Literature Reviews

The effects of micromobility modes on sustainability goals are mixed. A literature review by
McQueen et al [1] defined micromobility modes as “small, lightweight human-powered or electric vehicles operated at low speeds, including docked and dockless e-scooters and bike share systems,” and found mixed results of the modes’ effects across three key sustainability goals – reducing greenhouse gas emissions, equitable and reliable operations, and enhancement of the human experience. Regarding greenhouse gas emissions, the review concluded that micromobility modes have the potential to decrease emissions when serving as a substitute for automobile trips. One way this can occur is by complementing transit; making it more accessible and convenient and therefore more competitive with automobile trips. However, the review also found that micromobility trips often replace walking or transit trips, thus increasing emissions [2].

Municipalities see a human benefit to offering alternative modes. Research around perceptions of new mobility has found them to be a pleasant experience, especially for electrified mobility, although many of the studies are focused on e-bikes [3], [4]. Additionally, a significant amount of research focuses on the integration of micromobility with public transportation. The body of work related to this topic generally spans four study areas - policy, sustainability, interactions between shared micromobility and public transit, and infrastructure [5]. Improving first/last mile access and network efficiency is also a major focus area [6], [7]. Future research should focus on sustainability through business models analysis, comparing public and private operations and how best to navigate regulatory burdens surrounding the deployment of such services.

Several researchers have tried to understand the effects of ride-hailing on transportation system performance related metrics such as vehicle miles traveled (VMT) [1], [2], [3]. Most studies are in agreement that Transportation Network Companies increase VMT and decrease public transit ridership [1], [2], [3], [4]. For example, Wu and MacKenzie (2021) used the 2017 National Household Travel Survey (NHTS) along with causal inference to estimate the effect of ride-hailing services on VMT. They concluded that a net 7.8 million daily VMT or 2.8 billion annual VMT were added nationwide due to ride-hailing services at the time of the 2017 NHTS data collection [1]. Other studies have tried to understand the effect of congestion pricing strategies on ride-hailing ridership [1]. For example, Zheng et al. (2023) estimated the effects of ride-hailing congestion pricing policy on ridership in Chicago and concluded that the policy led to a growth in shared trips and a decline in single trips. Some studies have also tried to understand the effects of ride-hailing on transit and other modes of transportation [1], [2], [3].
Current opportunities for future research include: 1) using more updated data (e.g., 2022 NHTS) to assess the effects of ride-hailing on VMT and travel behavior, as the impact of ride-hailing services changes dynamically, and 2) assessing the impact of ride-hailing services in rural areas and less studied regions of the country, which could provide insights for local and state policies.

On-demand delivery services have been shown to have a significant impact on transportation systems, both positively and negatively [1]. On the positive side, modern delivery services could reduce shopping trips to physical stores and related energy consumption [2] and greenhouse gas emissions [3]. Emissions from delivery services vary based on delivery scheduling [4], service coverage area [5], engine type (e.g., combustion or electric), and efficiency of cooling equipment [6]. On the negative side, the increasing number of delivery vehicles adds to crash risk in the transportation system, particularly for road users [7]. In addition, the delivery vehicles compete for limited curbside space in the urban area [8], [9].

Research on the impact of robotic delivery services on transportation systems is predominantly theoretical, due to scarce empirical evidence. The City of Pittsburgh [10] operated a six-month pilot program with Kiwibot and deployed a limited number of devices (less than 10 at any time) to deliver packages. Different from package delivery robots, which mostly operate on sidewalks and have a limited influence on the road traffic, future autonomous delivery vehicles could exert a huge impact on the traffic systems. Studies showed mixed results about the effects of autonomous vehicles on traffic flow efficiency, both positive and negative, depending on their modeling conditions [11].

Carsharing can increase the efficiency of the transportation system by allowing multiple individuals to access a single vehicle that uses a single parking space [1]. As with many transportation modes, carsharing serves different needs under different conditions within a broader transportation system. For example, carsharing works well in communities with low vehicle ownership rates [2] or co-located with bus services and areas that have mobility constraints in accessing metro services [3]. Additional research is needed to determine what specific pricing conditions and when and how public or privately-operated carsharing can be sustainable.

Demand-responsive transit (DRT) and microtransit optimization has been studied using models and theoretical networks. From a strategic design perspective, continuous approximations of demand over time and space in highly theoretical networks were used to determine optimal flexible service types as a function of demand density [1], [2], [3], [4]. For tactical decision making, studies have used optimization methods in highly theoretical networks to optimize slack times [5], [6], longitudinal velocities [7], service cycle times [8], and compulsory stop selection and sequence [9]. Finally, from an operations standpoint, previous studies have evaluated policies such as dynamic stations [10], flag stops [4], point deviations[11], and optimal cycle lengths [12] in off-line settings. Few studies have also evaluated real-time operational strategies, such as optimal shuttle departure times [13] and routing/stopping decisions for rail connector services [14]. Generally, previous studies consider highly simplified or theoretical network conditions (e.g., grid networks, uniform travel times and uniform trip types), which can lead to suboptimal decision-making and unrealistic performance estimates. Though there are a number of DRT or microtransit pilots throughout the country, analysis and evaluation of real-world microtransit systems do not necessarily improve the overall system performance on efficiency, accessibility and financial sustainability. There is potential for DRT and microtransit service to be improved by innovative technologies, such as real-time demand prediction, real-time ride requests, coordination with both fixed-route mainline public transit and privately operated ride-hailing or mobility service. Both technologies of sensing, communication and service, and AI-powered algorithms could improve DRT and microtransit performance.

Studies show that Mobility-as-a-Service (MaaS) could decrease the use and ownership of private vehicles and support a switch to active travel modes and transit [1], [2], [3]. However, the magnitude of this switch is not comprehensively explored among the literature [2]. According to one simulation study, MaaS could reduce emissions by up to 54 percent, depending on the modeling scenarios [4]. Another simulation study showed that MaaS could reduce transport-related energy consumption because of the introduction of car-sharing and bike-sharing services [5]. Another study suggested that MaaS could reduce vehicle miles traveled and related negative externalities [6].

Several research directions are promising for future studies. First, there are limited studies on what drives people to use MaaS, highlighting a need to explore user incentives to adoption. Understanding these factors can inform more targeted service design and marketing strategies. Second, modeling the integration of multi-modal travel within MaaS is crucial. This could offer insights into optimizing traffic flows and enhancing the environmental and social benefits of MaaS. Third, the collaborative mechanism between the public and private sectors in the MaaS ecosystem requires further examination. Investigating how these entities can better cooperate could foster the broader application of MaaS solutions.

Many researchers have used agent-based simulation to assess the effects of Automated Vehicles (AV)s on transportation system operations and efficiency (e.g., congestion and Vehicle Miles Traveled (VMT)) [1], [2], [3], [4], [5], [6], [7]. For example, Yan et al. (2020) simulated and then evaluated the performance of a shared autonomous vehicle fleet serving requests across the Minneapolis-Saint Paul region [1]. Yan et al. [1], [2], [3], [4], [5], [6], [7] estimated that the average shared AV could serve at most 30 person-trips per day with less than a 5 minute wait time but generates 13 percent more VMT. Yan et al. [1], [2], [3], [4], [5], [6], [7] also concluded that dynamic ridesharing could reduce shared AV VMT by 17 percent on average and restricting shared AV parking on the busiest streets could generate up to 8 percent more VMT.
Other methods such as static traffic assignment models and scenario analysis, have also been used to to understand the effect of AVs on congestion and VMT [8], [9], [10], [11], [12], [13]. For example, Harper et al. (2016) estimated the upper bound increase in travel with AVs for the non-driving, elderly, and people with travel-restrictive medical conditions by creating demand wedges and assuming that these traditionally underserved populations would travel as much as younger and/or healthier populations [9]. Harper et al. (2016) estimated that vehicle automation addressing latent demand for underserved population could increase VMT by as much as 14 percent, with females and non-drivers making up most of this increase [9].

Most studies are in agreement that AVs are likely to increase VMT and congestion, due to increased trip making, the ability for AVs to search for more distant and cheaper parking, and the additional VMT generated from people switching from personally owned vehicles to shared autonomous vehicles, generating empty travel [5], [9], [14]. Current opportunities for future research in this area include: 1) simulating AVs considering a heterogeneous population of travelers with different values of travel time (VOTT) and 2) incorporating parking to estimate the impact of AVs on transportation system operations [15].

The University of California Institute of Transportation Studies recently released a technical report that summarizes Universal Basic Mobility (UBM) pilot programs in California along various design dimensions, including eligibility requirements, monetary assistance value, and allowable travel modes [1]. For example, Los Angeles, CA offered 2,000 residents $150 per month for use of public transit, private taxi, transportation network company (e.g., Uber), electric bikeshare, and carshare. The Pittsburgh, PA program gave 50 residents unlimited access to transit and bikeshare along with a monthly credit for scooter and carshare [2]. Other U.S. cities that have implemented a UBM pilot include Portland, OR; Sacramento, CA; Oakland, CA; and Stockton, CA.

Evaluations of most UBM programs are still underway, though some results are available for Oakland and Portland. The Oakland Department of Transportation and Alameda County Transportation Commission surveyed 66 participants pre-program and mid-program, and they observed that 66 percent of these participants used the extra mobility funds for commuting. They also found that 90 percent of funds were spent on transit, and the number of participants who self-reported driving as their primary mode declined by 6 percent for commuting trips [3]. Researchers at Portland State University also evaluated the Portland program based on surveys. Their results revealed that participants had positive UBM perceptions: 89 percent of participants reported greater travel flexibility and 66 percent of participants reported the ability to reach work-related activities that would have been otherwise unreachable. Regarding travel mode shift, over 50 percent of participants agreed that they increased their usage frequency of Uber/Lyft, taxi, bikeshare, and e-scooter [4].

In addition to survey results, policymakers would benefit from studies that analyze how UBM affects system-level efficiency, accessibility and equity. However, there is limited completed research to this end. Most studies focus on analysis based on surveys that are only reflective of stated preferences from participants. Those stated preferences may not be generalizable or accurate in practice, and they are limited to a small spatio-temporal scope. Research gaps lie in tracking and understanding the actual (revealed) preferences of UBM participants, in regards to how UBM, by various levels of support, enables those participants to select mobility options to improve efficiency, accessibility and equity. In particular, research is needed to understand how those improvements vary by neighborhood and population groups. This would help public agencies and private service providers to jointly design a UBM program that is tailored for population groups with a vital business model to scale/group in the future.

Connected vehicles (CVs), connected autonomous vehicles (CAVs), and Vehicle-to-Everything (V2X) technologies can improve transportation system operations and efficiency. For example, Guler et. al [1] used simulations to assess the potential for CVs to optimize intersection efficiency, finding that CVs could reduce delays by up to 60 percent. The reductions in intersection delays were up to 7 percent greater for CAVs compared to CVs controlled by drivers [1]. Platooning technology can reduce fuel consumption and smooth traffic oscillation [2], and there is a growing body of research around autonomous intersection management [3], [3], [4], [5]. Vehicle to infrastructure (V2I) technology can also improve traffic efficiency by optimizing traffic signal control [6].

A significant body of research exists on how to optimize traffic and safety using connective technology, but it is primarily based on simulations since real-world data is limited [1].

Autonomous vehicles (AVs) applications can be categorized into a) private autonomous vehicles, b) shared autonomous taxis and c) heavy duty autonomous vehicles like trucks and buses. Research studies [1] based on simulation and hypothetical models suggest that connected and autonomous vehicles (CAVs) will result in increased vehicle miles, shift from active travel to more autonomous vehicle (AV) travel, and more urban sprawl. Thus, these technologies seem to be in conflict with sustainability goals. However, AVs can be environmentally friendly and have social equity benefits if used for public transport, shuttles and shared use mobility. Based on public perception [1], AVs are well accepted for public transport among the public as opposed to being used as personal vehicles.
Mouratidis & Cobeña Serrano [2] analyzed the intention for using autonomous buses within a case study area to see user perceptions of AVs. They observed that travelers would be willing to adopt autonomous buses if these offer more frequent departures. López-Lambas & Alonso [1] observed autonomous buses to decrease congestion, intersection wait time and reduce emissions as factors to influence perception of acceptance of these technologies.
Automated applications for trucks have received a lot of attention over the years due to ease of platooning with freight and the potential network wide operation and fuel consumption benefits. Z. Wang et al. [3] tested connected eco-driving system on heavy duty trucks in Carson, California on two corridors with six intersections. They observed smoothing of speed profiles when trucks approached signalized intersections and showed 9 percent and 4 percent fuel savings in acceleration and deceleration phases. Lee et al. [4] analyzed the safety and mobility of different market penetrations of truck platoons using simulation. They observed that safety improved with 2.5 percent higher speed difference by increasing market penetration rate of truck platoons. M. Wang et al., [5] tested truck platooning under different penetration rates in a simulation environment and observed truck platooning to reduce congestion and improve throughput at higher market penetration rates. They observed significant variability in merging speed under saturated traffic.
The literature has focused more on the benefits of AVs in personal vehicles than the use of heavy duty AVs. Additional research is needed regarding the use of heavy duty AVs for public transport, as well as potential system impacts of automated trucking.