Multivariate adaptive regression splines and neural network models for prediction of pile drivability

Piles are long, slender structural elements used to transfer the loads from the superstructure through weak strata onto stiffer soils or rocks. For driven piles, the impact of the piling hammer induces compression and tension stresses in the piles. Hence, an important design consideration is to check that the strength of the pile is sufficient to resist the stresses caused by the impact of the pile hammer. Due to its complexity, pile drivability lacks a precise analytical solution with regard to the phenomena involved.In situations where measured data or numerical hypothetical results are available, neural networks stand out in mapping the nonlinear interactions and relationships between the system’s predictors and dependent responses. In addition, unlike most computational tools, no mathematical relationship assumption between the dependent and independent variables has to be made. Nevertheless, neural networks have been criticized for their long trial-and-error training process since the optimal configuration is not known a priori. This paper investigates the use of a fairly simple nonparametric regression algorithm known as multivariate adaptive regression splines（MARS）, as an alternative to neural networks, to approximate the relationship between the inputs and dependent response, and to mathematically interpret the relationship between the various parameters. In this paper, the Back propagation neural network（BPNN） and MARS models are developed for assessing pile drivability in relation to the prediction of the Maximum compressive stresses（MCS）, Maximum tensile stresses（MTS）, and Blow per foot（BPF）. A database of more than four thousand piles is utilized for model development and comparative performance between BPNN and MARS predictions.

Integrated Control Strategy for CVT Powertrains with Consideration of Vehicle Drivability

During shift,power flow is not interrupted in powertrains equipped with continuously variable transmission（CVT）.When hard acceleration is commanded,engine speed will flare and corresponding torque will be consumed,which leads to a drop in vehicle drive torque and also the vehicle acceleration.This is the reason why CVT vehicles have poor drivability during hard acceleration maneuver.Conventional method such as torque compensation doesn＇t always work due to the limited backup torque of engine.According to this,means to evaluate the drivability of CVT vehicles are studied,affect factors of drivability are analyzed in detail.Hard acceleration process of CVT vehicle is studied by theoretical analysis,based on which engine torque and ratio change rate of CVT are identified as two key control parameters that decide the drivability of CVT vehicles during hard acceleration maneuver.Therefore,a control strategy based on restricting the change rate of CVT ratio together with torque compensation is proposed,and two different algorithms to establish the limitation of ratio change rate are proposed.These two algorithms are simulated and compared with each other,results indicate that drop of vehicle acceleration is eliminated evidently by limit the change rate of CVT ratio,but small ratio change rate also results in a longer time to finish the accelerate process,an algorithm to decide a proper ratio change rate is needed in order to tune these different characteristics.In order to get better control effects,a new fuzzy logic based algorithm is proposed to decide a proper ratio change rate during kick down conditions,simulation and experiment results indicate that,the amount of vehicle acceleration decrease is reduced from about 1 m/s2 to almost 0,in the mean time the accelerate process only delayed for about 0.3 s.The proposed control strategy and algorithm can effectively tune the characteristics of CVT equipped vehicle during kick down conditions.

Soil Plugging Effect on Drivability Prediction of Offshore Platform Piles

Field measurements of driving resistances and heights of soil core during driving were made offshore and onshore of steel pipe piles. Measured data show that the height of soil core varies differently for piles of different diameters with the increase of penetration. Dynamic plugging could be assumed never to occur for steel pipe piles with diameters over 900 mm. Soil resistances at the time of continuous driving (SRD) are back analyzed from blow counts with an empirical distribution of resistances suppported by many early dynamic measurements. A method of predicting SRD is finally suggested.

Desktop calibration of automatic transmission for passenger vehicle

Desktop calibration of automatic transmission(AT) is a method which can reduce cost, enhance efficiency and shorten the development periods of a vehicle effectively. We primary introduced the principle and approach of desktop calibration of AT based on the condition of coupling characteristics between engine and torque converter and obtained right point exactly. It is shown to agree with experimental measurements reasonably well. It was used in different applications abroad based on AT technology and achieved a good performance of the vehicle compared with traditional AT technology which primary focuses on the drivability, performance and fuel consumption.

YOLOP:You Only Look Once for Panoptic Driving Perception

A panoptic driving perception system is an essential part of autonomous driving.A high-precision and real-time perception system can assist the vehicle in making reasonable decisions while driving.We present a panoptic driving perception network(you only look once for panoptic(YOLOP))to perform traffic object detection,drivable area segmentation,and lane detection simultaneously.It is composed of one encoder for feature extraction and three decoders to handle the specific tasks.Our model performs extremely well on the challenging BDD100K dataset,achieving state-of-the-art on all three tasks in terms of accuracy and speed.Besides,we verify the effectiveness of our multi-task learning model for joint training via ablative studies.To our best knowledge,this is the first work that can process these three visual perception tasks simultaneously in real-time on an embedded device Jetson TX2(23 FPS),and maintain excellent accuracy.To facilitate further research,the source codes and pre-trained models are released at https://github.com/hustvl/YOLOP.

Driving Space for Autonomous Vehicles

Driving space for autonomous vehicles(AVs)is a simplified representation of real driving environments that helps facilitate driving decision processes.Existing literatures present numerous methods for constructing driving spaces,which is a fundamental step in AV development.This study reviews the existing researches to gain a more systematic understanding of driving space and focuses on two questions:how to reconstruct the driving environment,and how to make driving decisions within the constructed driving space.Furthermore,the advantages and disadvantages of different types of driving space are analyzed.The study provides further understanding of the relationship between perception and decision-making and gives insight into direction of future research on driving space of AVs.