Fig. 1 presents a conceptual design of next-generation highways in cold climates, where a number of CI entities are envisioned. These, along with V2X and ICT capabilities, are expected to greatly enhance the spatial and temporal resolutions of road weather data and thus better inform stakeholders such as highway operators, emergency responders, truck drivers and other highway users. They can also augment the sensing capabilities and confidence level of CVs/AVs, reduce the cost and uncertainties of sensing, enrich the sources of information for both CVs/AVs and unconnected vehicles, and improve the response time. The following sections will further discuss some entities, including: energy-harvesting, connected roadways and roadside infrastructure assets (RIAs); self-sensing or anti-icing pavement; connected Environmental Sensing Stations (ESS) and connected FAST system, followed by ICT capabilities. Other CI entities will not be discussed in detail, including: weather-responsive traffic signals, smart work zone equipment, and non-contact static charging for vehicles. In this work, we also do not discuss AV-enabling infrastructure components, such as smart signage for automatic driving, standardized pavement markings for machine vision, and magnetic nails and reflective striping for lane-keeping.
Energy-harvesting, connected roadways and RIAs
The massive number of roadway mileages and vast amount of lands in the right-of-way present a great opportunity to capture and utilize the energy dissipated from the ambient roadway system, such as mechanical energy from vehicle or wind loadings and thermal energy from the sun or earth. Such energy harvesting is particularly beneficial for roadways in remote, off-grid areas where the lifeline of CI applications is endangered by the lack of access to power. In the U.S. alone, there are 2.6 million miles of paved roads and highways, of which approximately 93 percent has an asphalt surface [Mohamed Jaafar 2019] and 3,000 (linear) miles are equipped with noise barrier [Poe et al. 2017]. The “(unpaved) land cover in close proximity to the National Highway System” in the U.S. has been estimated to be “roughly 68 percent, or 3.4 acres” [Earsom et al. 2010]. A suite of energy-harvesting technologies are available, as illustrated in Fig. 2 [Wang et al. 2018]. These technologies entail the use of piezoelectric materials, micro wind turbines, photovoltaic panels or solar cell roads, geothermal heat pumps, or pipe-pavement thermoelectric generator (PP-TEG) system. They could be potentially incorporated into either the connected roadways such as pavements and traffic signals, or RIAs such as noise barriers and structural snow fences, featuring a wide variety of cost, energy output and efficiency, service life, dimensions, maintenance requirements, recyclability, and other characteristics.
The last five years have seen increased interest and promising progress in demonstrating the use of piezoelectric materials to harvest deformation energy from asphalt pavement. Yet, the amount of energy harvested is limited but suitable for applications such as “powering wireless sensors embedded into pavement structure” [Roshani et al. 2017] and other microelectronics, “heating road surface on bridge deck for anti-icing, lighting, or powering traffic devices” [Wang et al. 2018]. A few representative advances in the development of piezoelectric energy harvester (PEH) technology is summarized in Table 1.
Wang et al. [2018] reviewed the energy-harvesting technologies in roadway and bridge applications. Technologies other than PEHs also have their own strengths and limitations. For instance, photovoltaic systems can produce high energy output but their use in roadways may complicate vehicle operations or pose a risk to traffic safety, and more research is needed to address such concerns. Geothermal heat pumps are considered a mature technology, which is “geologically and geographically limited” and most appropriate for safety-critical areas. PP-TEG produces a low energy output at high cost, and more research is needed to improve the system efficiency.
Table 1
Recent advances in PEH technology for roadways
Configuration | Optimal energy performance | Other performance considerations | Reference |
A stacked configuration of transducers | Not reported | Able to support loadings up to 150 kN and remained effective after 100,000 cyclic loadings | Yang et al. [2017] |
Bridge transducer with layered poling and electrode design | Output power: 2.1 mW at a resistive load of 400 kΩ, under 70 kPa (compression loading), 5 Hz (vibration frequency), output voltage: 556 V | Balanced the desire for improved energy output and the need for less risk of stress concentration | Jasim et al. [2017] |
Several PEH prototypes to be embedded into asphalt pavement | Output power: > 25 mW under the compression of 15 MPa (for each piezoelectric element) and 10 Hz; output voltage (nearly 20 V) and output current (> 100 µA) at a load of 3 kN | Higher frequency “represents higher traffic speed and greater traffic volume” (e.g., on Interstate highways), and leads to better output power. | Roshani et al. [2017] |
Several most commonly installed PEH technologies (surveyed) | Average output power: 3.1 mW (based on on-site evaluation) | One challenge was “only 14.43% of the applied loading was transmitted to the piezoelectric materials” and the electrical productivity of PEHs highly depends on “the axle configuration and magnitude of passing vehicles”. | Xiong and Wang [2016] |
One layer of “piezoelectric elements with a higher piezoelectric stress constant” and two layers of “more flexible conductive asphalt mixtures” | Output power: ranging from 1.2 mW to 300 mW at 30 Hz | The cost of electricity produced by this PEH can be as low as “$19.15/kWh at a high-volume roadway within a 15-year service life”. | Guo and Lu [2017] |
In addition to smart pavements, there are a number of conceptual scenarios that energy-harvesting technologies may be incorporated into the roadway infrastructure. For instance, portable micro wind turbines could be mounted on structural snow fences, traffic signals, and so on. Being a cost-effective technology to prevent blowing and drifting snow, snow fences can improve road safety and provide additional benefits, if designed and sited properly [Du et al. 2017]. Nabavi and Zhang [2016] reported three groups of portable wind energy harvesters, i.e., piezoelectric-, electromagnetic-, and electrostatic-based generators, with different wind-flow-trapping mechanisms and varying dimensions and energy conversion efficiencies. Photovoltaic panels can be strategically installed either in the right-of-way providing sufficient space for such a “distributed solar power plant” [Asanov et al. 2019], or integrated with noise barriers to produce a considerable amount of renewable energy [Poe et al. 2017]. Qiao et al. [2011] proposed the concept of a “smart microgrid that optimally utilizes the public right-of-way and roadway infrastructure to provide cost-effective, highly efficient, and reliable wind/solar electric power production, distribution, storage, and utilization”. The design entails a “grid-connected wind/solar hybrid generation system installed on the pole of a roadway/traffic signal light”. Considering the great temperature difference between the air and the relatively warm soil beneath pavement, another potential technology to explore is thermoelectric generators. With the thermal gradient in the opposite direction, this concept was demonstrated in south Texas [Datta et al. 2017] where a TEG prototype produced “an average of 10 mW of electric power continuously over a period of 8 h”.
Self-sensing or anti-icing pavement
Self-sensing pavements can serve as an integral part of connected roadway infrastructure or VII System, and the related research is still in the burgeoning stage. Han et al. [2013] reported the traffic detection performance of a self-sensing pavement being tested at the Minnesota Road Research Facility, USA, both in winter and in summer. This smart pavement was enabled by the admixed carbon nanotubes in concrete, and was able to “accurately detect the passing of different vehicles under different vehicular speeds and test environments”. Relative to conventional strain gauges, this self-sensing concrete exhibited advantages in its ease of installation and maintenance, compatibility with pavement structures, and durability. Liu et al. [2014] reported an exploratory study that suggests the potential use of conductive asphalt materials for self-sensing applications, because the different stages of damage evolution corresponded to certain change patterns in their electrical resistivity.
Anti-icing pavements are often not designed for CI applications, but they can contribute to the resilience of surface transportation system during snowy weather. A variety of technologies have been explored to enable anti-icing pavements, ranging from “anti-freezing pavements that rely on physical action, to high-friction in situ anti-icing polymer overlays, to asphalt pavements containing anti-icing additives, to heated pavements using energy transfer systems” [Shi et al. 2018]. All of them aim to “prevent or reduce the bond of ice or compacted snow to pavement or to prevent or treat winter precipitation”. Pan et al. [2015] and Zhang et al. [2020] presented a comprehensive review on the use of conductive and salt-releasing asphalt mixtures as anti-icing pavements, respectively. Note that none of these technologies have been widely adopted by transportation agencies, and this is mainly due to concerns over their long-term performance.
Connected ESS
RWIS-ESS could be further enhanced so as to serve as integral part of the CI solution and of the larger vehicle-infrastructure ecosystem, with the focus on design improvements and cost reductions. RWIS refers to “networks of ESS that observe the near-surface atmosphere and pavement surface and subsurface” [Albrecht et al. 2018], generally deployed at fixed locations to better inform WRM operations. Each ESS consist of various sensors that provide site-specific, real-time data on the meteorological conditions, RSC, and subsurface temperature, which altogether enables pro-active WRM practices such as anti-icing, improves the roadway level of service and resource allocation, and enhances traveler information, traffic management and emergency response [Strong and Fay 2007, Kwon et al. 2017]. Kwon et al. [2017] developed a framework to optimize the siting of RWIS stations in a given network to maximize “the coverage of accident-prone areas (while) minimizing the total estimation error”, addressing the needs by both WRM operators and the traveling public. Note that the mobile data collection by CVs/AVs will likely bolster the accuracy of road weather forecasting models, which in turn induces advances in the RWIS-ESS technology.
Connected FAST system
FAST has been employed by many roadway agencies to prevent or pro-actively mitigate black ice or bonding of compacted snow and ice to pavement (or bridge deck). Conceptually, the design of FAST system could be further enhanced to improve its connectedness and thus contribute to the vehicle-infrastructure ecosystem. Ye et al. [2013] surveyed the state of the art of FAST systems and confirmed their potential in delivering substantial benefits, such as less need for mobile operations and WRM materials and reduced crash frequency and traffic delay. This technology “works best for frost and light snow events”, but its application has been hindered by challenges in sensor malfunctioning, system maintenance, and training. A study by Veneziano et al. [2015] examined the safety effects of FAST systems operated by the Colorado Department of Transportation and recommended the deployment of such system at “high-traffic, high-crash severity locations”. For instance, FAST systems were able to contribute to “an annual reduction of 16–70% on urban Interstates, 31–57% on rural Interstates, and 19–40% on interchange ramps between Interstates”, when sited and operated properly.
Arguably, there are a number of roadside or CV solutions other than embedded pavement sensors that can augment the FAST’s detection capability and improve its reliability by providing additional information on the road surface condition (RSC). The technologies include non-invasive pavement friction sensors [Ewan et al. 2013], sensors installed on vehicles for thermal mapping [Todeschini et al. 2016, Hu et al. 2019] or RSC monitoring [Chen et al. 2011, Pu et al. 2019], and the combined use of mobile camera and smart phone [Linton and Fu 2016] or mobile RWIS technologies [Ye et al. 2012] for RSC monitoring. Most of them could be readily integrated with the AVL technology [Santiago-Chaparro et al. 2012] to unlock their potential in enabling more proactive, efficient and resilient WRM operations.
ICT capabilities
This commentary agrees that as part of the transportation cyber-physical system (TCPS), the physical infrastructure of roadways should be “upgraded with digital (ICT) infrastructure that evolves with increasing CV penetration levels (so as to) create an environment suitable for fostering beneficial V2I innovations” [Khan et al. 2019]. This is critical for enabling the timely and reliable sensing, processing and communication of the unprecedented amount of data available in the VII environment. Khan et al. [2019] summarized the typical roadway digital infrastructure components (as shown in Fig. 3), including RSUs, traffic signals, loop detectors, traffic cameras, and DMS communicating with both CVs/AVs and the backend infrastructure (transportation management center (TMC) servers and cloud servers) in a real-time or near-real-time fashion.
In the context of VII revolution, no single communication technology “(in the near future) can support such a variety of expected V2X applications for a large number of vehicles”, at least not efficiently [Abbound et al. 2016]. Currently, dedicated short-range communications (DSRC, featuring its 5.9 GHz frequency band) and cellular networks are the two main technologies for V2X communications. Abbound et al. [2016] reviewed various DSRC-cellular hybrid architectures and discussed the interworking challenges between the two technologies, with DSRC transceiver embedded in RSUs interacting with in-vehicle onboard units (OBUs) as well as the backhaul network (via cellular or wired Internet connections). Dey et al. [2016] developed a method to optimize available communication options for V2V and V2I applications in a heterogeneous wireless network consisting of DSRC and “other wireless technologies (e.g., Wi-Fi, LTE, and WiMAX)”. Nonetheless, Jenkins et al. [2017] concluded that “currently, DSRC for wireless access in vehicular environments (WAVE) (protocol) is the only way to provide V2X communications with high reliability and low latency” and thus suitable for dissemination of basic safety messages.
We are at the dawn of the 5G era, which stands for the 5th generation mobile network and features communication capabilities better meeting the needs of IoT and V2X applications, particularly safety-critical use cases [Boban et al. 2016]), i.e., higher data rates, ultralow latency, ultra-reliability, and increased availability. Boban et al. [2016] proposed an architecture of the 5G V2X network, embodied in a heterogeneous multi-radio V2X network designed to leverage the strengths of “cellular systems in centimeter (cmWave) and millimeter (mmWave) frequency bands, vehicular visible light communication (VVLC)” and DSRC/WAVE.
The computing infrastructure, either distributed or centralized, is also a key player in the TCPS to enable smart and seamless mobility, and should be planned, designed, deployed, operated and maintained properly [Khan et al. 2019]. To attain outstanding performance and reliability of V2X applications, innovations are also expected in the field of information technology in terms of system design, hardware and software, by tapping into recent advances in blockchain and edge computing [Liu et al. 2018], big data analytics [Darwish and Bakar 2018], machine learning and artificial intelligence [Vishnukuma et al. 2017], and so on. For instance, Raza et al. [2018] described an advanced ITS based on ultrahigh speed, ultralow latency 5G scenario, where networking technologies (cloud computing, mobile edge computing, and software defined networking) were integrated to provide ubiquitous network connectivity and efficient computing to support V2X communications. Among them, edge computing (a.k.a., fog computing) refers to “the enabling technologies allowing computation to be performance at the edge of the network, on downstream data on behalf of cloud services and upstream data on behalf of IoT services” [Shi et al. 2016], and it brings many advantages in “addressing the concerns of response time requirement, battery life constraint, bandwidth cost savings, and data safety and privacy”.