Search

The correct specification of reward models is a well-known challenge in reinforcement learning. Hand-crafted reward functions often lead to inefficient or suboptimal policies and may not be aligned with user values. Reinforcement learning from human feedback is a successful technique that can mitigate such issues, however, the collection of human feedback can be laborious. Recent works have solicited feedback from pre-trained large language models rather than humans to reduce or eliminate human effort, however, these approaches yield poor performance in the presence of hallucination and other errors. This paper studies the advantages and limitations of reinforcement learning from large language model feedback and proposes a simple yet effective method for soliciting and applying feedback as a potential-based shaping function. We theoretically show that inconsistent rankings, which approximate ranking errors, lead to uninformative rewards with our approach. Our method empirically improves convergence speed and policy returns over commonly used baselines even with significant ranking errors, and eliminates the need for complex post-processing of reward functions., Comment: 13 pages, 8 figures, The 2024 Conference on Empirical Methods in Natural Language Processing

Multi-Agent Reinforcement Learning (MARL) methods have shown promise in enabling agents to learn a shared communication protocol from scratch and accomplish challenging team tasks. However, the learned language is usually not interpretable to humans or other agents not co-trained together, limiting its applicability in ad-hoc teamwork scenarios. In this work, we propose a novel computational pipeline that aligns the communication space between MARL agents with an embedding space of human natural language by grounding agent communications on synthetic data generated by embodied Large Language Models (LLMs) in interactive teamwork scenarios. Our results demonstrate that introducing language grounding not only maintains task performance but also accelerates the emergence of communication. Furthermore, the learned communication protocols exhibit zero-shot generalization capabilities in ad-hoc teamwork scenarios with unseen teammates and novel task states. This work presents a significant step toward enabling effective communication and collaboration between artificial agents and humans in real-world teamwork settings., Comment: Accepted to Neurips 2024, 19 pages, 10 figures

Addressing safe and efficient interaction between connected and automated vehicles (CAVs) and human-driven vehicles in a mixed-traffic environment has attracted considerable attention. In this paper, we develop a framework for stochastic time-optimal trajectory planning for coordinating multiple CAVs in mixed-traffic merging scenarios. We present a data-driven model, combining Newell's car-following model with Bayesian linear regression, for efficiently learning the driving behavior of human drivers online. Using the prediction model and uncertainty quantification, a stochastic time-optimal control problem is formulated to find robust trajectories for CAVs. We also integrate a replanning mechanism that determines when deriving new trajectories for CAVs is needed based on the accuracy of the Bayesian linear regression predictions. Finally, we demonstrate the performance of our proposed framework using a realistic simulation environment., Comment: first submission 15 pages

Merging into dense highway traffic for an autonomous vehicle is a complex decision-making task, wherein the vehicle must identify a potential gap and coordinate with surrounding human drivers, each of whom may exhibit diverse driving behaviors. Many existing methods consider other drivers to be dynamic obstacles and, as a result, are incapable of capturing the full intent of the human drivers via this passive planning. In this paper, we propose a novel dual control framework based on Model Predictive Path-Integral control to generate interactive trajectories. This framework incorporates a Bayesian inference approach to actively learn the agents' parameters, i.e., other drivers' model parameters. The proposed framework employs a sampling-based approach that is suitable for real-time implementation through the utilization of GPUs. We illustrate the effectiveness of our proposed methodology through comprehensive numerical simulations conducted in both high and low-fidelity simulation scenarios focusing on autonomous on-ramp merging., Comment: 7 pages, 3 figures

In this paper, we develop a control framework for the coordination of multiple robots as they navigate through crowded environments. Our framework comprises of a local model predictive control (MPC) for each robot and a social long short-term memory model that forecasts pedestrians' trajectories. We formulate the local MPC formulation for each individual robot that includes both individual and shared objectives, in which the latter encourages the emergence of coordination among robots. Next, we consider the multi-robot navigation and human-robot interaction, respectively, as a potential game and a two-player game, then employ an iterative best response approach to solve the resulting optimization problems in a centralized and distributed fashion. Finally, we demonstrate the effectiveness of coordination among robots in simulated crowd navigation.

Crowd navigation has received increasing attention from researchers over the last few decades, resulting in the emergence of numerous approaches aimed at addressing this problem to date. Our proposed approach couples agent motion prediction and planning to avoid the freezing robot problem while simultaneously capturing multi-agent social interactions by utilizing a state-of-the-art trajectory prediction model i.e., social long short-term memory model (Social-LSTM). Leveraging the output of Social-LSTM for the prediction of future trajectories of pedestrians at each time-step given the robot's possible actions, our framework computes the optimal control action using Model Predictive Control (MPC) for the robot to navigate among pedestrians. We demonstrate the effectiveness of our proposed approach in multiple scenarios of simulated crowd navigation and compare it against several state-of-the-art reinforcement learning-based methods., Comment: 7 pages, 3 figures, 6 tables

This paper discusses the limitations of existing microscopic traffic models in accounting for the potential impacts of on-ramp vehicles on the car-following behavior of main-lane vehicles on highways. We first surveyed U.S. on-ramps to choose a representative set of on-ramps and then collected real-world observational data from the merging vehicle's perspective in various traffic conditions ranging from free-flowing to rush-hour traffic jams. Next, as our core contribution, we introduce a novel car-following model, called MR-IDM, for highway driving that reacts to merging vehicles in a realistic way. This proposed driving model can either be used in traffic simulators to generate realistic highway driving behavior or integrated into a prediction module for autonomous vehicles attempting to merge onto the highway. We quantitatively evaluated the effectiveness of our model and compared it against several other methods. We show that MR-IDM has the least error in mimicking the real-world data, while having features such as smoothness, stability, and lateral awareness., Comment: 8 pages, 9 figures, 1 table

A lane-change maneuver on a congested highway could be severely disruptive or even infeasible without the cooperation of neighboring cars. However, cooperation with other vehicles does not guarantee that the performed maneuver will not have a negative impact on traffic flow unless it is explicitly considered in the cooperative controller design. In this letter, we present a socially compliant framework for cooperative lane-change maneuvers for an arbitrary number of CAVs on highways that aims to interrupt traffic flow as minimally as possible. Moreover, we explicitly impose feasibility constraints in the optimization formulation by using reachability set theory, leading to a unified design that removes the need for an iterative procedure used in prior work. We quantitatively evaluate the effectiveness of our framework and compare it against previously offered approaches in terms of maneuver time and incurred throughput disruption., Comment: 6 pages, 2 figures

We derive optimal control policies for a Connected Automated Vehicle (CAV) and cooperating neighboring CAVs to carry out a lane change maneuver consisting of a longitudinal phase where the CAV properly positions itself relative to the cooperating neighbors and a lateral phase where it safely changes lanes. In contrast to prior work on this problem, where the CAV "selfishly" only seeks to minimize its maneuver time, we seek to ensure that the fast-lane traffic flow is minimally disrupted (through a properly defined metric). Additionally, when performing lane-changing maneuvers, we optimally select the cooperating vehicles from a set of feasible neighboring vehicles and experimentally show that the highway throughput is improved compared to the baseline case of human-driven vehicles changing lanes with no cooperation. When feasible solutions do not exist for a given maximal allowable disruption, we include a time relaxation method trading off a longer maneuver time with reduced disruption. Our analysis is also extended to multiple sequential maneuvers. Simulation results show the effectiveness of our controllers in terms of safety guarantees and up to 16% and 90% average throughput and maneuver time improvement respectively when compared to maneuvers with no cooperation., Comment: arXiv admin note: substantial text overlap with arXiv:2203.17102

In this paper, we extend a framework that we developed earlier for coordination of connected and automated vehicles (CAVs) at a signal-free intersection by integrating a safety layer using control barrier functions. First, in our motion planning module, each CAV computes the optimal control trajectory using simple vehicle dynamics. The trajectory does not make any of the state, control, and safety constraints active. A vehicle-level tracking controller employs a combined feedforward-feedback control law to track the resulting optimal trajectory from the motion planning module. Then, a barrier-certificate module, acting as a middle layer between the vehicle-level tracking controller and physical vehicle, receives the control law from the vehicle-level tracking controller and using realistic vehicle dynamics ensures that none of the state, control, and safety constraints becomes active. The latter is achieved through a quadratic program, which can be solved efficiently in real time. We demonstrate the effectiveness of our extended framework through a numerical simulation., Comment: 8 pages, 5 figures, 1 table

In this letter, we consider a transportation network with a 100\% penetration rate of connected and automated vehicles (CAVs) and present an optimal routing approach that takes into account the efficiency achieved in the network by coordinating the CAVs at specific traffic scenarios, e.g., intersections, merging roadways, and roundabouts. To derive the optimal route of a travel request, we use the information of the CAVs that have already received a routing solution. This enables each CAV to consider the traffic conditions on the roads. The solution of any new travel request determines the optimal travel time at each traffic scenario while satisfying all state, control, and safety constraints. We validate the performance of our framework through numerical simulations. To the best of our knowledge, this is the first attempt to consider the coordination of CAVs in a routing problem., Comment: 6 pages, 5 figures

Vehicle platooning using connected and automated vehicles (CAVs) has attracted considerable attention. In this paper, we address the problem of optimal coordination of CAV platoons at a highway on-ramp merging. We present a single-level constrained optimal control framework that optimizes fuel economy and travel time of the platoons while satisfying the state, control, and safety constraints. We also explore the effect of delayed communication among the CAV platoons and propose a robust coordination framework to enforce lateral and rear-end collision avoidance constraints in the presence of bounded delays. We provide a closed-form analytical solution of the optimal control problem with safety guarantees that can be implemented in real time. Finally, we validate the effectiveness of the proposed control framework using a high-fidelity commercial simulation environment.

The steady increase in the number of vehicles operating on the highways continues to exacerbate congestion, accidents, energy consumption, and greenhouse gas emissions. Emerging mobility systems, e.g., connected and automated vehicles (CAVs), have the potential to directly address these issues and improve transportation network efficiency and safety. In this paper, we consider a highway merging scenario and propose a framework for coordinating CAVs such that stop-and-go driving is eliminated. We use a decentralized form of the actor-critic approach to deep reinforcement learning$-$multi-agent deep deterministic policy gradient. We demonstrate the coordination of CAVs through numerical simulations and show that a smooth traffic flow is achieved by eliminating stop-and-go driving. Videos and plots of the simulation results can be found at this supplemental $\href{https://sites.google.com/view/ud-ids-lab/MADRL}{\text{site}}$., Comment: 6 pages, 6 figures

In this paper, we investigate the problem of a last-mile delivery service that selects up to $N$ available vehicles to deliver $M$ packages from a centralized depot to $M$ delivery locations. The objective of the last-mile delivery service is to jointly maximize customer satisfaction (minimize delivery time) and minimize operating cost (minimize total travel time) by selecting the optimal number of vehicles to perform the deliveries. We model this as an assignment (vehicles to packages) and path planning (determining the delivery order and route) problem, which is equivalent to the NP-hard multiple traveling salesperson problem. We propose a scalable heuristic algorithm, which sacrifices some optimality to achieve a reasonable computational cost for a high number of packages. The algorithm combines hierarchical clustering with a greedy search. To validate our approach, we compare the results of our simulation to experiments in a $1$:$25$ scale robotic testbed for future mobility systems., Comment: 7 pages, 8 figures

Deriving optimal control strategies for coordination of connected and automated vehicles (CAVs) often requires re-evaluating the strategies in order to respond to unexpected changes in the presence of disturbances and uncertainties. In this paper, we first extend a decentralized framework that we developed earlier for coordination of CAVs at a signal-free intersection to incorporate replanning. Then, we further enhance the framework by introducing a priority-aware resequencing mechanism which designates the order of decision making of CAVs based on theory from the job-shop scheduling problem. Our enhanced framework relaxes the first-come-first-serve decision order which has been used extensively in these problems. We illustrate the effectiveness of our proposed approach through numerical simulations., Comment: 6 pages, 4 figures

Emerging mobility systems, e.g., connected and automated vehicles (CAVs), shared mobility, and electric vehicles, provide the most intriguing opportunity for enabling users to better monitor transportation network conditions and make better decisions for improving safety and transportation efficiency. However, before connectivity and automation are deployed en masse, a thorough evaluation of CAVs is required-ranging from numerical simulation to real-world public roads. Assessment of the performance of CAVs in scaled testbeds has recently gained momentum due to the flexibility they offer to conduct quick, repeatable experiments that could go one step beyond simulation. This article introduces the Information and Decision Science Lab's Scaled Smart City (IDS$^3$C), a 1:25 research and educational scaled robotic testbed that is capable of replicating different real-world urban traffic scenarios. IDS$^3$C was designed to investigate the effect of emerging mobility systems on safety and transportation efficiency. On the educational front, IDS$^3$C can be used for (a) training and educating graduate students by exposing them to a balanced mix of theory and practice, (b) integrating the research outcomes into existing courses, (c) involving undergraduate students in research, (d) creating interactive educational demos, and (e) reaching out to high-school students. In our exposition, we also present a real-time control framework that can be used to coordinate CAVs at traffic scenarios such as crossing signal-free intersections, merging at roadways and roundabouts, cruising in congested traffic, passing through speed reduction zones, and lane-merging or passing maneuvers. Finally, we provide a tutorial for applying our framework in coordinating robotic CAVs to a multi-lane roundabout scenario and a transportation corridor in IDS$^3$C., Comment: 35 pages, 9 figures

The increasing demand for emerging mobility systems with connected and automated vehicles has imposed the necessity for quality testing environments to support their development. In this paper, we introduce a Unity-based virtual simulation environment for emerging mobility systems, called the Information and Decision Science Lab's Scaled Smart Digital City (IDS 3D City), intended to operate alongside its physical peer and its established control framework. By utilizing the Robot Operation System, AirSim, and Unity, we constructed a simulation environment capable of iteratively designing experiments significantly faster than it is possible in a physical testbed. This environment provides an intermediate step to validate the effectiveness of our control algorithms prior to their implementation in the physical testbed. The IDS 3D City also enables us to demonstrate that our control algorithms work independently of the underlying vehicle dynamics, as the vehicle dynamics introduced by AirSim operate at a different scale than our scaled smart city. Finally, we demonstrate the behavior of our digital environment by performing an experiment in both the virtual and physical environments and comparing their outputs., Comment: 6 pages, 8 figures

In this paper, we extend a framework that we developed earlier for coordination of connected and automated vehicles (CAVs) at a signal-free intersection to incorporate uncertainty. Using the possibly noisy observations of actual time trajectories and leveraging Gaussian process regression, we learn the bounded confidence intervals for deviations from the nominal trajectories of CAVs online. Incorporating these confidence intervals, we reformulate the trajectory planning as a robust coordination problem, the solution of which guarantees that constraints in the system are satisfied in the presence of bounded deviations from the nominal trajectories. We demonstrate the effectiveness of our extended framework through a numerical simulation., Comment: 8 pages, 6 figures

Connected and automated vehicles (CAVs) can alleviate traffic congestion, air pollution, and improve safety. In this paper, we provide a decentralized coordination framework for CAVs at a signal-free intersection to minimize travel time and improve fuel efficiency. We employ a simple yet powerful reinforcement learning approach, an off-policy temporal difference learning called Q-learning, enhanced with a coordination mechanism to address this problem. Then, we integrate a first-in-first-out queuing policy to improve the performance of our system. We demonstrate the efficacy of our proposed approach through simulation and comparison with the classical optimal control method based on Pontryagin's minimum principle., Comment: 8 pages, 5 figures, 2 tables

In this paper, we establish a decentralized optimal control framework for connected and automated vehicles (CAVs) crossing multiple adjacent, multi-lane signal-free intersections to minimize energy consumption and improve traffic throughput. Our framework consists of two layers of planning. In the upper-level planning, each CAV computes its optimal arrival time at each intersection recursively along with the optimal lane to improve the traffic throughput. In the low-level planning, we formulate an energy-optimal control problem with interior-point constraints, the solution of which yields the optimal control input (acceleration/deceleration) of each CAV to cross the intersections at the time specified by the upper-level planning. Moreover, we extend the results of the proposed bi-level framework to include a bounded steady-state error in tracking the optimal position of the CAVs. Finally, we demonstrate the effectiveness of the proposed framework through simulation for symmetric and asymmetric intersections and comparison with traditional signalized intersections., Comment: 14 pages, 9 figures, 5 tables

Roundabouts in conjunction with other traffic scenarios, e.g., intersections, merging roadways, speed reduction zones, can induce congestion in a transportation network due to driver responses to various disturbances. Research efforts have shown that smoothing traffic flow and eliminating stop-and-go driving can both improve fuel efficiency of the vehicles and the throughput of a roundabout. In this paper, we validate an optimal control framework developed earlier in a multi-lane roundabout scenario using the University of Delaware's scaled smart city (UDSSC). We first provide conditions where the solution is optimal. Then, we demonstrate the feasibility of the solution using experiments at UDSSC, and show that the optimal solution completely eliminates stop-and-go driving while preserving safety., Comment: 6 Pages, 4 Figures, 1 table

In this paper, we provide a hierarchical coordination framework for connected and automated vehicles (CAVs) at two adjacent intersections. This framework consists of an upper-level scheduling problem and a low-level optimal control problem. By partitioning the area around two adjacent intersections into different zones, we formulate a scheduling problem for each individual CAV aimed at minimizing its total travel time. For each CAV, the solution of the upper-level problem designates the arrival times at each zones on its path which becomes the inputs of the low-level problem. The solution of the low-level problem yields the optimal control input (acceleration/deceleration) of each CAV to exit the intersections at the time specified in the upper-level scheduling problem. We validate the performance of our proposed hierarchical framework through extensive numerical simulations and comparison with signalized intersections, centralized scheduling, and FIFO queuing policy., Comment: 17 pages, 7 figures, 3 tables

Energy consumption is the major contributor associated with large and growing transportation cost in logistics. Optimal vehicle routing approaches can provide solutions to reduce their operating costs and address implications on energy. This paper outlines a solution to the single-depot capacitated vehicle routing problem with the objective of minimizing daily operation cost with a homogeneous fleet of delivery vehicles. The problem is solved using Simulated Annealing, to provide optimal routes for the vehicles traveling between the depot and destinations. Simulation results demonstrate that the proposed approach is effective to recommend an optimal route and reduce operation cost. Supplementary information and video of our proposed approach can be found at: https://sites.google.com/view/ud-ids-lab/last-mile, Comment: Consists of 8 pages, containing 12 figures and 2 tables

In this article, we demonstrate a zero-shot transfer of an autonomous driving policy from simulation to University of Delaware's scaled smart city with adversarial multi-agent reinforcement learning, in which an adversary attempts to decrease the net reward by perturbing both the inputs and outputs of the autonomous vehicles during training. We train the autonomous vehicles to coordinate with each other while crossing a roundabout in the presence of an adversary in simulation. The adversarial policy successfully reproduces the simulated behavior and incidentally outperforms, in terms of travel time, both a human-driving baseline and adversary-free trained policies. Finally, we demonstrate that the addition of adversarial training considerably improves the performance \eat{stability and robustness} of the policies after transfer to the real world compared to Gaussian noise injection., Comment: 6 pages, 4 figures

The implementation of connected and automated vehicle (CAV) technologies enables a novel computational framework to deliver real-time control actions that optimize travel time, energy, and safety. Hardware is an integral part of any practical implementation of CAVs, and as such, it should be incorporated in any validation method. However, high costs associated with full scale, field testing of CAVs have proven to be a significant barrier. In this paper, we present the implementation of a decentralized control framework, which was developed previously, in a scaled-city using robotic CAVs, and discuss the implications of CAVs on travel time. Supplemental information and videos can be found at https://sites.google.com/view/ud-ids-lab/tfms.

In this paper, we provide a decentralized optimal control framework for coordinating connected and automated vehicles (CAVs) in two interconnected intersections. We formulate a control problem and provide a solution that can be implemented in real time. The solution yields the optimal acceleration/deceleration of each CAV under the safety constraint at "conflict zones," where there is a chance of potential collision. Our objective is to minimize travel time for each CAV. If no such solution exists, then each CAV solves an energy-optimal control problem. We evaluate the effectiveness of the efficiency of the proposed framework through simulation., Comment: 8 pages, 5 figures, IEEE CONFERENCE ON CONTROL TECHNOLOGY AND APPLICATIONS 2019

Using deep reinforcement learning, we train control policies for autonomous vehicles leading a platoon of vehicles onto a roundabout. Using Flow, a library for deep reinforcement learning in micro-simulators, we train two policies, one policy with noise injected into the state and action space and one without any injected noise. In simulation, the autonomous vehicle learns an emergent metering behavior for both policies in which it slows to allow for smoother merging. We then directly transfer this policy without any tuning to the University of Delaware Scaled Smart City (UDSSC), a 1:25 scale testbed for connected and automated vehicles. We characterize the performance of both policies on the scaled city. We show that the noise-free policy winds up crashing and only occasionally metering. However, the noise-injected policy consistently performs the metering behavior and remains collision-free, suggesting that the noise helps with the zero-shot policy transfer. Additionally, the transferred, noise-injected policy leads to a 5% reduction of average travel time and a reduction of 22% in maximum travel time in the UDSSC. Videos of the controllers can be found at https://sites.google.com/view/iccps-policy-transfer., Comment: To be published at the International Conference on Cyber Physical Systems (ICCPS) 2019. 10 pages, 9 figures