Machine learning applications in activity-travel behaviour research: a review

ABSTRACT

This paper reviews the activity-travel behaviour literature that employs Machine Learning (ML) techniques for empirical analysis and modelling. Machine Learning algorithms, which attempt to build intelligence utilizing the availability of large amounts of data, have emerged as powerful tools in the fields of pattern recognition and big data analysis. These techniques have been applied in activity-travel behaviour studies since the early ’90s when Artificial Neural Networks (ANN) were employed to model mode choice decisions. AMOS, an activity-based modelling system developed in the mid-’90s, has ANN at its core to model and predict individual responses to travel demand management measures. In the dawn of 2000, ALBATROSS, a comprehensive activity-based travel demand modelling system, was proposed by Arentze and Timmermans using Decision Trees. Since then researchers have been exploring ML techniques like Support Vector Machines (SVM), Decision Trees (DT), Neural Networks (NN), Bayes Classifiers, and more recently, Ensemble Learners to model and predict activity-travel behaviour. A large number of publications over the years and an upward trend in the number of published articles over time indicate that Machine Learning is a promising tool for activity-travel behaviour analysis and prediction. This article, first of its kind in the literature, reviews these studies and explores the trends in activity-travel behaviour research that apply ML techniques. The review finds that mode choice decisions have received wide attention in the literature on ML applications. It was observed that most of the studies identify the lack of interpretability as a serious shortcoming in ML techniques. However, very few studies have attempted to improve the interpretability of the models. Further, some studies report the importance of feature engineering in ML-based studies, but very few studies adopt feature engineering before model development. Spatiotemporal transferability of models is another issue that has received minimal attention in the literature. In the end, the paper discusses possible directions for future research in the area of activity-travel behaviour modelling using ML techniques.     

An ontological model to support communications of situation-aware vehicles

In an Intelligent Transport System (ITS) environment, the communication component is of great importance to support interactions between vehicles and roadside infrastructure. Previous studies have focused on the physical capability and capacity of the communication technologies, but the equally important development of suitable and efficient semantic content for transmission received notably less attention. Ontology is one promising approach for context modelling in ubiquitous computing environments, and in the transport domain it can be used both for context modelling and semantic contents for vehicular communications. This paper explores the development of an ontological model implementing relative geo-semantic information messages to support vehicle-to-vehicle communications. The proposed ontology model contains classes, objects, their properties/relations as well as some functions and query templates to represent and update the information of dynamic vehicles, inter-vehicle interactions and behaviour. This model was developed through a scenario enabling the evaluation of traffic conflict resolution approaches, by implementing a set of decision-making processes for intelligent vehicles. Given the scope of the proposed ontology modelling, it shows how vehicular communications can be used to update each vehicle’s context model. This work can be easily extended for more complex interactions among vehicles and the infrastructure.     

Centrifuge modelling of a combined pile-type anchor subjected to general inclined loading

One of the main challenges associated with anchoring offshore platforms is the logistic needed to install them in deep water. The growing need for high capacity anchors has prompted the development of new systems where the feasibility of transporting them out to sea is a crucial consideration. With respect to torpedo piles used by Petrobras in Brazil, new platforms installed in deep water required robust anchoring systems, and a cluster of small torpedoes was considered instead of one large torpedo. However, torpedo pile installation is hindered by deep water streams, which can deviate their trajectory during the fall. This is of great concern for the design of torpedo pile clusters because the distance between the piles cannot be controlled during installation. The present study aims to determine the variation in the load capacity of a cluster or array of single piles as a function of the distance between the piles and the load direction after installation on the seabed. To that end, a series of centrifuge tests were carried out on pile clusters installed in loose sand, varying the distance between the piles and the loading direction. The results show considerable interactions between the adjacent piles depending on the load direction and pile spacing. The optimal spacing for maximum cluster efficiency was determined, and comparisons were made with values from the literature for capped pile clusters.     

Design,modelling, and analysis of a large floating dock for spar floating wind turbine installation

Installation of floating wind turbines at the offshore site is a challenging task. A significant part of the time efficiency and costs are related to the installation methods which are sensitive to weather conditions. This study investigates a large floating dock concept, which can be used to shield a floating wind turbine during installation of tower, nacelle, and rotor onto a spar foundation. In this paper, the concept is described in detail, and a design optimisation is carried out using simple design constraints. Hydrodynamic analysis and dynamic response analysis of the coupled system of the optimum dock and spar are conducted. Two spars of different sizes are considered, and the motion responses of the spars with and without the dock in irregular waves are compared. Through analysis of the motion spectra and response statistics, dynamic characteristics of the coupled system is revealed. The present design of the dock reduces the platform-pitch responses of the spars and potentially facilitates blade mating, but may deteriorate the heave velocity of the spars in swell conditions. Finally, future design aspects of the floating dock are discussed.     

Experimental and numerical study of localized pitting effect on compressive behavior of tubular members

Locally pitted tubular members are usually considered as stub columns to assess the ultimate strength. However, it is not suitable for those with relatively larger slenderness ratios as their failure behavior is more complex and closely related to corrosion features of localized pitting. This paper presents compressive column tests on locally pitted tubular members of a moderate slenderness ratio. Corrosion pits were artificially introduced on local surface of the members, forming corrosion patches with various corrosion features. A numerical modelling method was proposed to reproduce the test specimens. Localized pitting damage was proven to cause substantial declines in the load deformation capacity and ultimate strength, and have a significant effect on the failure mode. The failure of a pitted member is mostly initiated by local buckling after yielding occurs in the corrosion patch, concurrent with pitting closure, and even shear cracking of member wall due to the perforated pits. Moreover, shape change of the corrosion patch most likely results in the failure mode to alter from column buckling to local buckling or interactive buckling. The shape ratio of the corrosion patch is one of the critical factor to influence the ultimate strength of locally pitted members. The proposed modelling method is applicable for extensive stochastic simulations so as to develop an empirical formula and to clarify the probabilistic characteristics of ultimate strength.     

Repeatability of ice-tank tests with broken ice

This study investigates the repeatability of ice-tank tests with broken ice. Ice-tank test campaigns normally do not perform multiple repetitions of tests with the same initial conditions. Therefore, the repeatability of ice-tank tests with broken ice is not well understood. Data from two test campaigns are analysed. The first test campaign studied the interaction between a 4-legged structure with a vertical waterline and several broken and intact level ice conditions. In the second test campaign, a ship hull geometry was tested. We analyse selected test cases from each test campaign. The ice-tank tests are reproduced using a 3-D discrete element method (DEM) model. Each analysed test case is simulated 20 times. The only difference between each simulation is the initial position of the ice floes. The numerical simulation results show that changes in the initial floe positions can cause large changes in the statistical properties of the ice load. Often, a single random interaction event can be identified that is responsible for the change in the results. Such interaction events can cause additional floe accumulation ahead of the structure, thereby influencing the load statistics for a large portion of the interaction length. The observed events occur both in the numerical simulations and in the physical ice-tank tests. This result indicates that ice-tank tests with broken ice have a poor repeatability; a change in an uncontrolled condition, such as the exact initial floe positions, can lead to a large variation in the experimental results.     

Impacts of automated vehicles on travel behaviour and land use: an international review of modelling studies

ABSTRACT

Automated vehicles (AVs) could completely change mobility in the coming years and decades. As AVs are still under development and gathering empirical data for further analysis is not yet possible, existing studies mainly applied models and simulations to assess their impact. This paper provides a comprehensive review of modelling studies investigating the impacts of AVs on travel behaviour and land use. It shows that AVs are mostly found to increase vehicle miles travelled and reduce public transport and slow modes share. This particularly applies to private AVs, which are also leading to a more dispersed urban growth pattern. Shared automated vehicle fleets, conversely, could have positive impacts, including reducing the overall number of vehicles and parking spaces. Moreover, if it is assumed that automation would make the public transport system more efficient, AVs could lead to a favouring of urbanisation processes. However, results are very sensitive to model assumptions which are still very uncertain (e.g. the perception of time in AVs) and more research to gain further insight should have priority in future research as well as the development of the models and their further adaptation to AVs.     

A theoretical justification of the sine with dwell manoeuvre

The sine with dwell (SWD) manoeuvre has received much attention within the context of vehicle stability testing. This manoeuvre is used in a test developed by the NHTSA, designed to certificate electronic stability control systems in light vehicles. The test is used in legislations as well as in consumer tests all over the globe. The SWD manoeuvre was designed using test vehicles on a test track and experimentally validated. The paper at hand uses optimal control theory to theoretically justify the use of the SWD manoeuvre to produce a severe lateral motion and over-steering based on steering input. It is shown that a manoeuvre similar to the SWD manoeuvre can be obtained from an optimal control problem using simple vehicle dynamics models. The optimal control method is further used to analyse the manoeuvre's robustness with respect to vehicle dimensions and tyre properties. It is shown that the manoeuvre is robust in dimensions, which theoretically motivates its application for various sizes of vehicles.     

基于ADAMS的轮式装载机工作装置动态载荷分析

以反转六连杆轮式装载机ZL50G为样机,运用ADAMS软件建立样机的三维实体模型,对轮式装载机的工作装置进行动态载荷分析,为轮式装载机工作装置的研究提供一种可靠的方法.分析结果为轮式装载机工作装置的设计与优化提供了理论依据.     

Key factors influencing the global passenger transport dynamics using the AIM/transport model

A bottom-up passenger transport model named AIM (Asia-pacific Integrated Model)/Transport model is developed by incorporating behavioral parameters and transportation technological details. This model is based on discrete based choice modelling covering 17 global regions soft-linked with the AIM/CGE (Computable General Equilibrium) model. In this paper, the model is used to assess the impact of various factors like travel time, energy efficiency improvement, load factor, mode preference along with environmental awareness factors on transport demand, energy and emissions. The modelling assessment results show that travel speed and land-use patterns have significant impact on the travel demand. High occupancy rate and shift towards the mass-transit system result in energy and emissions reduction. Implementation of carbon tax aligned with the two-degree target results in a 22% cumulative emission reduction from 2005 to 2100 relative to the baseline case. However, the reduction potential can be increased to 42% by combining behavioral and technology related mitigation options like mass-transit system speed improvement, transit oriented development, efficiency improvement, preference towards eco-friendly technologies and high vehicle occupancy.