level controller and a lower level controller regulates the tire forces through the actuators. The tire force is controlled in a recursive way based on vehicle state measurement, which may not be sufficient for fast response. For immediate force tracking, we introduce a direct tire force generation method that uses a nonlinear inverse tire model, a pseudo-inverse model of vehicle dynamics and the relationship between longitudinal force and brake pressure.     

Multi-Component Model of Diesel Sprays Under High Injection Pressure

The combustion efficiency of a diesel engine depends not only on spray characteristics but also on fuel-air mixing characteristics. Based on the original spray model, a new spray model is established in this paper to accurately predict the diesel spray, and then a multi-component evaporation model is added into it. The model takes the influence of component concentration gradient and species on its evaporation rate in the liquid phase into account. This paper studies the spray characteristics (spray penetration, spray angle and spray morphology) and fuel-air mixing characteristics (spray area, spray volume and air entrainment mass) using the spray model, and the results are compared with the experimental results. The comparison shows that the simulated spray penetration and spray angle are close to the experimental results with the average deviations less than 3%. Moreover, this paper studies the spray area, spray volume and air entrainment using empirical formula under different conditions. And the maximum deviations of the spray volume, spray area and air entrainment mass are less than 5% as compared with the test values. Overall, this spray model can predict the diesel spray characteristics and fuel-air mixing characteristics under high injection pressure accurately.     

Feature Selection,Deep Neural Network and Trend Prediction

The literature generally agrees that longer-horizon (over a month) predictions make more sense than short-horizon ones. However, it’s an especially challenging task due to the lack of data (in unit of long horizon) and economic data have a low S/N ratio. We hypothesize that the stock trend is largely dictated by driving factors which are filtered by psychological factors and work on behavioral factors: representative indicators from these three aspects would be adequate in trend prediction. We then extend the Stepwise Regression Analysis (SRA) algorithm to constrained SRA (cSRA) to carry out a further feature selection and lag optimization. During modeling stage, we introduce the Deep Neural Network (DNN) model in stock prediction under the suspicion that economic interactions are too complex for shallow networks to capture. Our experiments indeed show that deep structures generally perform better than shallow ones. Instead of comparing to a kitchen sink model, where over-fitting can easily happen with a shortage of data, we turn around and use a model ensemble approach which indirectly demonstrates our proposed method is adequate.