Integrated Approach to Variable Speed Limits and Ramp Metering

Abstract

Traffic congestion continues to remain a serious problem in most countries, with significant economic and environmental losses. ‘Intelligent Transport Systems (ITS)’ that utilize advanced information and communication technology for managing road infrastructure, including vehicles and users, have shown great potential in dealing with this problem. However, most current traffic management measures, like route guidance, ramp metering and variable speed limits are designed and implemented independently. Potentially counter-productive when deployed simultaneously, integration of these measures can benefit from synergic effects. The aim of this work is to integrate a near-future variable speed control strategy with ramp metering, to operate efficiently at a freeway merging section. From a policy perspective, this is a challenge because freeways and urban network are managed by different authorities. Currently, a spillback of ramp queue to the urban network results in a deactivation of ramp metering, and an abrupt release of on-ramp vehicles onto the freeway. This is undesirable for the freeway control measure. Therefore, the development of the integrated strategy considers both, an improvement in freeway efficiency, and at the same time queue reg- ulation at the on-ramp. The research arch is initiated by conducting a literature survey, wherein the theoretical and control as- pects of available speed control approaches are studied first. Subsequently, different queue management approaches used in current ramp metering strategies are reviewed to identify the desirable features for a queue controller. Finally, integration approaches for combining ramp metering with variable speed limits are looked at. It is seen that optimization-based integration strategies may offer huge potential to improve freeway effi- ciency but these strategies are computationally complex for systems available in practice. Therefore, a so- phisticated reactive variable speed limits strategy - COoperative Speed Control ALgorithm (COSCAL v2) is chosen for integration with ramp metering. COSCAL v2 is a counteractive control strategy that resolves moving jams to recover loss in efficiency from capacity drop. It uses shockwave theory concepts to speed limit the exact the number of vehicles required to first resolve a jam, and next, to stabilize the traffic while recovering traffic throughput. The theory development is conducted in two stages from a simplistic to a more advanced approach that relaxes some of the initial assumptions. The basic strategy is developed to guarantee jam resolution under the condition of a constant ramp metering flow during the activation of COSCAL v2. Additionally, ramp queue constraints are not included, which implies that in order to prevent a spillback, a sufficiently high metering flow must be chosen. The theoretical formulation to determine when and how many vehicles must be speed-limited, as in COoperative Speed Control ALgorithm (COSCAL) v2, is preserved here. The core principle involves instantaneously speed-limiting the vehicles entering the moving jam. In this way, the inflow into the jam is reduced as compared to the flow exiting the jam, resulting in jam resolution. In the integrated strategy, the theory to determine the most upstream speed-limited vehicle that resolves the jam, given a constant ramp metering flow, is formulated. Similarly, the theory for determining the last vehicle speed-limited for stabilization is extended for the additional ramp flow. For this, COSCAL v2 checks for a maximum (average) density in the speed-limited area to ensure traffic stability. The value of the target density determines the freeway flow arriving at the on-ramp. Then, the integrated approach determines reduced target densities required to accommodate the ramp flow. It is also described how these densities can be implemented in practice, and how errors its realization can be compensated within the feedback structure of the scheme. However, this approach is not feasible for time-dependent ramp metering flows. This is essentially because accommodating a varying ramp flow will require implementing different densities in space-time, which is both difficult to determine and implement in the COSCAL v2 algorithm. Therefore, in the advanced strategy the theory of COSCAL v2 is extended for compatibility to a more sophisticated ramp metering approach. In the new approach, rather than regulating density the flow arriving at the on-ramp is directly controlled by determining when and where vehicles should enter and exit speed-limits. This addition- ally improves the effectiveness of COSCAL v2. Further, a new ramp metering strategy using cumulative curves is developed to include: queue-length constraints; policy restrictions on ramp metering flows; and limitations on the minimum freeway flow that can be achieved with speed limits given the traffic state mea- surements. One of the main advantages of the advanced strategy is its responsiveness to changing ramp flows, and the utilization of freeway speed-limitation to prevent a queue spillback on the ramp. Following the development of the two integration strategies, the advanced approach is translated to an al- gorithm. This step ensures that a controller can be designed to realize traffic behaviour according to the developed theory. Here, the complete strategy is divided into five modules that together determine the spatial extent of speed limits at any given time. Similarly to the algorithmic formulation of COSCAL v2, unique driving modes for the different detection and actuation tasks of the algorithm are used to communi- cate the control scheme for implementation. The control strategy thus developed is tested by means of simulation. Of the five modules, three modules related to the speed control strategy are implemented in this work. Simulation runs under different de- mand conditions are performed to verify the behavioural performance of the approach. Jam waves - that otherwise occur when ramp metering and COSCAL v2 are implemented independently - are prevented in the integrated approach. The results offer an initial indication of potential improvement from the strat- egy. However, a more comprehensive quantitative evaluation is recommended. Furthermore, some critical developments to the approach following the simulation study are identified. In conclusion, this work demonstrates two ways in which COSCAL v2 can be integrated with ramp me- tering (RM); how cumulative curves can be used for integration of control measures; and how efficiency gains on the freeway and urban network can be balanced. The simulations offer a proof of concept for the working of the advanced integrated strategy. However, the basic strategy should be both implemented and tested further.