Psychological constructs in driving automation: a consensus model and critical comment on construct proliferation

As automation in vehicles becomes more prevalent, the call for understanding the behaviour of the driver while driving an automated vehicle becomes more salient. Although a variety of driver behaviour models exist, and various psychological constructs have been said to be influenced by automation, an empirically testable psychological model of automated driving has yet to be developed. Building upon Stanton and Young's model of driving automation, this article presents an updated model of interrelated psychological constructs. The proposed model was created based upon a systematic literature search of driving automation papers and a subsequent quantification of the number of reported links between a selected set of psychological constructs. A secondary aim of this article is to reach consensus in the use of psychological constructs regarding driving automation. Henceforth special attention is paid to resolving the issue of construct proliferation.     

Correcting evaluation bias of relational classifiers with network cross validation

Recently, a number of modeling techniques have been developed for data mining and machine learning in relational and network domains where the instances are not independent and identically distributed (i.i.d.). These methods specifically exploit the statistical dependencies among instances in order to improve classification accuracy. However, there has been little focus on how these same dependencies affect our ability to draw accurate conclusions about the performance of the models. More specifically, the complex link structure and attribute dependencies in relational data violate the assumptions of many conventional statistical tests and make it difficult to use these tests to assess the models in an unbiased manner. In this work, we examine the task of within-network classification and the question of whether two algorithms will learn models that will result in significantly different levels of performance. We show that the commonly used form of evaluation (paired t-test on overlapping network samples) can result in an unacceptable level of Type I error. Furthermore, we show that Type I error increases as (1) the correlation among instances increases and (2) the size of the evaluation set increases (i.e., the proportion of labeled nodes in the network decreases). We propose a method for network cross-validation that combined with paired t-tests produces more acceptable levels of Type I error while still providing reasonable levels of statistical power (i.e., 1−Type II error).     

Effects of Using Alternative Extreme Pressure (EP) and Anti-Wear (AW) Additives with Oxy-Nitrided Samples

Oxy-nitriding is a widely used industrial process aiming to improve the tribological properties and performance of components. Previous studies have shown the effectiveness of the treatment with friction and wear performance, but very few have focussed on optimising this behaviour. The lubrication properties of several EP and AW additives were examined to investigate their effectiveness in improving the tribological properties of the layers formed after treatment. Previous studies showed the presence of an oxide layer on the sample could improve the effectiveness of the sulphurised olefin (SO) and tricresyl phosphate (TCP) additives. The friction and wear behaviour of oxy-nitrided samples were analysed using a tribometer and surface profiler. Scanning electron microscope, energy-dispersive X-ray spectroscopy and X-ray photoelectron spectroscopy were employed to identify the morphologies and chemical compositions of the treated surface before and after testing. No real effect on friction was observed when using the SO or TCP additives, mostly due to lack of interaction with the less reactive iron nitride layer and their roles as anti-wear additives. However, when the zinc dialkyldithiophosphate-containing lubricant was used, a higher friction coefficient was observed. Greater improvements in anti-wear properties with the presence of additives in comparison with only using base oil were reported, with the TCP additive producing the lowest wear rates. The study effectively demonstrated that the additive package type used could impact the tribological and tribochemical properties of oxy-nitrided surfaces.     

Joint optimisation of tracking capability and price in a supply chain with endogenous pricing

Tracking systems have been widely used to resolve the issues of product recall and food safety. Thus far, few researches have been done on designing the tracking capability from the perspective of supply chain. In this paper, using the traceable unit size at the manufacturer level to measure the tracking capability, we propose a non-convex non-linear programming to jointly optimise the tracking capability and price considering the tracking cost and recall cost in a supply chain with endogenous pricing. Results show that, in both centralised and decentralised supply chains, there is a unique tracking capability and retailing/wholesale price with closed-form solutions to optimise the supply chain profit. When the cost ratio (unit tracking cost/unit recall cost) is sufficiently large and small, the optimal tracking strategy is barcode tracking and unit tracking, respectively, and otherwise, the optimal tracking strategy is batch tracking with an economic traceable unit size which depends on the cost ratio, quality inspection threshold, supply defection rate and the supplier’s tracking capability. Furthermore, in the context of large and small cost ratio, we find that improving tracking capability will enlarge and mitigate the effect of double marginalisation, respectively. In particular, we find that the strict tracking regulation policy is more robust than the subsidy policy to improve the supply chain tracking capability.     

Human performance under two different command and control paradigms

The paradoxical behaviour of a new command and control concept called Network Enabled Capability (NEC) provides the motivation for this paper. In it, a traditional hierarchical command and control organisation was pitted against a network centric alternative on a common task, played thirty times, by two teams. Multiple regression was used to undertake a simple form of time series analysis. It revealed that whilst the NEC condition ended up being slightly slower than its hierarchical counterpart, it was able to balance and optimise all three of the performance variables measured (task time, enemies neutralised and attrition). From this it is argued that a useful conceptual response is not to consider NEC as an end product comprised of networked computers and standard operating procedures, nor to regard the human system interaction as inherently stable, but rather to view it as a set of initial conditions from which the most adaptable component of all can be harnessed: the human.     

Dielectric function of palladium capped zirconium thin films as a function of absorbed hydrogen

We report measurements of the optical transmission, between 240 and 1040 nm, and electrical resistivity of polycrystalline zirconium thin films as they absorb hydrogen. Both are measured as H₂ pressure is increased up to 880 mbar, at room temperature. Films, 20–22 nm thick, are deposited on fused quartz substrates by e-beam evaporation at 5.3 × 10^?7 mbar base pressure and covered with a 8.0 nm Pd over-coat. The morphology of the films is studied by means of AFM images. The complex refractive indices of Zr and Pd are extracted numerically from the transmission spectra by using a spectral projected gradient method for different hydrogen pressures. The corresponding dielectric functions for various Zr hydrogen concentrations are described with the parametric Drude-Lorentz and Brendel-Bormann (DL & BB) models. The Acceptance-Probability-Controlled Simulated Annealing approach is applied to calculate the parameters of the DL & BB model. This allows us to describe the effect of increasing hydrogen absorption on these parameters and in derived quantities, like the relaxation time and the effective mass of conduction electrons, the electrical resistance, the Fermi energy, and the electronic density of states at the Fermi level.     

Alternating current electrical field effects on lettuce (Lactuca sativa) growing in hydroponic culture with and without cadmium contamination

The aim of the present study was to investigate the effect of a 10 Hz alternating current (10 Hz 1 V cm⁻¹) and a 50 Hz alternating current (50 Hz 1 V cm⁻¹) on the lettuce plant (Lactuca sativa) growing in a hydroponic (soil-free) culture. Thirty lettuce plants were pre-germinated, and then 15 of them were treated with cadmium solution (CdCl₂) of 5 mg/L in concentration. Ten plants (five plants with Cd and five plants without Cd) were subjected to a 10 Hz alternating current (AC) electrical field; 10 plants were subjected to a 50 Hz AC field. The rest of the plants were used as a control. The lettuce plants were harvested after a growth of 60 days. The chlorophyll content, biomass and metal content of the lettuce plants were determined. The biomass of the plants growing in non-contaminated medium was 28 and 106% higher under the 10 and the 50 Hz AC fields respectively compared to the control. Although the plant biomass was reduced by the presence of Cd in the growth medium, the biomass of the plants growing in Cd contaminated medium was 40 and 63% higher respectively for 10 and 50 Hz AC field compared to the plant growing in Cd contaminated medium without electrical treatment. Increased uptake of Cd in the plant shoot was found with the 50 Hz AC field. Significant accumulation and uptake of Cu in plant roots and shoots was found under both electrical treatments.