2.1. Setting Appropriate Speed Limits

Since the correlation between speed and road crash casualties is strongly evidenced [13, 34], it can be assumed that most casualties can be avoided if drivers comply with speed limits. In general, the “speed limit” refers to the highest speed at which drivers on the road in question can safely drive [35]. It can take two forms: static speed limits, which do not change based on the prevailing traffic and weather conditions, and dynamic speed limits (DSLs), which inform the driver by displaying the speed limit that might change considering the actual traffic and weather conditions. In both cases, speed limits are intended to accomplish two functions. Firstly, their primary function is to provide information on the appropriate speed and, therefore, reduce the hazards caused by driver speed selection [36]. If properly set, speed limits can support drivers’ estimation of the safe speed and remind them of the nature of the roadway and the risk level it poses to them and vulnerable road users [37]. Recent policy approaches further reinforce this principle. Initiatives such as the adoption of default 20 mph (30 km/h) speed limits in urban areas have demonstrated a reduction in road crashes, fatalities, and injuries [38]. This emphasizes the role of appropriately set speed limits in enhancing road safety.

Secondly, the other purpose of setting speed limits is to establish a foundation for enforcement and sanctions against individuals who exceed the limit and endanger others. However, drivers are generally observed to exceed the speed limit on roads. For example, according to a survey in Europe by the OECD [14], nearly half of the drivers travel faster than the speed limit, while 10%–20% of drivers travel at a speed of 10 km/h higher than the speed limit. For the majority of drivers to adhere to the speed limit, the road environment and the speed limit should ideally be consistent and coherent [39]. Therefore, setting appropriate speed limits consistent with the road environment and drivers’ expectations is one of the most important but difficult tasks to achieve improved speed-related outcomes.

Globally, speed limits are established based on various perspectives and approaches. Although there is no one-way approach that can be adopted to establish the appropriate speed limit for a given road section, the most commonly utilized approaches include engineering philosophy, harm minimization philosophy, optimization philosophy, and expert-based systems [40, 41]. Engineering philosophy is a two-stage process commonly used in New Zealand and Canada [40], where a base speed limit is established based on the 85^th percentile operating speed or the design speed, and it is then adjusted based on traffic and road conditions to determine the appropriate limit. However, this approach makes the unrealistic assumption that drivers choose their speed with adequate and objective consideration of road safety issues and is thought to cause a gradual increase in driver operating speed [41, 42]. On the other hand, in the harm minimization approach, also known as the safe system approach, speed limits are determined considering the types of collisions that are expected to occur, the resulting impact forces, and the human body’s capacity to tolerate these forces. It has been applied in some nations, with the most prominent examples being Sweden (Vision Zero) and the Netherlands (Sustainable Safety) [41]. Although the safe system approach to speed limit setting is conceptually appealing, its implementation may face financial hurdles. This is because the speed limits set based on this approach are mostly lower and require extensive enforcement and road modifications. In contrast, the optimization approach intends to establish speed limits to reduce societal costs, such as travel time, crashes, noise, pollution, and fuel usage [43]. All these cost items are considered to develop a total cost model as a function of the posted speed limit. Ultimately, the minimum total cost represents the lowest social cost of transportation under specific conditions. However, this approach is rarely employed, as the models are mostly complex to interpret and implement, requiring a specialized skill in mathematical modeling.

The expert-based system is the emerging approach to setting speed limits, supported by innovative models and tools that take into account the geometry, safety, and operational elements of roads, aiming to provide a unified and consistent approach to determining speed limits on road sections. It is a knowledge-based computer program that uses a method similar to the one used by experts to solve complex problems. This approach commonly employs various methods and algorithms for setting speed limits, including rule-based, machine-learning, and fuzzy logic approaches. Fuzzy logic, as highlighted in the work of References [44, 45], is a prominent approach because of its ability to handle uncertainty and imprecision in decision-making processes. Specifically, a fuzzy expert system that incorporates membership functions and fuzzy rules derived from expert evaluations was developed to set speed limits on Brazilian highways, demonstrating the effectiveness of fuzzy systems in this domain [41]. Rule-based and machine learning approaches are also employed in expert-based systems to set speed limits. The rule-based approach uses predefined rules, while the machine learning approach uses algorithms that help simplify decision-making processes [46]. The benefits of employing these systems are proven in practice, which demonstrates their robustness, consistency, and reliability, as well as their simplicity in replication across a variety of contexts [40]. The Australian XLIMITS application families, which were later developed for the United States and New Zealand, are notable examples of this method [47]. These models could be adopted in other countries after calibration, although they require a comprehensive data set encompassing road safety data, road geometry, traffic volume, and other characteristics of the road section, which may diminish their applicability in countries with poor road crash data management systems.

In conclusion, although speed limits are widely utilized and accepted globally, no standardized and concrete methodology is available that can support the selection of the appropriate speed limit for a specific road section. Although speed limit guidelines of various countries describe several criteria to be considered when selecting speed limits, including functional classification of roads, consistency and continuity in alignment, sight distance, intersection and interchange zones, crash data, vehicle type, load, and weather conditions [13, 48], they mainly do not provide a clear and objective methodology for quantifying appropriate speed limits. Setting speed limits based on such subjective criteria could lead to a lack of consistency and uniformity among the speed limits established by various practitioners. This may confuse drivers and cause them to lose faith in the whole speed limit system [49]. In this regard, although it is complex and requires further development, the expert-based approach seems ideal, as it utilizes a set of objective and readily observable criteria that characterize the road geometry, safety, and operating conditions to determine appropriate speed limits. Table 1 summarizes the strengths and weaknesses of the different approaches to speed limit setting.

2.2. Traffic Calming Measures (TCMs)

In areas where straight and/or wider roadway designs encourage speeding, interventions limited to speed limit signposting are not often successful in reducing vehicular speed and demand complementary measures. In such circumstances, the most common approach practiced by traffic engineers is the use of different self-enforcing physical measures such as speed bumps, speed humps, speed tables, roundabouts, chicanes, speed cushions, lane narrowing, and raised crosswalks, commonly known as “TCMs,” to calm vehicular speeds for the safety of all road users. The design of these measures aims to provide a road environment that forces road users to drive at a safe speed. Hence, their effectiveness is largely dependent not only on the type of TCM but also on the dimensions and spacing between consecutive devices. The effectiveness of TCMs can be evaluated using measurements of reductions in traffic speed, road crashes, noise pollution, and air pollution. However, crash and speed reduction measures are the most researched impacts of TCMs in the literature. Moreover, speed reduction is frequently used to evaluate the effectiveness of TCMs [18, 50, 51] because of the lack of reliable crash data to evaluate their safety effects. As speed is strongly correlated with the occurrence and severity of road crashes, it seems practical to use speed reduction as a surrogate measure of road safety when evaluating the effectiveness of TCMs.

A range of studies have explored the effectiveness of TCMs in reducing vehicle speed and improving road safety. Liu et al. [52] evaluated the impact of transverse rumble strips combined with a pedestrian crosswalk on average speed and crash frequency on rural roads with a speed limit of 60 km/h. The study has concluded that a 9.2 km/h reduction in average speed and about a 25% reduction in the frequency of road crashes can be achieved after the installation of rumble strips near pedestrian crossings on rural roads. Similarly, a detailed analysis by Afukaar et al. [15] revealed a 35% reduction in road traffic crashes and a 55% and 76% reduction in fatalities and serious injuries, respectively, after installing rumble strips on national roads. Obeng et al. [50] further supported these findings, demonstrating a significant impact of rumble strips in 36 spots with a spacing of 2.5 to 5 m on reducing vehicle speeds by up to 30% along highways. We may infer from these studies that installing rumble strips on national or rural highways will significantly reduce vehicle speed, even though their application in noise-sensitive areas has to be limited because of their potential to increase road noise.

Furthermore, other sets of traffic calming devices with a much better speed reduction effect are speed bumps and speed humps, widely used in several countries. The effectiveness of speed humps in achieving speed limit compliance by driver’s has been well proven [16, 27, 53–55]. For example, Mashrosa et al. [16] evaluated the impact of round and flat speed humps, and both were effective in reducing vehicle speed, with a reduction of 46% and 52%, respectively. Similarly, Bassani et al. [55] assessed the effectiveness of 12 trapezoidal speed humps in Italy at an average spacing of 14 m. For a given stretch of road, they reported that trapezoidal speed humps significantly reduce speed and improve the safety of vulnerable road users. Abdulmawjoud et al. [53] used vehicle data such as vehicle headways, travel durations, delays, and driving speed at different places to study the effectiveness of flat-topped humps, double bumps, and single bumps situated in Iraq and they found that such speed humps can reduce vehicle speed by 60% to 71%. More specifically, the study discovered that double-speed bumps spaced at 10 m distance from each other can reduce drivers’ speeds by 66%. However, the placement of speed bumps with such close spacing may increase the frequency of rear-end crashes [31], which may reduce the safety benefit of speed reduction unless a before-and-after crash analysis is conducted, although under-researched. In addition, speed bumps may also contribute to the process of pavement deterioration on roads with a high proportion of trucks because of induced vibrations and the bouncing suspension of trucks [56].

Moreover, road sections equipped with speed tables and pedestrian crosswalks were also found to significantly reduce the average speed of vehicles, the percentage of vehicles exceeding the speed limit, and pedestrian crashes [17, 57, 58]. A study by Damsere-Derry et al. [17] revealed that in areas with speed tables, the proportion of drivers exceeding the speed limit was less than 30%, while in areas without such devices, the percentage was higher than 60%. Hence, speed tables and raised crosswalks for pedestrians could increase safety by reducing vehicle speeds and improving how well drivers give way to pedestrians. These findings are consistent with the conclusions of several past studies [57, 58]. Therefore, installing these devices in high-risk areas can significantly enhance pedestrian safety and contribute to overall speed management efforts.

Besides, speed cushions, the modified versions of speed humps, are designed to minimize their effect on emergency response vehicles. After installing speed cushions in Italy, Berloco et al. [18] found a reduction of 35%–44% in the average vehicle speed and a 29%–39% reduction in the 85^th percentile operating speed of traffic. The use of speed cushions is considered more effective and versatile, as they have been found to significantly reduce traffic speed better than other TCMs [59], considerably decrease traffic volume [60], and improve road safety [61]. Moreover, speed cushions offer an advantage over other competent TCMs by allowing emergency vehicles to pass with minimal disruption, thus balancing speed control with essential mobility. These attributes make speed cushions a more valuable tool, particularly in areas with high traffic volumes and frequent emergency responses.

In conclusion, the effectiveness of various types of TCMs has been explored in several studies, which have reported the benefits associated with their implementation, such as reducing the average vehicular speed, the number of road crashes, overtaking frequencies, and the need for speed enforcement measures [15–18]. However, these measures can also have some negative impacts because of their potential to increase fuel consumption, environmental pollution, and emergency response time. In this regard, studies have shown that TCMs can increase fuel consumption by 40%–50% [28] and pollutant emission rates by 20%–60% [29]. In addition, frequent deceleration and acceleration maneuvers because of inadequate spacing between speed bumps and humps can lead to additional dangerous behavior as a consequence of a lack of speed uniformity [62]. According to Mohanty et al. [31], because of the faulty design of speed-reducing devices, drivers tend to decelerate frequently, which causes more wear and tear on the vehicle, leading to a higher probability of rear-end collisions. It is also attributed to the fact that the emergency response time can increase by 9.6 s for ambulances per TCM encountered [30].

Nevertheless, some alternatives can be used to minimize such limitations, such as the use of speed cushions instead of speed humps to reduce the delay experienced by time-critical emergency response vehicles. This is because cushions are designed to pass vehicles with a wider wheelbase, such as emergency service vehicles, relatively unhindered, targeting calm passenger vehicles. As a result, delays experienced by emergency vehicles because of speed cushions are expected to vary from zero to a few seconds [63], compared to delays of 1–20 s because of speed humps, raised intersections, and pedestrian crosswalks [64]. This illustrates how TCMs might be adapted to reduce the negative impacts on emergency response time; however, speed cushions can still be problematic for relatively small vehicles such as police cars, fire response vehicles, or even ambulances. Overall, the effectiveness of TCMs in improving road safety is supported by research; however, their implementation can lead to unwanted outcomes in terms of environmental pollution, delays, and pavement damage, specifically when poorly designed and located. Hence, further developments in the design and installation of TCMs are still required. Table 2 summarizes the strengths and weaknesses of the most common TCMs.

2.3. Speed Enforcement

Speed limit enforcement is a significant challenge for almost all countries worldwide. A study among OECD countries has shown that up to 80% of drivers were driving above speed limits, while a similar proportion of speeding drivers have been found in low- and middle-income countries [2, 65]. Apprehension and deterrence are fundamental objectives of speed enforcement, commonly accomplished by enacting traffic laws, enforcing them, and penalizing violators [66]. Hence, speed enforcement and sanctions are always needed to ensure compliance with speed limits because some drivers will always be noncompliant [67], and such measures are frequently reported to make significant contributions to safety improvements [19, 21, 68]. Furthermore, the effectiveness of speed enforcement relies not only on the consistency of law enforcement but also on the type of enforcement method employed and the severity of penalties, which can serve as a strong deterrent to potential violators.

As a countermeasure against speeding, various speed enforcement tools and methods have been used over the past decades, which can be broadly categorized as manual (man-based) and automatic (camera-based) systems. The former comprises surveillance through hidden police officers in a specific area (stationary approach) or the need to have them travel in a vehicle (mobile approach) [6]. In this approach, various stationary or mobile enforcement tools, such as photo radar, radar, and laser detectors, are frequently utilized to regulate vehicle speed. The latter requires a special camera equipped with speed measuring technology that can automatically record the speed of vehicles, take pictures of speeding cars, and send violation information straight to a ticketing office. It comprises conventional speed cameras, which measure vehicle speed at one point, and a SSES that determines the average vehicle speed over a long distance.

Speed enforcement by the police has been used for several years as a speed management strategy to influence drivers to comply with speed limits. Police enforcement for deterrence is instantaneous, as officers typically issue speeding tickets immediately [68]. Police officers often communicate educational messages to offenders to emphasize the risks of speeding, possibly resulting in a positive and long-lasting impact on their driving behavior [69]. This method has been considered an efficient strategy for discouraging violators and increasing safety [68]. Studies have shown that speed enforcement by police can reduce the number of speed violators, road crash occurrences, and fatalities. For example, Chen et al. [20] concluded that when speed enforcement programs were implemented, there was a significant reduction in speed limit violators, followed by a 25% decrease in speed-related crashes and a 17% reduction in road crash fatalities. Even better effects were observed in Australia by implementing randomized and scheduled police enforcement, where a 32% reduction in fatal and injury crashes was achieved [19]. However, the successful implementation of this approach depends on the consistent presence of police officers on the roads. Hence, its applicability is limited because of the extensive manpower requirement, the possible risk of crashes for police officers working in high-traffic environments, and the space requirement to park the vehicle.

In contrast, conventional speed cameras, commonly known as automatic speed cameras, were introduced to mitigate the limitations of police-based speed enforcement methods and reduce the requirement for on-site police presence. Generally, early studies reported a positive effect of deploying conventional speed cameras to increase the percentage of speeders apprehended and reinforce deterrence, leading to a decrease in travel speeds, crash rates, and fatality rates [21]. For example, implementing automated speed cameras in Austria reduced injury crashes and fatalities by 33.3% and 48.8%, respectively [22]. A similar study in Italy by Montella et al. [23] also reported a crash reduction of 19% for all crashes and 51% for fatal crashes. However, it has been claimed that drivers reduce their speed at locations with speed cameras and then accelerate afterward [70, 71]. Hence, the effectiveness of conventional speed camera enforcement systems has been criticized in light of such enforcement avoidance behavior, which increases the risk of road crash involvement because of sudden changes in driving speed close to enforcement areas.

To overcome such problems, a relatively advanced approach has been devised called the SSES, which is gaining popularity in several highly motorized countries. It is also known as point-to-point (P2P) speed enforcement, average speed enforcement, or trajectory control, depending on the jurisdiction in question. Unlike conventional speed camera enforcement methods, the SSES involves the determination of the average speed of each driver over a road section and hence encourages speed limit compliance over a longer distance. This type of speed enforcement system constitutes a series of speed cameras and detectors installed at several locations along the road. At each location, the system collects image and registration data for every crossing vehicle and matches them using automatic number plate recognition (ANPR) technology. Then, the average speeds of vehicles between two consecutive locations are calculated. An offense is considered committed when the average speed of a vehicle exceeds the speed limit stated in at least one of the sections under inspection, not just when it is traveling above the limit in a particular spot.

Numerous studies have consistently demonstrated that SSES is a promising strategy for improving road safety by reducing average vehicle speeds, decreasing speed variance, and enhancing overall traffic conditions [72–75]. For instance, implementing SSES on Korean expressways has led to an estimated 43% reduction in road crash occurrences, and the effect was ensured immediately after the installation of SSES [73]. Similar effects on average speed, speed variance, and road crash rate were reported because of the introduction of SSES in work zones [74] and on national roads [75]. Moreover, studies have also highlighted the ancillary benefits of implementing the SSES approach, such as homogeneous traffic flow because of lower speed variance [76], reduced fuel consumption, and reduced air pollution [77]. These findings consistently indicate that the SSES approach to speed enforcement is effective in encouraging safer driving behavior and improving overall road safety. Table 3 provides a summary of the strengths and weaknesses of the different types of speed enforcement strategies.

3.1. Intelligent Speed Adaptation (ISA)

Drivers are more likely to comply with the speed limit if they are aware of it and notified when they exceed it. In this regard, ISA is one of the most promising features of ADAS technologies that supports the driver by bringing the speed limit information to the vehicle’s dashboard. The system uses in-vehicle technology to continuously identify the vehicle’s position and employs sign recognition cameras to detect speed limit signs or speed limit databases in conjunction with GPS positioning to identify legal speed limits [78]. Then, it compares the vehicle speed to the speed limit at that locality and offers some in-vehicle feedback to the driver. Based on the type of in-vehicle feedback provided, the ISA system can take three different forms: advisory ISA, which provides an audio or visual alert to the driver when the speed limit is exceeded; voluntary (supportive) ISA, which prevents the driver from exceeding the limit but is overridable; and mandatory (limiting) ISA, which automatically limits the maximum travel speed of the vehicle to the speed limit and cannot be overridden. These ISA variants are expected to increase drivers’ compliance with speed limits while also increasing the credibility and effectiveness of speed enforcement, although differently [79].

However, fitting ISA devices in vehicles is not the only requirement for the technology to perform its function, assisting drivers to comply with speed limits. It requires infrastructure supportive of the technology in identifying the local speed limit in either of the two ways: (1) using a digital or cellular network to enable the vehicle to access speed limit database maps and to receive live updates about the speed limit or, in the absence of such cellular connectivity, (2) using an in-vehicle camera to enable the vehicle to detect the physical posted speed limit signs [80]. Therefore, in areas without cellular connectivity, the ISA system will depend on the physical speed limit sign infrastructure to perform its function. Hence, in specific areas where the physical speed limit sign is not posted and/or lacks cellular connectivity, the ISA technology will not be able to assist drivers in keeping to the limit. In general, state-of-the-art ISA systems employ a combination of camera detection systems, use up-to-date digital speed limit maps, and have a global navigation satellite system (GNSS), which has the highest real-life performance and reliability [32].

Although real-time data integration using speed limit information from cameras or GPS to ensure accurate and up-to-date speed advisory functions is commonly employed, various modeling and computational techniques have been used in the development and implementation of ISA systems. Fuzzy logic, rule-based systems, human-machine interaction models, simulation and modeling methods, and optimization techniques are frequently employed [81–84]. Fuzzy logic enables the system to handle imprecise data and make decisions based on degrees of truth rather than strict binary values [83], while rule-based systems offer a structured framework for establishing logical rules that regulate speed adaptation, taking into account factors such as speed limits and road conditions [84]. On the other hand, optimization techniques aim to improve the efficiency of ISA systems by finding optimal solutions for complex nonlinear processes [85]. Furthermore, it is imperative to incorporate human-machine interaction models to ensure driver acceptance and usability of ISA systems. This is evident in research that examines drivers’ behavior and acceptance of ISA systems [83]. By integrating these techniques, engineers could develop ISA systems that improve road safety and promote driver compliance with speed limits.

Recently, promising studies have been conducted on advancing the features of the ISA system to provide feedback to drivers when traveling at an unsafe speed for the prevailing sight conditions [82, 86, 87]. These are novel ISA systems designed to reduce the risk of crashes because of limited visibility along curves, named ISA for visibility (V-ISA). However, they are under development and are only tested in a simulated driving environment [86]. These new ISA systems are proposed in three variants with specific feedback modes: (1) V-ISA1, which displays visual information when drivers are driving at inappropriate speeds; (2) V-ISA2, which alerts the driver with an alert sound; and (3) V-ISA3, which intervenes directly to change vehicle speed [82].

Using a driving simulator, Hazoor et al. [86] examined the effectiveness of each of the three V-ISA variants with regard to driving speed decisions along road curves with limited sight distances. They employed a linear mixed-effect model on 60 vehicle drivers recruited for the research investigation. Accordingly, all V-ISA variants performed effectively in reducing vehicle speed at entry points, with no noticeable detrimental effect on driver lateral behavior. The V-ISA intervening variants were observed to be most successful in addressing sight limitations. The findings of these studies suggest that V-ISA may assist drivers in adjusting their operating speed based on the prevailing sight conditions, thereby creating safer conditions for driving. Further developments and evaluation of the new ISA system will be required to assess the implications of adopting the system, its acceptability, and its effect on drivers’ speed behavior on different types of roads and complex scenarios using a larger population dataset. Since the technology has never been implemented in actual vehicles [86], installing the V-ISA system on operational vehicles and field testing in a real road environment will be crucial to observe the impact of the system for full validation and general implementation on the design of new generation vehicles. Generally, research is also needed to evaluate the long-term effects of the V-ISA system on drivers’ behavioral adaptation phenomena and performance.

To sum up, research on the effectiveness of ISA systems on drivers’ speeding behavior has shown promising results, although their effectiveness varies depending on the type of ISA variant implemented. For example, an advisory form of ISA has been shown to help drivers avoid misjudgments and errors and improve driver compliance with speed limits [88, 89]. This type of ISA variant is commonly promoted because of its less intrusiveness and better acceptance potential by drivers, although the more intervening (limiting and supportive) form of ISA has a greater impact on compliance with the speed limit [32]. As a result, installing this variant of ISA technology in all new vehicles is being promoted in Europe and several other countries, such as Australia and the USA. If installed in all vehicles in Europe, the mandatory or limiting variant of ISA is predicted to reduce road crash occurrences and road crash deaths by 30% and 20%, respectively [90], while in Australia, these figures are estimated to be 30% and 28%, respectively [91]. Although studies have highlighted that drivers who participated in ISA studies have experienced some negative impacts, such as too much reliance on the ISA system [92] and speed compensation in locations where ISA is not active [14], generally, the ISA systems are found to have positive effects on road safety, such as a reduction in mean speed, speed variation, speed limit violations, and road crashes. Table 4 summarizes the strengths and weaknesses of the different variants of the ISA systems.

3.2. Variable Speed Limits (VSLs)

Conventional speed management strategies provide a static speed limit appropriate for the particular road section at average traffic, road, and weather conditions [93] but cannot indicate the safe speed for the prevailing traffic, road, and weather conditions. The emerging ITS solution overcomes such limitations and offers the DSL system intending to optimize traffic flow, enhance road safety, and adapt to changing conditions on the road. The system has been implemented in countries such as the United Kingdom, Australia, and the Netherlands and was observed to be capable of dynamically adjusting itself based on real-time traffic, road, and weather conditions [94]. As a result, the implementation of DSL systems represents a significant advancement in managing road traffic speed by providing more context-sensitive speed limits. By integrating real-time data, these systems enhance traffic safety and efficiency, further addressing the limitations of conventional static speed limits.

By definition, the terms “VSL” and “DSL” are often used synonymously, as both refer to speed limits that can vary based on specific road conditions compared to static posted speed limits. Nevertheless, there is still a distinction in emphasis or context between the two, as DSLs are more responsive and updated in real time based on current conditions, while VSLs may change, but the changes made are often scheduled or manually implemented based on more general considerations. DSLs function autonomously by employing algorithms to determine and set speed limits. These algorithms may span from basic logical operations and lookup tables to sophisticated and advanced models. However, the distinction is not often rigorously followed in the academic literature, and the term VSL is frequently used to represent the role of DSL. For the sake of consistency with the academic literature, in this context, when mentioning VSL further, we interpret it as an automated system where the speed limit on a specific road or highway can be dynamically adjusted based on the prevailing road, traffic, and weather conditions.

Although mostly limited to main roads, the performance of VSL systems has been evaluated in several studies in terms of reducing traffic congestion, fuel consumption, and improving road safety using different approaches, such as fuzzy logic [95], safety performance functions (SPFs) [96], and traffic theory-based heuristics [97]. In their carefully designed study, Coppola et al. [95] used a fuzzy logic-based VSL system to control speed limits in a connected traffic environment. The simulation analysis revealed that the proposed approach has improved traffic safety and efficiency by reducing speed variation, increasing average speed and traffic volume, and reducing fuel consumption. On the other hand, Hasan et al. [96] evaluated the impact of implementing VSL systems on road safety of highways using short-term SPF. They found a significant reduction in road crashes from 15.97% to 26.14% under different traffic conditions. Similarly, the implementation of a VSL system activated by roadside telecommunication antennas in a connected vehicle environment resulted in efficient control even at low market penetration rates (MPR) of connected and automated vehicles (CAVs) [97].

Furthermore, VSL systems have been found to have positive effects on aggregated drivers’ speed-related behavior, such as mean speed, speed difference, and the percentage of small space headways, leading to potential benefits to road safety. For example, Qu et al. [98] analyzed the safety impacts of the VSL system on such aggregated driving behavior and proved that the system can significantly decrease the mean speed, the speed difference, and the percentage of small space headways, indicating that potential traffic safety benefits can be achieved by changing the appropriate VSL values that match the prevailing traffic conditions. These emerging ITS solutions are helpful to reduce the burdens on drivers in determining the appropriate speeds suitable to the current environmental and situational circumstances [99] and significantly reduce fuel consumption and crash frequency [97]. Therefore, VSLs are expected to reflect the safe speed limits and improve the credibility of the speed limit system at large.

Moreover, recent developments in ITS and the emergence of CAVs would lead to the implementation of VSLs not only on main roads but also on other types of roads and the integration of VSLs with ISA systems for better safety performance. This will improve the limitations of the ISA system, which assists drivers by merely considering the static posted speed limits without considering the prevailing traffic volume, road layout, or environmental conditions. Integrating VSL and ISA involves using real-time information to control vehicle speed [81]. Hence, ISA systems can directly reduce a vehicle’s speed based on information obtained from the traffic control system installed on the road infrastructure. This information is transmitted to vehicles with ISA communication capabilities, allowing them to adjust their speed accordingly [100]. Exploring the potential of integrating VSL with ISAs is a novel area of investigation for researchers in the field of transportation. In this regard, Gámez Serna and Ruichek [101] proposed a dynamic speed adaptation (DSA) method that integrates VSL with ISA systems to automatically adjust vehicle speed based on road curvature and speed limit information. The study concluded that DSA improved road safety and passenger comfort by helping drivers respect speed limits and significantly reducing lateral errors on sharp curves. This integration will promote drivers’ compliance with speed limits and improve road safety.

In general, VSL systems have been demonstrated to improve traffic safety and operational efficiency in congested highways, particularly under adverse weather conditions [102, 103]. Although the effectiveness of VSL controls can be limited by the heterogeneity of human drivers [104], the integration of CAVs will enhance the performance of VSL systems by reducing driver heterogeneity and improving safety [105]. Studies have also shown that the implementation of VSL systems resulted in significant reductions in total crash counts, particularly rear-end crashes, and improvements in traffic speed differentials [106]. Despite the interest in evaluating the effect of the VSL and ISA systems, few studies have considered the potential benefits of integrating VSLs with ISAs in improving road safety and traffic management. Integrating the two systems could offer a more adaptive and efficient approach to managing road traffic speed by ensuring that drivers receive real-time speed recommendations while also being assisted in maintaining speed limit compliance.

3.3. Toward Intelligent Traffic Calming Devices

Although there is a great deal of research supporting several benefits of using TCMs in reducing the average speed of vehicles, the number of road crashes, overtaking frequencies, the need for speed enforcement, and relatively low installation and maintenance costs [27, 31, 54], they also have certain drawbacks. These include the risk of vehicle damage and subsequent maintenance costs, reduced driving comfort, delays for emergency and public transport vehicles, increased travel times and traffic congestion, problems with snow removal, the formation of pavement distresses before and after such devices, and an increase in fuel consumption, traffic noise, and greenhouse gas emissions when braking and accelerating [107]. However, given their lower installation and maintenance costs and simplicity of installation [108], traditional TCMs are frequently used and are likely to continue to be predominantly implemented.

Given these limitations, further developments in the application of ITS solutions for traffic calming are required to improve the shortcomings of conventional TCMs based on vehicle-to-vehicle and vehicle-to-infrastructure communications. Consequently, in recent years, there has been a significant effort to address the aforementioned issues through the development of intelligent speed bumps [109–111]. One such improvement is the development of dynamically active speed bumps, commonly called intelligent or adaptive bumps [111, 112], which activate and perform their usual functions when a speeding vehicle is detected. The intelligent bumps can identify the speed and vehicle type of the approaching traffic using some detection system, such as road loop sensors or radars. If a road user is detected driving above the speed limit, the bump will be activated to reduce vehicle speed, acting as a conventional speed bump, but if the approaching vehicle is complying with the limit, the bump will be inactive. There are also ways to adapt intelligent bumps to recognize selected vehicle types, such as emergency response vehicles, and allow these vehicles to pass without obstruction [112].

Previous research on modifications of speed bumps also proposed different designs exhibiting electrical, mechanical, and hydraulic structures [111, 113, 114]. Electronic speed bumps employ accelerometers to quantify the vibrations produced by the vehicle’s wheels as they go over them. It is then used to regulate the retraction of the speed bump. On the other hand, mechanical and hydraulic structures emphasize the methods for elevating and regulating the arched portions of the bump. Recently, the concept of using liquid-based speed bumps has been introduced by two manufacturers from Spain and England [115]. In this case, the degree of hardness of the liquid inside the speed bump is adjusted depending on the vehicle’s speed. The liquid bump serves as a mechanism to decelerate drivers exceeding the posted speed limit while allowing compliant drivers to pass without discomfort.

Moreover, a manufacturing company in Sweden has proposed an active dynamic speed bump, the so-called Actibump, which serves the same purpose and is activated by the speed of an approaching vehicle [116]. The detection system uses a speed radar or an inductive loop to estimate the speed of an incoming vehicle. The Actibump is positioned flat as long as the speed of the incoming vehicle is at or below the speed limit. The surface panel of the device obliquely sinks a few centimeters into the road if the driver is traveling at excessive speed. The system can also detect emergency vehicles and allow them to pass unhindered. These intelligent bumps have been implemented in Sweden, Iceland, Norway, and Australia in recent years [115], although their performance in achieving the desired outcomes is hardly evaluated. A recent study by Lin and Ho [111] introduced an intelligent speed bump system that performs adaptively using vehicle identification or vehicle speed to allow emergency vehicles and other vehicles driving within the speed limit to be unaffected by the speed bump. They designed and implemented a prototype speed bump system and highlighted the potential of implementing intelligent speed bumps for road infrastructure.

Although these developments are interesting, research in the academic literature regarding the performance of intelligent traffic calming devices is not comprehensive, and most of them have not been investigated in a real-world environment. In this regard, the effect of Actibumps on reducing vehicle speed was explored in a study conducted in Sweden by Nilsson [117], where the average speed reduction at different sites after installing the Actibumps ranged from 3.7 to 11.1 km/h. The percentage of drivers violating the 30 km/h speed limit was 75% and 21% before and after implementation, respectively. Overall, further research should investigate the development and performance of such intelligent traffic calming devices in a real-world environment. As research and technological advancements progress, such developments could lead to more intelligent traffic calming solutions that improve drivers’ speed limit compliance while minimizing unnecessary discomfort for road users who adhere to speed limits.