Behaviourally relevant road categorisation: a step towards self-explaining rural roads

An interpretable prediction model of illegal running into the opposite lane on curve sections of two-lane rural roads from drivers' visual perceptions

Illegal running into the opposite lane (IROL) on curve sections of two-lane rural roads is a frequently hazardous behavior and highly prone to fatal crashes. Although driving behaviors are always determined by the information from drivers' visual perceptions, current studies do not consider visual perceptions in predicting the occurrence of IROL. In addition, most machine learning methods belong to black-box algorithms and lack the interpretation of prediction results. Therefore, this study aims to propose an interpretable prediction model of IROL on curve sections of two-lane rural roads from drivers' visual perceptions. A new visual road environment model, consisting of five different visual layers, was established to better quantify drivers' visual perceptions by using deep neural networks. In this study, naturalistic driving data was collected on curve sections of typical two-lane rural roads in Tibet, China. There were 25 input variables extracted from the visual road environment, vehicle kinematics, and driver characteristics. Then, XGBoost (eXtreme Gradient Boosting) and SHAP (SHapley Additive exPlanation) methods were combined to build a prediction model. The results showed that our prediction model performed well, with an accuracy of 86.2% and an AUC value of 0.921. The average lead time of this prediction model was 4.4 s, sufficient for drivers to respond. Due to the advantages of SHAP, this study interpreted the impacting factors on this illegal behavior from three aspects, including relative importance, specific impacts, and variable dependency. After offering more quantitative information on the visual road environment, the findings of this study could improve the current prediction model and optimize road environment design, thereby reducing IROL on curve sections of two-lane rural roads.

Self-explaining roads: What does visual cognition tell us about designing safer roads?

In 1995, Theeuwes and Godthelp published a paper called "self-explaining roads," in which they argued for the development of a new concept for approaching safe road design. Since this publication, self-explaining roads (SER) became one of the leading principles in road design worldwide. The underlying notion is that roads should be designed in such a way that road users immediately know how to behave and what to expect on these roads. In other words, the environment should be designed such that it elicits adequate and safe behavior. The present paper describes in detail the theoretical basis for the idea of SER and explains why this has such a large effect on human behavior. It is argued that the notion is firmly rooted in the theoretical framework of statistical learning, subjective road categorization and the associated expectations. The paper illustrates some successful implementation and describes recent developments worldwide.

Procedure for the Identification of Existing Roads Alignment from Georeferenced Points Database

The aim of this research is to look for an automated, economical and fast method able to identify the elements of an existing road layout, whose original geometric design could date back to distant ages and could have undergone major modifications over the years. The analysis has been directed towards the Italian two-lane rural roads; the national public company ANAS made available its graph, obtained from high-performance surveys, that represents about 90% of these roads’ network. The graph is made up of a collection of georeferenced points but does not recognize or describe the geometric elements making up the roadway. Consequently, it has been necessary to design and develop an original procedure, subsequently implemented in a programming platform, able to identify the characteristics of the several parts, which constitute the reference axes of the existing roads. This research focuses on the horizontal geometry assessing the coherence, consistency and homogeneity of the roads’ layout, through the ex post application of the regulatory model for the design verification. If road sections are identified in which some conditions are not significantly met, further investigation should be conducted in order to ensure road safety and to plan any road upgrading activities.

Quantifying the Effects of Visual Road Information on Drivers' Speed Choices to Promote Self-Explaining Roads

Roads should deliver appropriate information to drivers and thus induce safer driving behavior. This concept is also known as “self-explaining roads” (SERs). Previous studies have demonstrated that understanding how road characteristics affect drivers’ speed choices is the key to SERs. Thus, in order to reduce traffic casualties via engineering methods, this study aimed to establish a speed decision model based on visual road information and to propose an innovative method of SER design. It was assumed that driving speed is determined by road geometry and modified by the environment. Lane fitting and image semantic segmentation techniques were used to extract road features. Field experiments were conducted in Tibet, China, and 1375 typical road scenarios were picked out. By controlling variables, the driving speed stimulated by each piece of information was evaluated. Prediction models for geometry-determined speed and environment-modified speed were built using the random forest algorithm and convolutional neural network. Results showed that the curvature of the right boundary in “near scene” and “middle scene”, and the density of roadside greenery and residences play an important role in regulating driving speed. The findings of this research could provide qualitative and quantitative suggestions for the optimization of road design that would guide drivers to choose more reasonable driving speeds.

Epidemiology of fatal crashes in an underdeveloped city for the decade 2008-2017

The road traffic accident is an important public health issue affecting social development and public safety. The purpose of this study is to describe the situation of traffic accidents, and to explore the relationship between fatal traffic accidents and risk factors in Suzhou city of Anhui province in China. The accident data was obtained from the Traffic Police Detachment of the Suzhou Public Security Bureau. Other vehicle, population and road information are derived from Suzhou Statistical Yearbook. Descriptive analysis was used to summarize road traffic accident participants' characteristics. The binary logistic regression model was used to analyze the factors that affected fatal traffic accident. There are 7,795 cases involved 18,774 road traffic participants. By logistic regression, we found that in Suzhou, overloading makes traffic accidents more likely to be the most fatal traffic accidents (adjusted OR = 5.33, 95%CI: 2.60-10.93). Speeding, drive after drinking, fatigue driving and pedestrian reasons might also increase the probability of fatal traffic accidents. These findings would help transportation authorities identify dangerous driving behaviors and take the necessary measures to improve road safety in undeveloped areas.

A driving simulation study to examine the impact of available sight distance on driver behavior along rural highways

The available sight distance (ASD) is the maximum length of the roadway ahead visible to the driver. It is a fundamental factor in road geometry principles and is used by road designers to ensure safe driving conditions. However, designers do not know how a specific ASD may affect the longitudinal and transversal behavior of drivers engaged in negotiating curves. This paper focuses on analyzing driver longitudinal behavior along rural highways curves with limited visibility. A number of virtual sight condition scenarios were recreated and tested in the driving simulator. Three tracks were designed with various combinations of radii and sight obstructions (a continuous wall) along the roadside located at various offsets from the lane centerline, combinations which resulted with a minimum ASD of 56.6 m. Roadside factors capable of influencing the risk perception of drivers (e.g., traffic barriers, posted speed limit signs, vegetation) were all excluded from the simulations. Results indicate that speed and trajectory dispersion from the lane centerline depend linearly on ASD in the investigated range of curve radii (from 120 to 430 m). In general, when ASD increases, so does speed and the trajectories tend to be less dispersed around the lane centerline. As a result, in safety terms, any variation in ASD will have the polar opposite effect on safety related parameters. Furthermore, different curves with similar ASD values resulted in different speed and lateral control behaviors. Analysis from ANOVA support the same findings; in addition, radius, curve direction, and distance from trajectory to sight obstruction have been identified as significant independent parameters. Road designers should adjust the ASD and these parameters when seeking to encourage drivers to adopt appropriate behaviors. To optimize safe driving conditions, ASD should be designed so that it is slightly greater than the required sight distance, since excessive ASD values may encourage drivers to drive at inappropriate speeds.

Quantifying visual road environment to establish a speeding prediction model: An examination using naturalistic driving data

Speeding is one of the major contributors to traffic crashes. To solve this problem, speeding prediction is recognized as a critical step in a pre-warning system. While previous studies have shown that speeding is affected by road environmental design, research in predicting speeding behavior through road environment features has not yet been conducted. Furthermore, there is a large discrepancy between actual and perceived road environmental information given that a driver's visual perception plays a crucial role as the dominant source of information in determining driver's behavior. Thus, this paper aims to establish a speeding prediction model based on quantifying the visual road environment to improve the design of pre-waring systems, which can predict whether drivers are going to speed and provide them with visual or/and audio warnings about their current driving speed and the speed limit prior to the occurrence of speeding behavior. Twenty input variables derived from three categories including visual road environment parameters, vehicle kinematic features, and driver characteristics were considered in the proposed speeding prediction model. Especially, the road environmental design factors consisting of the visual road geometry and visual roadside environment as perceived by the driver's eyes were quantified using a visual road environment model. Field experiments were conducted to collect naturalistic driving data concerning speeding behavior on the typical two-lane mountainous rural highways in five provinces of China. Random Forests, an ensemble learning method for regression and classification, were applied to build the speeding prediction model and variable importance was calculated. Additionally, logistic regression was used as a supplement to further investigate factors impacting on speeding behavior. A speeding criterion was defined with two levels in this study: a lower level (exceeding the posted speed limit) and a higher level (10% above the posted speed limit). Under both levels of the speeding criterion, the speeding prediction model performed well with high accuracy (over 85%). This model could use the value of the variables obtained from the current position to predict drivers' speeding behavior at the future position located a sighting distance away. This interval was sufficient for a pre-warning system to give a speeding warning that a driver with normal perception-reaction time (around 2.5 s) could respond to. Findings in this study can be used to effectively predict speeding in advance and help to reduce speeding-related traffic accidents.

Middle-aged Drivers' subjective categorization for combined alignments on mountainous freeways and their speed choices

Road geometric design is a fundamental factor that impacts driving speed. Previous research generally paid attentions to the influences of specific road characteristics (e.g. curvature) on driving behaviors. Limited studies have focused on how drivers identify different alignments and how they further take the varying speed choices. This study aims at filling the gap by investigating the subjective categorization of road alignments based on middle-aged driver groups. A total of sixteen participants with ages ranging from 23 to 40 years were recruited. Participants were first asked to undertake naturalistic driving tests on a four-lane divided mountainous freeway while photos of the road and the driving speed were collected. Participants were then asked to subjectively sort the photos of the road into piles, within each pile we considered their driving behaviors would be similar. Finally, questionnaire survey was conducted in terms of comfort, safety, speed choice and sight distance. The picture grouping revealed three distinct and non-overlapping subjective categories of road alignment. And driver's ratings about comfort and safety were significantly different between these categories. The category with the largest sight distance and highest speed choice turned out to have the lowest rating in comfort and safety (note that the rating scales for comfort and safety had reversed polarity such that low numbers indicated high comfort and high safety). Statistical evidences indicated that the drivers have developed underlying mental schema about road alignment. Therefore, their speed choices on combined alignment were further investigated. The difference between actual driving speed and driver's expected speed showed close relation to the ratings and significant difference between two of the categories. Road with large absolute value in speed difference informed inconsistency between geometric design and driver's expectation from the aspects of drivers' perception and expectation of the road. The findings provided insight into how middle-aged driver views and categorizes road alignment. And it was found that the drivers relied on visual characteristics of the alignment to distinguish the categories instead of separate horizontal and vertical geometric parameters. It was implied that more considerations should be taken into driver's perception of road during alignment design to improve road safety.

Effect of novel divergence markings on conflict prevention regarding motorcycle-involved right turn accidents of mixed traffic flow

INTRODUCTION

In Taiwan, segregated traffic flow countermeasures have long been in place. Although these facilities have decreased the numbers of motorcycle left-turn collisions, right-angle collisions, and sideswipe collisions, they have also induced serious right-turn accidents. The purpose of this research was to evaluate an intervention intended to decrease conflicts and motorcycle-involved crashes. In this study, the reasons why the motorcycle accident rate is higher at intersections with slow lanes than at those without slow lanes are presented, and the theory of the self-explaining road was applied to create divergence markings for a mixed traffic flow environment. An intervention that guides motorcycles and cars into appropriate locations at intersections was applied to three intersection approaches.

METHOD

The intervention effectiveness was evaluated by comparing the number of accidents at the intersections before and after the implementation of improvement measures. Moreover, video recordings were used to analyze the traffic distributions at the cross-sections of intersections. T-test was adopted to examine whether the traffic flows at the cross-sections of the intersections before and after the intervention were statistically different. In addition, this research applied the post-encroachment time (PET), the time between the first road user leaving the encroachment zone and the second road user arriving in it, to evaluate traffic conflicts. Finally, the PET and severity index between a straight-through motorcycle and a right-turn vehicle were analyzed.

RESULTS

PET increased by 3.2%-20.4%, and the rates of right-turn collisions, sideswipe collisions, and rear-end collisions decreased by 64.3%, 77.3%, and 61.5% respectively.

CONCLUSIONS

Eliminating the slow traffic lane and setting divergence markings may not effectively cause vehicles in different driving directions to drive in the proper locations in the lanes. However, divergence markings both reduce the rate of right-turn collisions and decrease the incidence of sideswipe and rear-end collisions. Practical applications: The proposed design method may be a good design reference for countries having a high motorcycle density.

Impact of perceptual countermeasures on driving behavior at curves using driving simulator

OBJECTIVE

The probability of crash occurrence on horizontal curves is 1.5 to 4 times higher than that on tangent sections. A majority of these crashes are associated with human errors. Therefore, human behavior in curves needs to be corrected.

METHODOLOGY

In this study, 2 different road marking treatments, optical circles and herringbone patterns, were used to influence driver behavior while entering a curve on a 2-lane rural road section. A driving simulator was used to perform the experiment. The simulated road sections are replicas of 2 real road sections in Flanders.

RESULTS

Both treatments were found to reduce speed before entering the curve. However, speed reduction was more gradual when optical circles were used. A herringbone pattern had more influence on lateral position than optical circles by forcing drivers to maintain a safe distance from opposing traffic in the adjacent lane.

CONCLUSION

The study concluded that among other low-cost speed reduction methods, optical circles are effective tools to reduce speed and increase drivers' attention. Moreover, a herringbone pattern can be used to reduce crashes on curves, mainly for head-on crashes where the main problem is inappropriate lateral position.

Influence of drivers' psychological risk profiles on the effectiveness of traffic calming measures

Road traffic injuries represent a serious public health problem and are one of leading cause of death, injury and disability around the world. Road accidents are often caused by an accumulation of factors; however, drivers appear to be by far the most decisive one. The driver's behaviour is complex and depends on reflex (or involuntary) and voluntary driving actions. The first class of actions (reflex actions) are typical human reactions that remain inaccessible to awareness and refer to the direct interaction between the road user and the characteristics of the road and its surrounding environment. Conversely, voluntary actions are conscious behaviours adopted on the basis of planned decisions. Both types of driving actions act simultaneously and the interaction between them and their relative effects on road safety are an aspect not yet well examined. The main objective of this study was to provide, by means of a driving simulation experiment, an insight on this interaction by evaluating the influence of some psychological characteristics on the effectiveness of different types of traffic calming measures at pedestrian crossings, designed according to the Human Factors principles. Fifty-eight participants drove a virtual urban route while data on their performance, as they approached five configurations of pedestrian crossings equipped with different physical and perceptual treatments, were collected. The participants were preliminarily characterized by means of two psychological questionnaires, which allowed the identification of three distinct groups of drivers belonging to three risk profiles (careful, worried, and at risk). The three groups of drivers reacted differently to the proposed engineering treatments, confirming the clustering identified by the preliminary analysis. The results showed that the proposed traffic calming measures are effective on all psychological sub-groups of drivers, with different effectiveness. These first results support that, in the considered driving environment (pedestrian crossings), the Human Factors approach, with which traffic calming measures can be arranged, could be effective, even if different psychological sub-groups are differently affected.

Transitions within a safe road system

As drivers move through the road transport system they are exposed to a range of different situations and road conditions in a relatively short space of time. Drivers' expectations about what will happen on different types of roads have strong effects on their speed choices, and where they look and what they attend to. As a result it is important to assist drivers to change their expectations when they transition from one road type to another. In this experiment we investigated the effectiveness of different centreline road markings in preparing for a horizontal curve as drivers moved from a motorway to a two-lane rural country road. Fifty individuals were recruited to participate in a video-based simulated driving task to compare three centreline marking types in terms of their effects on speed choice and reactions to a driving hazard (horizontal curve). Although a complex marking previously associated with high risk produced the largest speed reductions during the transition from the motorway, it was the centreline more traditionally associated with rural country roads (dashed white centreline) that was associated with the best hazard reactions post-transition (brake reaction time and speed reduction before a horizontal curve). The findings demonstrated that the look of a road needs to convey a clear and unambiguous message to drivers. The transition to a two-star rural road is best achieved by making the road look like a typical two-star road as soon as possible.

Controlled before-after intervention study of suburb-wide street changes to increase walking and cycling: Te Ara Mua-Future Streets study design

BACKGROUND

Achieving a shift from car use to walking, cycling and public transport in cities is a crucial part of healthier, more environmentally sustainable human habitats. Creating supportive active travel environments is an important precursor to this shift. The longevity of urban infrastructure necessitates retrofitting existing suburban neighbourhoods. Previous studies of the effects of street changes have generally relied on natural experiments, have included few outcomes, and have seldom attempted to understand the equity impacts of such interventions.

METHODS

In this paper we describe the design of Te Ara Mua - Future Streets, a mixed-methods, controlled before-after intervention study to assess the effect of retrofitting street changes at the suburb scale on multiple health, social and environmental outcomes. The study has a particular focus on identifying factors that improve walking and cycling to local destinations in low-income neighbourhoods and on reducing social and health inequities experienced by Māori (Indigenous New Zealanders) and Pacific people. Qualitative system dynamics modelling was used to develop a causal theory for the relationships between active travel, and walking and cycling infrastructure. On this basis we selected outcomes of interest. Together with the transport funder, we triangulated best evidence from the literature, transport policy makers, urban design professionals and community knowledge to develop interventions that were contextually and culturally appropriate. Using a combination of direct observation and random sample face to face surveys, we are measuring outcomes in these domains of wellbeing: road-user behaviour, changes to travel mode for short trips, physical activity, air quality, road traffic injuries, greenhouse gas emissions, and perceptions of neighbourhood social connection, safety, and walking and cycling infrastructure .

DISCUSSION

While building on previous natural experiments, Te Ara Mua - Future Streets is unique in testing an intervention designed by the research team, community and transport investors together; including a wide range of objective outcome measures; and having an equity focus. When undertaking integrated intervention studies of this kind, a careful balance is needed between epidemiological imperatives, the constraints of transport funding and implementation and community priorities, while retaining the ability to contribute new evidence for healthy, equitable transport policy. The study was retrospectively registered as a clinical trial on 21 June 2018 in the ISCRTN registry: ISRCTN89845334 http://www.isrctn.com/ISRCTN89845334.

Understanding international road safety disparities: Why is Australia so much safer than the United States?

Despite similarities to the US in terms of transportation, land use, and culture, Australia kills 5.3 people per 100,000 population on the roads each year, as compared to the US rate of 12.4. Similar trends hold when accounting for distance driven and the number of registered cars. This paper seeks to understand what is behind the road safety disparities between these two countries. The results suggest that a number of inter-related factors seem to play a role in the better road safety outcomes of Australia as compared to the US. This includes Australia's strategies related to seat belt usage and impaired driving as well as their efforts to help curb vehicle speeds and reduce exposure. Design-related differences include a much greater reliance on roundabouts and narrower street cross-sections as well as guidelines that encourage self-enforcing roads. Policy-related differences include stronger and more extensive enforcement programs, restrictive licensing programs, and higher driving costs. Combined with a more urban population and multimodal infrastructure, Australia tends to discourage driving mileage and exposure while encouraging safer modes of transportation such as transit, at least more so than in most of the US. Australia also enacted their version of Vision Zero - called the Safe System Approach - more than a decade before similar policies began cropping up in US cities. While it is difficult to attribute recent road safety successes to any specific policy, Australia continues to expand their lead on the US in terms of safety outcomes and is a road safety example worthy of consideration.

Driving on urban roads: How we come to expect the 'correct' speed

The subjective categories that drivers use to distinguish between different road types have been shown to influence the speeds they choose to drive but as yet we do not understand the road features that drivers use to make their discriminations. To better understand how drivers describe and categorise the roads they drive, 55 participants were recruited to drive a video of familiar urban roads in a driving simulator at the speed they would drive these roads in their own cars (using the accelerator and brake pedal in the driving simulator to adjust their speed). The participants were then asked to sort photos of the roads they had just driven into piles so that their driving would be the same on all roads in one pile but different to the other piles. Finally, they answered a series of questions about each road to indicate what speed they would drive, the safe speed for the road, their speed limit belief as well as providing ratings of comfort, difficulty and familiarity. Overall, drivers' categorisation of roads was informed by a number of factors including speed limit belief, road features and markings (including medians), road width, and presence of houses, driveways and footpaths. The participants' categories were congruent with what they thought the speed limits were, but not necessarily the actual speed limits. Mismatches between actual speed limits and speed limit beliefs appeared to result from category-level expectations about speed limits that took precedence over recent experience in the simulator. Roads that historically had a 50km/h speed limit but had been reduced to 40km/h were still regarded as 50km/h roads by the participants, underscoring the point that simply posting a sign with a lower speed limit is not enough to overcome drivers' expectations and habits associated with the visual appearance of a road. The findings provided insights into how drivers view and categorise roads, and identify specific areas that could be used to improve speed limit credibility.

Traffic and Environmental Cues and Slow-Down Behaviors in Virtual Driving

This study used a driving simulator to investigate whether the presence of pedestrians and traffic engineering designs that reported to have reduction effects on overall traffic speed at intersections can facilitate drivers adopting lower impact speed behaviors at pedestrian crossings. Twenty-eight men (M age = 39.9 yr., SD = 11.5) with drivers' licenses participated. Nine studied measures were obtained from the speed profiles of each participant. A 14-km virtual road was presented to the participants. It included experimental scenarios of base intersection, pedestrian presence, pedestrian warning sign at intersection and in advance of intersection, and perceptual lane narrowing by hatching lines. Compared to the base intersection, the presence of pedestrians caused drivers to slow down earlier and reach a lower minimum speed before the pedestrian crossing. This speed behavior was not completely evident when adding a pedestrian warning sign at an intersection or having perceptual lane narrowing to the stop line. Additionally, installing pedestrian warning signs in advance of the intersections rather at the intersections was associated with higher impact speeds at pedestrian crossings.

Road user behaviour changes following a self-explaining roads intervention

The self-explaining roads (SER) approach uses road designs that evoke correct expectations and driving behaviours from road users to create a safe and user-friendly road network. Following the implementation of an SER process and retrofitting of local and collector roads in a suburb within Auckland City, lower speeds on local roads and less variation in speed on both local and collector roads were achieved, along with a closer match between actual and perceived safe speeds. Preliminary analyses of crash data shows that the project has resulted in a 30% reduction crash numbers and an 86% reduction in crash costs per annum, since the road changes were completed. In order to further understand the outcomes from this project, a study was carried out to measure the effects of the SER intervention on the activity and behaviour of all road users. Video was collected over nine separate days, at nine different locations, both before and after SER construction. Road user behaviour categories were developed for all potential road users at different location types and then used to code the video data. Following SER construction, on local roads there was a relatively higher proportion of pedestrians, less uniformity in vehicle lane keeping and less indicating by motorists along with less through traffic, reflecting a more informal/low speed local road environment. Pedestrians were less constrained on local roads following SER construction, possibly reflecting a perceptually safer and more user-friendly environment. These behaviours were not generally evident on collector roads, a trend also shown by the previous study of speed changes. Given that one of the objectives of SER is to match road user behaviour with functionally different road categories, the road user behaviour differences demonstrated on different road types within the SER trial area provides further reinforcement of a successful SER trial.

Supporting drivers in forming correct expectations about transitions between rural road categories

In order to support drivers in forming the right expectations on the road, road categories are being made recognisable and predictable in the Netherlands. The present study investigated which of the selected road layouts can make rural road categories most recognisable for road users, especially in transitions from one road category to another. A second objective was to study whether explicit information could contribute to a better recognisability of transitions. The experiment was performed with a series of photographs showing sections of two road categories with an intersection in between. The road layout of road categories varied in markings and separation of driving direction (within-subjects factor). Informed and non-informed participants (between-subjects factor) had to indicate their expectations regarding speed limit and access restriction of each road section, before and after a transition. The results show that for transitions between distributor and through roads, the physicality of separation of driving direction is a better distinctive characteristic than the currently used edge marking. The green centre marking on through roads also enhances recognisability, but only with additional information. As far as transitions between distributor and access roads are concerned, the results demonstrate that this type of transitions is better recognised when no markings on access roads are present. Physical separation of driving directions on distributor roads also improves recognisability, although this layout is associated with higher speed limits. Providing explicit information has in general a positive effect on the reconisability of transitions. Implications are discussed in the light of potential safety effects.

Using endemic road features to create self-explaining roads and reduce vehicle speeds

This paper describes a project undertaken to establish a self-explaining roads (SER) design programme on existing streets in an urban area. The methodology focussed on developing a process to identify functional road categories and designs based on endemic road characteristics taken from functional exemplars in the study area. The study area was divided into two sections, one to receive SER treatments designed to maximise visual differences between road categories, and a matched control area to remain untreated for purposes of comparison. The SER design for local roads included increased landscaping and community islands to limit forward visibility, and removal of road markings to create a visually distinct road environment. In comparison, roads categorised as collectors received increased delineation, addition of cycle lanes, and improved amenity for pedestrians. Speed data collected 3 months after implementation showed a significant reduction in vehicle speeds on local roads and increased homogeneity of speeds on both local and collector roads. The objective speed data, combined with residents' speed choice ratings, indicated that the project was successful in creating two discriminably different road categories.