Traffic Operations and Management

Ecological Velocity Planning through Signalized Intersections: A Deep Reinforcement Learning Approach

Pozzi, Andrea
Bae, Sangjae
Choi, Yongkeun
Borrelli, Francesco
Raimondo, Davide M.
Moura, Scott
2020

The use of infrastructure-to-vehicle communication technologies can enable improved energy efficient autonomous driving. Traditional ecological velocity planning methods have high computational burden, particularly when plug-in hybrid electric vehicles are considered. Consequently, in order to retrieve an optimal velocity profile in real time, it is necessary to rely on significant approximations.In this paper, the aforementioned issue is addressed by exploiting deep reinforcement learning in order to "learn" an eco-driving velocity planner for a plug-in hybrid electric vehicle within a...

Risk-Aware Lane Selection on Highway with Dynamic Obstacles

Bae, Sangjae
Isele, David
Fujimura, Kikuo
Moura, Scott J.
2021

This paper proposes a discretionary lane selection algorithm. In particular, highway driving is considered as a targeted scenario, where each lane has a different level of traffic flow. When lane-changing is discretionary, it is advised not to change lanes unless highly beneficial, e.g., reducing travel time significantly or securing higher safety. Evaluating such “benefit” is a challenge, along with multiple surrounding vehicles in dynamic speed and heading with uncertainty. We propose a realtime lane-selection algorithm with careful cost considerations and with modularity in design. The...

Reinforcement Learning Versus PDE Backstepping and PI Control for Congested Freeway Traffic

Yu, Huan
Park, Saehong
Bayen, Alexandre
Moura, Scott
Krstic, Miroslav
2022

We develop reinforcement learning (RL) boundary controllers to mitigate stop-and-go traffic congestion on a freeway segment. The traffic dynamics of the freeway segment are governed by a macroscopic Aw–Rascle–Zhang (ARZ) model, consisting of 2 \times 2 quasi-linear partial differential equations (PDEs) for traffic density and velocity. The boundary stabilization of the linearized ARZ PDE model has been solved by PDE backstepping, guaranteeing spatial L<sup>2</sup> norm regulation of the traffic state to uniform density and velocity and ensuring that traffic oscillations are...

HumanLight: Incentivizing Ridesharing via Human-Centric Deep Reinforcement Learning in Traffic Signal Control

Vlachogiannis, Dimitris
Wei, Hua
Moura, Scott
Macfarlane, Jane
2024

Single occupancy vehicles are the most attractive transportation alternative for many commuters, leading to increased traffic congestion and air pollution. Advancements in information technologies create opportunities for smart solutions that incentivize ridesharing and mode shift to higher occupancy vehicles (HOVs) to achieve the car lighter vision of cities. In this study, we present HumanLight, a novel decentralized adaptive traffic signal control algorithm designed to optimize people throughput at intersections. Our proposed controller is founded on reinforcement learning with the...

Traffic Delay at Unsignalized Intersections: Clarification of Some Issues

Daganzo, Carlos F.
1977

Investigations in this area have been directed at finding the delay to a motorist who arrives at an intersection and wishes to cross a high-priority traffic stream. In this paper two conceptual errors that have appeared in some past publications are identified and corrected.

The Uniqueness of a Time-dependent Equilibrium Distribution of Arrivals at a Single Bottleneck

Daganzo, Carlos F.
1985

Motorists going through a bottleneck during the morning rush hour have to time their departure times to ensure they arrive to work at a reasonable time. Traffic and congestion levels at the bottleneck depend on the motorists' work schedule and the disutility of unpunctuality. This paper shows that, under certain conditions, there is only one equilibrium order of arrivals; an order under which motorists do not have an incentive to jockey for position in the queue.

Modeling Distribution Problems with Time Windows: Part I

Daganzo, Carlos F.
1987

This paper shows how distribution problems with delivery time constraints can be modeled approximately with just a few variables. Its objective is neither to develop a scheduling algorithm nor an exact predictive method; rather, it is to illustrate some trade-offs and principles that can be used for planning and algorithm development. A workday is divided into time periods. Time windows are modeled by specifying the period in which a vehicle should visit each customer. (The companion paper explores scenarios where many customers do not specify a time window, and thus, it is advantageous...

Modeling Distribution Problems with Time Windows. Part II: Two Customer Types

Daganzo, Carlos F.
1987

This paper extends the results of a previous study concerning distribution with time windows. It is recognized that customers who do not need to be allocated to a time window should receive different service than the rest. Three strategies were studied to accomplish that: stratified service, discriminating service, and staggered and discriminating service. Of these, the last strategy yields the lowest local distribution distance per customer, a distance which can be less than half that for the strategy explained in the previous paper (joint service). Stratified service, however, can yield...

The Break-Bulk Role of Terminals in Many-to-Many Logistic Networks

Daganzo, Carlos F.
1987

This paper examines the structure of many-to-many logistics networks. Using as little data as possible, it attempts to answer macroscopic questions such as: How many terminals should be used? Should they be used at all? What should be the frequency of service? Although such a problem could be formulated with a large number of parameters and data, we show that near-optimal network structures can be characterized by two dimensionless constants which can be determined from the data (e.g., from the value of the items carried, the number of origins, the size of the service region, etc …). The...

Restricting Road Use Can Benefit Everyone

Daganzo, Carlos F.
1992

This paper seeks congestion reduction schemes that do not penalize anyone. It shows that a combination of rationing and pricing (unlike congestion pricing alone) can benefit everyone even if the collected revenues are not returned to the population. The simple conditions under which this is possible are identified. Little data are needed to choose a proper policy. Examples are given.