Distributed Nonlinear Model Predictive Control and Metric Learning for Heterogeneous Vehicle Platooning with Cut-in/Cut-out Maneuvers

Abstract: Vehicle platooning has been shown to be quite fruitful in the transportation industry to enhance fuel economy, road throughput, and driving comfort. Model Predictive Control (MPC) is widely used in literature for platoon control to achieve certain objectives, such as safely reducing the distance among consecutive vehicles while following the leader vehicle. In this paper, we propose a Distributed Nonlinear MPC (DNMPC), based upon an existing approach, to control a heterogeneous dynamic platoon with unidirectional topologies, handling possible cut-in/cut-out maneuvers. The introduced method addresses a collision-free driving experience while tracking the desired speed profile and maintaining a safe desired gap among the vehicles. The time of convergence in the dynamic platooning is derived based on the time of cut-in and/or cut-out maneuvers. In addition, we analyze the improvement level of driving comfort, fuel economy, and absolute and relative convergence of the method by using distributed metric learning and distributed optimization with Alternating Direction Method of Multipliers (ADMM). Simulation results on a dynamic platoon with cut-in and cut-out maneuvers and with different unidirectional topologies show the effectiveness of the introduced method.

• The paper introduces a novel Distributed Nonlinear Model Predictive Control framework that enables dynamic platoon management in heterogeneous vehicle groups.
• It incorporates real-time metric learning using ADMM to optimize safety, fuel economy, and driving comfort during cut-in/cut-out maneuvers.
• Simulations validate timely convergence and stability, demonstrating the system's capability to adapt to rapid changes in platoon composition.

Distributed Nonlinear Model Predictive Control and Metric Learning for Heterogeneous Vehicle Platooning with Cut-in/Cut-out Maneuvers

Introduction

This paper presents a detailed analysis and implementation of a Distributed Nonlinear Model Predictive Control (DNMPC) system for managing platoons of heterogeneous vehicles performing dynamic maneuvers such as cut-ins and cut-outs. The system is designed to maintain collision-free operation, track desired speed profiles, and ensure safe inter-vehicle distances using various unidirectional communication topologies.

Figure 1: Unidirectional topology: (a) PF, (b) PLF, (c) TPF, and (d) TPLF, with vehicle 0 as the LV.

System Modeling

The system models the longitudinal dynamics of a platoon consisting of a Lead Vehicle (LV) and multiple Follower Vehicles (FVs). Each vehicle's dynamics are described by its position, velocity, and integrated driving torque. The model incorporates vehicle-specific parameters such as mass, drag coefficient, and tire radius.

Distributed Nonlinear Model Predictive Control

The control strategy utilizes DNMPC with prediction and control horizons to optimize vehicle control inputs. This optimization adheres to constraints ensuring desired inter-vehicle distances and velocity tracking while accommodating dynamic platoon changes due to cut-in and cut-out maneuvers.

Dynamic Platoon Control

The authors extend DNMPC to handle dynamic platooning scenarios, addressing challenges associated with heterogeneous vehicle characteristics and communication topologies. The paper derives convergence times for the system following dynamic maneuvers, ensuring quick adaptation to changes in platoon composition while maintaining system stability and performance.

Metric Learning and Driving Experience Analysis

The paper leverages distributed metric learning to assess driving experience metrics such as comfort and fuel economy. This is achieved through ADMM optimization, ensuring the learning process aligns with theoretical stability guarantees provided by DNMPC.

Optimization Formulation

Key to this approach is formulating an optimization problem that encapsulates control objectives and constraints across vehicles, ensuring the optimization aligns with stability constraints dictated by the system's underlying control strategy.

Simulations

The authors conduct simulations to validate their approach, using a scenario of a platoon subjected to cut-in and cut-out maneuvers. Results demonstrate the system's capability to efficiently manage dynamic transitions in platoon structure and maintain stability.

• Convergence Analysis: The simulations confirm predicted convergence times, indicating the system's effectiveness in ensuring timely adaptation to dynamic changes.
• Driving Experience: The study uses metric learning to project driving dynamics onto learned subspaces, highlighting key experiential factors such as smoothness and control efficacy.

Conclusion

The research establishes a comprehensive framework for managing heterogeneous vehicle platoons with DNMPC, integrating metric learning to enhance system performance. The paper suggests future work could explore string stability and lateral dynamics to further enhance platoon control strategies.

In conclusion, the advancements in DNMPC outlined in this paper hold significant promise for improving vehicle platooning in dynamic traffic environments. The integration of metric learning with traditional control approaches offers a novel pathway to optimizing driving experience and operational efficiency in autonomous vehicle systems.