Feature selection using Ant Colony Optimization (ACO) and Road Sign Detection and Recognition (RSDR) system

Road Sign Detection and Recognition (RSDR) is aimed to enable drivers maintain basic functionality with the aim of identifying and notifying driver through the existing restrictions so that the process is a success on the present widened road. Examples for RSDR include ‘traffic light ahead’ or ‘pedestrian crossing’ signs. An innovative RSDR system has been introduced which comprises of pre-processing, edge detection, feature extraction, features selection and Ensemble Fuzzy Support Vector Machine (EFSVM) classifier. Feature selection is carried out successfully by deployment of Ant Colony Optimization (ACO) algorithm to determine most prominent and definitive features. These features are then fed into the ensemble SVM to enable both road side traffic detection as well as recognition. Suggested system’s performance is analyzed and evaluated with respect to road signs having a capable recognition rate.     

Bicycle fatalities: Trends in crashes with and without motor vehicles in The Netherlands

Boufous and Olivier (2015) recently found an increasing trend of police-reported cyclist deaths in single-vehicle crashes while a decreasing trend was found for fatal bicycle-motor vehicle crashes. This study sets out to conduct the same trend analyses for the Netherlands using causes of death statistics. The results show that for 1996–2014, the number of cyclist deaths in motor vehicle collisions per km cycled (i.e. the risk) has declined by 3.8% per annum (95% CI: −4.6%, −3.0%) while the risk of cyclist deaths in crashes without motor vehicles increased by 7.0% per annum (95% CI: 5.5%, 8.7%). The large majority of cyclist deaths at the beginning of the study period were due to bicycle-motor vehicle crashes. By 2014, however, the share with motor vehicles was only slightly higher than the share without motor vehicles. We also conducted trend analyses controlling for age by stratifying into nine age groups. After controlling for age, the average annual decrease of cyclist deaths in crashes with motor vehicles was 5.4% (95% CI: −6.0% to − 4.7%). The average annual increase of deaths in crashes without motor vehicles was 4.4% (95% CI: 3.1–5.7%). These trends are more favourable than those without controlling for age which can be explained by more cycling among the elderly who have an elevated risk. We conclude that, similar to Australia, cyclist deaths following motor vehicle crashes in the Netherlands have decreased while deaths without motor vehicles have increased.     

The implications of low quality bicycle paths on the gaze behaviour of young learner cyclists

In a recent study, Vansteenkiste et al. (2014) described how low quality bicycle paths cause an apparent shift of visual attention from distant environmental regions to more proximate road properties. Surprisingly, this shift of visual attention was not accompanied by an adaptation in cycling speed. The current experiment investigated to what extent these findings are applicable for young learner bicyclists (aged 6–12 years). Since young learner bicyclists do not yet have mature visual and motor skills, it was expected that the implications of a poor road surface would be larger for them than for experienced adult bicyclists. In general, children looked less to the road and more to task irrelevant regions, but the magnitude of the shift of visual attention when cycling on a low quality bicycle track was similar to that of adults. Although children cycled slower than adults, they did not cycle slower on the low quality track compared to the high quality track. Overall, our results suggest that children displayed a different visual-motor strategy than adults, characterized by lower cycling speeds and a different visual behaviour, and that they responded in a similar way to a low quality bicycle path as adults.     

Age group differences in self-reported aggressive driving perpetration and victimization

Aim

This study examined whether the predictors of self-reported aggressive driving perpetration and victimization vary across age groups.

Method

Based on data from a general-population telephone survey conducted from July 2002 through June 2005, three groups of drivers were examined: 18–34 year-olds (N = 1522), 35–54 year-olds (N = 2726), and 55 years of age or older (N = 1883). For each age group sample, logistic regression analyses examined self-reported aggressive driving perpetration and victimization in the last 12 months by measures of driving exposure, heavy drinking, cannabis use, and drinking-driving, while controlling for demographic factors.

Results

The prevalence of aggressive driving perpetration within the past year was highest for the youngest age group (51%), followed by the middle-aged group (37%), and then the oldest age group (18%). The same pattern of results was found for prevalence of aggressive driving victimization (54%, 47%, and 31%, respectively). Controlling for demographic factors, the predictors of perpetration were generally consistent across the age groups. The logistic regression model for the youngest drivers revealed that those who reported stressful driving, heavy drinking, and cannabis use had significantly increased odds of reporting perpetration of aggressive driving. For middle-aged and older drivers, stressful driving, driving on busy roads, cannabis use, and driving after drinking were associated with perpetration. In addition, increased mileage contributed to perpetration in the oldest group. The findings for victimization by aggressive driving were similar. The logistic regression model for the youngest age group identified stressful driving, cannabis use, and higher annual mileage as being associated with victimization. For the oldest age group, these same variables were significant predictors of victimization, in addition to driving on busy roads. The logistic regression for the middle-aged group identified the same predictors as that of the oldest age group; however, interestingly driving after drinking was found to predict lower victimization among middle-aged drivers.

Conclusions

Although the prevalence of aggressive driving perpetration and victimization declined with age, the factors that contributed to aggressive driving remained generally stable across the lifespan. The results suggest that efforts to reduce aggressive driving among young drivers may prove to be effective for drivers from all age groups.     

An examination of message-relevant affect in road safety messages: Should road safety advertisements aim to make us feel good or bad?

Drawing upon the multiple roles of affect posited by Elaboration Likelihood Model, the current paper examines the effectiveness of message-relevant affect. Specifically, humourous and fear-evoking anti-drink driving messages are examined in terms of perceptions of relative influence on self and others (i.e., the third-person effect) and their performance on a range of persuasion outcomes. The influence of involvement, response efficacy, and gender on persuasion outcomes is also examined. Participants (N = 201) viewed two advertisements and completed two questionnaires: the first, assessed pre-exposure attitudes and behaviour and immediate-post exposure attitudes and intentions; the second, 2–4 weeks later, assessed attitudes and behaviour. The results revealed, as predicted, interactions of the key variables and evidence of the greater persuasiveness of negative appeals immediately after exposure whilst greater improvement of positive appeals over time. The findings highlight the importance of continuing the exploration of positive appeals as a persuasive alternative to negative appeals.     

Generality of a strength model for three conditions of repeated recall

A Markovian model of repeated recall is presented that is based on four memory processes: (a) a presentation increment that strengthens the memory trace of a word when the word is shown to a subject, (b) a recall increment that strengthens the trace when the subject successfully recalls the word, (c) a cueing decrement that weakens the trace as contextual cues for retrieving a word change over successive trials of attempted recall, and (d) a sampling rule that translates the strength of a memory trace into a probability of recall. Mathematically, the model predicts the relative frequencies of the recall patterns that are defined by sequences of recalls and forgets for individual words over trials. According to the model, the recallability of a word depends on its state, which is a function of (a) whether the word has been recalled for the first time, and if so, then (b) how many consecutive forgets of the word immediately precede the current trial, (c) how many successful recalls of the word have accumulated, and (d) how many previous trials of attempted recall have elapsed. Paradigms that were well described by the model were (a) no reminding, in which words were presented once and then recalled repeatedly without additional presentations; (b) restricted reminding, in which words were presented before each trial of attempted recall up to their first successful recall but not thereafter; and (c) complete reminding, in which all words were presented befor each attempted recall, whether previously recalled or not. Differences in recall among these conditions were explained in terms of differences in values of the model's parameters.     

Homotopic callosal inhibition

Previous theories concerning the function of the corpus callosum have included a “topographic” column-to-column excitatory theory and a diffuse “regional” inhibitory theory. Here it is shown that a topographic inhibitory model, in conjunction with a postulate concerning surround inhibition among cortical columns, produces complementary patterns of cortical column firing which, although qualitatively different, do not imply quantitative metabolic asymmetries. The cognitive implications of such homotopic inhibition are discussed.     

Central auditory processing : IV. Ear dominance—Spatial and temporal complexity

Ear dominance for dichotically presented tones was measured in 63 right-handed subjects when the frequency difference (Δf) was small compared to the center frequency (f_c) and again when it was large. Although two-thirds of the population exhibited a left-ear dominance in both conditions, a shift toward right-ear dominance occurred when the Δf was increased. An additional study, employing the alternating tone illusion described by Deutsch, revealed the same general effect, i.e., a shift toward right-ear dominance with increasing values of Δf/f_c. The results of these experiments, coupled with a review of previously published data of other dichotic experiments, indicate that as the ratio of Δf/f_c increases, the subjective complexity of the sound image increases, and there is a progressive emergence of a “right-ear advantage” (or ear dominance). A tentative explanation relates these results to the effects of anatomical asymmetries of primary and auditory association cortex and the efferent temporal lobe enhancement mechanism described by R. Efron, P. H. Crandall, B. Koss, P. L. Divenyi, and E. W. Yund (Brain and Language, 1983, 19, 254–263.     

Central auditory processing: I. Ear dominance—A perceptual or an attentional asymmetry?

The phenomenon of ear dominance for pitch described by Efron and Yund has been attributed by them to an asymmetry of sensory origin in the binaural integration of dichotic tone pairs. An explanation of this phenomenon in terms of an attentional bias is rejected on the basis of two experiments where the possibility of such bias was excluded. These and other experiments indicate that a simple explanation of this ear dominance in terms of a hemispheric specialization in the processing of tonal stimuli also must be rejected.