Search

This paper proposes a framework for deep Long Short-Term Memory (D-LSTM) network embedded model predictive control (MPC) for car-following control of connected automated vehicles (CAVs) in traffic mixed with human-driven vehicles (HDVs) and CAVs. The framework consists of: 1) lead HDV trajectory prediction through D-LSTM; and 2) CAV car-following control via MPC based on the predicted vehicle trajectory. For the trajectory prediction, two D-LSTM structures are developed based on the availability of preceding vehicle information: 1) ‘sufficient’ historical information of the position and speed of multiple vehicles ahead; and 2) ‘insufficient’ information where preceding vehicle information is unavailable (e.g., due to failed communication). Based on the prediction, a distributed MPC is designed for each scenario by incorporating the predicted trajectory into state space construction. The proposed D-LSTM models are trained and tested with the NGSIM data for validation. Numerical simulation results for various traffic conditions suggest that the proposed strategies perform better than traditional MPC methods in terms of control objective cost reduction, smoother control, and stabilizing effect. The results also indicate that the sufficient information case outperforms the insufficient information case as expected, which highlights the importance of stable communication.

This paper investigates the longitudinal control of vehicle platoons with a focus on external disturbances, parameter uncertainties, and communication failures. When vehicle platoons encounter vehicle to vehicle wireless communication failures, the preceding vehicle’s acceleration is unavailable, which degrades the performance of the system. However, the existing achievements can not be directly used to address the aforementioned three issues. To this end, a generalized extended state observer-based control (GESOBC) law is contrived. First, the parameter uncertainties and the external disturbances are together regarded as a lumped disturbance. Then, a generalized extended state observer is designed to estimate the lumped disturbance and the preceding vehicle’s acceleration, separately. Based on the estimation, a composite controller consisting of a state feedback control part and an estimation-based feedforward control part is developed. Furthermore, it is proved that the proposed GESOBC method can guarantee the exponentially bounded stability of the individual vehicle and the input to state string stability of the whole vehicle platoon. Finally, numerical simulations are conducted to demonstrate the effectiveness and feasibility of the proposed method.

Ion‐imprinted polymers attract a lot of attention and are actively studied because of easy preparation and their high selectivity. Zn is known to be essential mineral in human body; however, when it is absorbed into our body, it is hardly discharged completely and causes various disease or side effects. To suggest new method for selective Zn extraction, Zn ion‐imprinted polymer (ZIIP) was synthesized by bulk polymerization and NZIIP (non‐ZIIP) was prepared as a control. Then, ZIIP and NZIIP were characterized by Fourier‐transform infrared spectroscopy, Scanning electron microscope, and X‐ray photoelectron spectroscopy to investigate structural changes during polymerization. Adsorption isotherm studies indicated that sigmoidal Langmuir isotherm equation was fitted well and the maximum adsorption capacity and affinity of Zn on ZIIP was superior to NZIIP. Dithizone was used to chelate with Zn in supernatant for colorimetric determination. ZIIP‐solid‐phase extraction potentially used as Zn ion filter was investigated showing good removal efficiency. Zn ion‐imprinted polymer selectively binds Zn ion with dithizone for colorimetric determination.

Truck parking is a crucial element of the United States’ transportation system as it provides truckers with safe places to rest and stage for deliveries. Demand for truck parking spaces exceeds supply and shortages are especially common in and around urban areas. Freight operations are negatively affected as truck drivers are unable to park in logistically ideal locations. Drivers may resort to unsafe practices such as parking on ramps or in abandoned lots. This report seeks to examine the potential parking availability of vacant urban parcels by establishing a methodology to identify parcels and examining whether the identified parcels are suitable for truck parking. Previous research has demonstrated that affordable, accessible parcels are available to accommodate truck parking. When used in conjunction with other policies, adaptation of urban sites could help reduce the severity of truck parking shortages. Geographic information system parcel and roadway data were obtained for one urban area in each of the 10 Mid America Association of Transportation Officials region states. Area and proximity filters were applied followed by spectral analysis of satellite imagery to identify candidate parcels for truck parking facilities within urban areas. The automated processes created a ranked short list of potential parcels from which those best suited for truck parking could be efficiently identified for inspection by satellite imagery. This process resulted in a manageable number of parcels to be evaluated further by local knowledge metrics such as availability and cost, existing infrastructure and municipal connections, and safety.

Chronic fatigue syndrome (CFS) reduces the health-related quality of life in the working-age population; however, studies have rarely investigated this group. A mixture of Cervus elaphus Linnaeus, Angelica gigas Nakai, and Astragalus membranaceus Bunge (CAA) may be an effective anti-fatigue supplement. However, few clinical trials have explored the anti-fatigue effects of herbal medicines in human participants. Therefore, this study aimed to investigate the effects of the CAA herbal complex on muscle fatigue and endurance capacity in a randomized, placebo-controlled, double-blind trial.

This paper presents a behavioral car following model, named the chained asymmetric behavior model, that improves on the asymmetric behavior model. This model is inspired by the empirical observation that vehicles react proportionately to the magnitude of disturbance experienced when traversing through a stop-and-go oscillation, deviating from a constant following behavior observed in equilibrium conditions. Findings from simulation experiments suggest that this “second-order” effect significantly affects traffic throughput and evolution under disturbances. Knowledge obtained from the model is leveraged toward designing control for connected automated vehicles in mixed traffic streams.

This study analyzes the behavior of heterogeneous connected and autonomous vehicles (CAVs) and proposes the best vehicle sequence for optimal platoon throughput and platoon formation. A spring-mass-damper (SMD) system is adopted for control of CAVs, and the control parameters are formulated in relation to the physical capabilities of vehicles. To gain insight, we consider three types of vehicle: passenger cars, mini-vans, and heavy-duty vehicles. For each type, we investigate the maximum platoon throughput and the clustering time, defined as the time to reach the target equilibrium state. We further investigate different sequences of vehicle types in a platoon to identify the optimal vehicle order that maximizes the throughput and minimizes clustering time. Findings suggest that the highest performance vehicle (in relation to acceleration capability) should be placed as the leader of a platoon and that the number of passenger cars behind heavy vehicles (e.g., semi-trailers) should be minimized in the platoon. In addition, we examine how the proportions of lower performance vehicles affect throughput and clustering times. The result suggests that the higher the proportions, the lower the throughput and the longer the clustering time. The lowest performance vehicle had the greatest effect.

This research presents an empirical analysis of lane flow distribution (LFD) in three-lane congested freeways. Data from 72 sites were analyzed to investigate LFD trends with respect to flow and the traffic and freeway characteristics that influence lane-specific traffic behavior. Regression analysis was employed to identify recurrent linear LFD trends with respect to flow and evaluate the factors that instigate the observed LFD patterns. It was found that varying LFD trends are just as prevalent as constant LFD trends and that LFD is not uniform. In addition, four recurrent combinations of varying LFD trends were observed among the sites. Regression models were estimated to represent class-specific linear LFD trends for sites with varying LFD. Results show that the estimated models represent the observed linear LFD trends reasonably well. It was also observed that proximity to freeway merges increases the probability of varying LFD trends, while the presence of HOV lanes reduces it.

Features of transitions between stationary traffic states for both congestion onset and dissipation were evaluated through empirical observations at two freeway sites. Results reveal that transition durations vary significantly between sites and that they can be rather long, with observed average transition durations of 42–45 min in one site. It was also observed that transition durations remain relatively constant within a queue, while transition rates vary as the transition zone travels upstream of the bottleneck. In addition, results also suggest that complex site geometry, particularly the presence of multiple lane reductions, has an effect on the spatial evolution of transition rates, while single merging and diverging maneuvers do not seem to affect the mechanism of transitions. Finally, it was observed that the presence of non-stationary traffic periods could have a considerable effect on the estimation of key traffic variables in locations where long transition durations are observed.

This paper sheds light on mixed-traffic dynamics considering the differences in driving characteristics, namely acceleration/deceleration rate, desired speed, and response time, between connected and autonomous vehicles (CAVs) and human-driven vehicles (HDVs). In light traffic, these differences were found to induce platoon formations, headed by vehicles with a lower acceleration rate and propensity not to exceed the desired speed (HDV in this study). Platoon formations lead to large inter-platoon spacing that can be utilized to accommodate cut-in vehicles. In a near-capacity condition, however, the differences in driving characteristics can induce voids and undermine traffic throughput when traffic is disturbed by merging vehicles. Based on these findings, a simple CAV control method is proposed based on the spring-mass-damper (SMD) system approach that directly considers the HDV behavior to mitigate disturbance propagation and throughput reduction. The main principle is to adjust the control parameters (lower spring coefficient and higher damping coefficient in the SMD control model) with an aim to control CAVs to absorb the cut-in impact (i.e., spacing shortage) before it reaches the first upstream HDV. A simulation experiment suggests the feasible region of the control parameters, subject to the recovery time, the number of controllable CAVs, and the cut-in impact.

This paper proposes a theoretical framework to control a platoon of connected and autonomous vehicles (CAVs) in the presence of cut-in movements. The control method is developed based on the spring–mass–damper (SMD) system concept and aims to improve the platoon efficiency and stability after a cut-in movement (e.g., lane change and merging from on-ramp). The method seeks to resolve a disturbance created by a cut-in vehicle by systematically setting two control parameters, spring constant and damping coefficient, of the SMD-based control model based on the prevailing traffic conditions. The control method is evaluated through a simulation based on the changes in speed and spacing, recovery time to reach the desired speed, disturbance propagation, and platoon flow. The simulation result shows that the control method can effectively reduce the disturbance caused by a cut-in movement and improve platoon flow.

This study presents a strategy for platoon formation and evolution of connected and autonomous vehicles (CAVs) in free-flow traffic. The proposed strategy is based on swarm intelligence, which describes bird flocking, fish schooling, and so on, in natural and artificial systems. In this concept, CAVs behave according to some rules to move together as a platoon without collisions. The rules are expressed by a spring–mass–damper system: CAV platoon formation and evolution are controlled by the spring constant and damping coefficient. Valid domains of these control parameters were derived on the basis of physical vehicle properties (e.g., bounded acceleration and deceleration) for realistic control. Furthermore, various relationships—maximum (in which the spring constant was set at its maximum for the given flow), quadratic, and cubic—between the control parameters and traffic flow were examined with simulations to obtain insight into desirable control parameter settings. The results suggest that the most efficient platooning can be achieved by the maximum relationship between the spring constant and flow with critical damping. However, the cubic relationship coupled with overdamping is more desirable in low-flow states to allow more freedom for vehicles to change lanes.

This paper examines CACC truck platooning on uphill grades. It was found that the design of CT policy should consider the effects of low crawl speeds on significant upgrades. Three simple solutions, which have different impacts on traffic flow efficiency, are proposed. Furthermore, truck platoons, controlled by a state-of-the-art CACC model, become asymptotically unstable beyond some critical grade. The errors are permanent, suggesting that trucks fail to re-engage after the upgrade. This occurs by complex interactions between the CACC control and the bounded acceleration capabilities of trucks. New control concepts are developed to complement the existing control model and achieve asymptotic (and string) stability. The instability mechanisms and new control concepts are not specific to the control model used.

This study presents a strategy for platoon formation and evolution of connected and autonomous vehicles (CAVs) in free-flow traffic. The proposed strategy is based on swarm intelligence, which describes bird flocking, fish schooling, and so on, in natural and artificial systems. In this concept, CAVs behave according to some rules to move together as a platoon without collisions. The rules are expressed by a spring–mass–damper system: CAV platoon formation and evolution are controlled by the spring constant and damping coefficient. Valid domains of these control parameters were derived on the basis of physical vehicle properties (e.g., bounded acceleration and deceleration) for realistic control. Furthermore, various relationships—maximum (in which the spring constant was set at its maximum for the given flow), quadratic, and cubic—between the control parameters and traffic flow were examined with simulations to obtain insight into desirable control parameter settings. The results suggest that the most efficient platooning can be achieved by the maximum relationship between the spring constant and flow with critical damping. However, the cubic relationship coupled with overdamping is more desirable in low-flow states to allow more freedom for vehicles to change lanes.

This paper presents empirical findings on car-following and lane-changing behavior involving heavy vehicles; trajectory data from the next generation simulation program were used in the study. It was found that when following passenger cars, heavy vehicles tended to reduce speed variations caused by traffic disturbances and thereby dampened traffic oscillations. In contrast, passenger cars following heavy vehicles tended to amplify traffic disturbances, although with lower probability and magnitude compared with the dampening effect. Moreover, heavy vehicles tended to discourage lane changes, especially behind them. This finding has convoluted implications: although reduced lane changes can improve traffic stability by preventing or reducing disturbances, large gaps can persist behind heavy vehicles and contribute to underutilization of road capacity.

This paper presents empirical findings on car-following and lane-changing behavior involving heavy vehicles; trajectory data from the next generation simulation program were used in the study. It was found that when following passenger cars, heavy vehicles tended to reduce speed variations caused by traffic disturbances and thereby dampened traffic oscillations. In contrast, passenger cars following heavy vehicles tended to amplify traffic disturbances, although with lower probability and magnitude compared with the dampening effect. Moreover, heavy vehicles tended to discourage lane changes, especially behind them. This finding has convoluted implications: although reduced lane changes can improve traffic stability by preventing or reducing disturbances, large gaps can persist behind heavy vehicles and contribute to underutilization of road capacity.

Field test results of a variable speed limit (VSL) control algorithm, a speed-controlling algorithm using shock wave theory (SPECIALIST), were analyzed to elucidate driver response and traffic flow evolution under VSL control. Successful VSL control was characterized by nearly constant, or decreasing, demand over time. In contrast, failed VSL control was attributed to (a) significant increase in demand (during control) and (b) significant net inflow from ramps. The demand increase was found to be the leading cause of the failed control, underscoring that the efficacy of the VSL control greatly relies on its ability to incorporate demand patterns during control. On the basis of these findings, some potential improvements are offered, including a parameter design strategy that incorporates demand patterns.

Field test results of a variable speed limit (VSL) control algorithm, a speed-controlling algorithm using shock wave theory (SPECIALIST), were analyzed to elucidate driver response and traffic flow evolution under VSL control. Successful VSL control was characterized by nearly constant, or decreasing, demand over time. In contrast, failed VSL control was attributed to (a) significant increase in demand (during control) and (b) significant net inflow from ramps. The demand increase was found to be the leading cause of the failed control, underscoring that the efficacy of the VSL control greatly relies on its ability to incorporate demand patterns during control. On the basis of these findings, some potential improvements are offered, including a parameter design strategy that incorporates demand patterns.

Continuum theory is used to explain why stop-and-go oscillations in congested freeway traffic change their amplitudes when they encounter the vehicular merging and diverging maneuvers that take place near ramps. The theory describes how oscillations diminish in amplitude when they propagate past a queued (and unmetered) on-ramp and how they grow when they propagate past an off-ramp. The premise is that merging (diverging) flows change in response to freeway oscillations and that these changes in flow dampen (amplify) oscillations. The theory's descriptions are simple and rational; all its inputs and outputs are directly observable; and its predictions are shown to match real data. The theory is tested against real data collected over multiple days from congested merge and diverge sites with videos and inductive loop detectors. For merges, predictions are found to agree with observation to within 10%, and for diverges, to within 12%. The paper thus resolves in a simple way a puzzling traffic feature reported in previous studies.

This paper provides additional empirical evidence confirming a recently proposed theory on the evolution of oscillations in congested traffic. It also proposes an improved method for computing the variation in oscillation amplitude, consisting in evaluating the oscillation amplitude along characteristic lines that travel at a constant wave speed. It is also shown that the theory is robust in that approximate input parameters can be used with little loss in accuracy. The paper, in addition, provides a finding on the evolution of oscillations in freeway segments with no entrances or exits. Although previous studies found an increase in oscillation amplitude in such segments, data in this study indicate that this is not the case in general. This finding can have important implications for understanding driver behavior in homogeneous freeway segments.

This paper provides additional empirical evidence confirming a recently proposed theory on the evolution of oscillations in congested traffic. It also proposes an improved method for computing the variation in oscillation amplitude, consisting in evaluating the oscillation amplitude along characteristic lines that travel at a constant wave speed. It is also shown that the theory is robust in that approximate input parameters can be used with little loss in accuracy. The paper, in addition, provides a finding on the evolution of oscillations in freeway segments with no entrances or exits. Although previous studies found an increase in oscillation amplitude in such segments, data in this study indicate that this is not the case in general. This finding can have important implications for understanding driver behavior in homogeneous freeway segments.

Data from two freeway locations show that absent lane changing, the relationships between congested speed and spacing are linear in most ranges of congested speed. The relationships are also found to be statistically indifferent across lanes, indicating that a single general model would suffice for all lanes. Moreover, speed-spacing relationships for different types of lane changers (discretionary versus mandatory due to merging and diverging) are also found to be statistically indifferent. Anticipation and relaxation periods, defined as durations of perturbed car-following behavior before and after a lane-change maneuver, were measured for lane changers and the vehicles immediately behind the followers. Lane changers exhibited anticipation and relaxation periods of 8 and 14 s, respectively. The followers in the initial lanes exhibited anticipation and relaxation periods of 4 and 8 s, which are smaller than the periods for the followers in the target lanes (12 and 16 s) as well as the periods for the lane changers. The findings suggest that car-following behavior on a freeway can be predicted according to Newell's simplified car-following theory, and a single model can be used without much loss of accuracy for different lanes. The findings also suggest that the anticipation and relaxation processes are independent of lane-changing types but are different for lane changers and followers.

Data from two freeway locations show that absent lane changing, the relationships between congested speed and spacing are linear in most ranges of congested speed. The relationships are also found to be statistically indifferent across lanes, indicating that a single general model would suffice for all lanes. Moreover, speed-spacing relationships for different types of lane changers (discretionary versus mandatory due to merging and diverging) are also found to be statistically indifferent. Anticipation and relaxation periods, defined as durations of perturbed car-following behavior before and after a lane-change maneuver, were measured for lane changers and the vehicles immediately behind the followers. Lane changers exhibited anticipation and relaxation periods of 8 and 14 s, respectively. The followers in the initial lanes exhibited anticipation and relaxation periods of 4 and 8 s, which are smaller than the periods for the followers in the target lanes (12 and 16 s) as well as the periods for the lane changers. The findings suggest that car-following behavior on a freeway can be predicted according to Newell's simplified car-following theory, and a single model can be used without much loss of accuracy for different lanes. The findings also suggest that the anticipation and relaxation processes are independent of lane-changing types but are different for lane changers and followers.

A systemwide adaptive ramp-metering (SWARM) system is being implemented in the Portland, Oregon, metropolitan area, replacing the previous pretimed ramp-metering system. SWARM has been deployed on six major corridors and operates during the morning and afternoon peak hours. This study entails a before and after evaluation of the benefit of the new SWARM system as compared with the pretimed system using the existing data, surveillance, and communications infrastructure. In particular, the objective of this study is to quantify the systemwide benefits in relation to savings in delay, emissions and fuel consumption, and safety improvements on and off the freeway due to the implementation of the SWARM system. A pilot study was conducted for 2 weeks on a 7-mi freeway corridor in an attempt to develop a strategic design for the future regional-level study. This paper discusses the selection process of the study corridor, the experimental design, and the results obtained from the pilot study.

A systemwide adaptive ramp-metering (SWARM) system is being implemented in the Portland, Oregon, metropolitan area, replacing the previous pretimed ramp-metering system. SWARM has been deployed on six major corridors and operates during the morning and afternoon peak hours. This study entails a before and after evaluation of the benefit of the new SWARM system as compared with the pretimed system using the existing data, surveillance, and communications infrastructure. In particular, the objective of this study is to quantify the systemwide benefits in relation to savings in delay, emissions and fuel consumption, and safety improvements on and off the freeway due to the implementation of the SWARM system. A pilot study was conducted for 2 weeks on a 7-mi freeway corridor in an attempt to develop a strategic design for the future regional-level study. This paper discusses the selection process of the study corridor, the experimental design, and the results obtained from the pilot study.

Data from four merge locations in northern California and Toronto, Ontario, Canada, unveil a notable feature of driver turn taking. It was observed that queued vehicles from the on-ramp and freeway traffic streams entered a congested merge in some (nearly) fixed ratio that was independent of the merge outflow. Drivers in competing traffic streams thus entered the merge by adopting some definite turn-taking behavior, and this behavior was not influenced by the severity of the exogenous flow restriction from downstream. The findings validate part of an existing theory of merging traffic and should be considered when any new such theories are developed.

Data from four merge locations in northern California and Toronto, Ontario, Canada, unveil a notable feature of driver turn taking. It was observed that queued vehicles from the on-ramp and freeway traffic streams entered a congested merge in some (nearly) fixed ratio that was independent of the merge outflow. Drivers in competing traffic streams thus entered the merge by adopting some definite turn-taking behavior, and this behavior was not influenced by the severity of the exogenous flow restriction from downstream. The findings validate part of an existing theory of merging traffic and should be considered when any new such theories are developed.