A Cooperative Optimal Control Framework for Connected and Automated Vehicles in Mixed Traffic Using Social Value Orientation

Le, Viet-Anh; Malikopoulos, Andreas A. · 2022 · OpenAlex-citations

DOI: 10.1109/cdc51059.2022.9993337

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the motion planning challenge for Connected and Automated Vehicles (CAVs) operating in mixed traffic environments alongside Human-Driven Vehicles (HDVs). Motivated by the reality that full CAV penetration is unlikely in the near future, the authors seek to improve coordination by incorporating human social preferences into control strategies. Specifically, the study utilizes Social Value Orientation (SVO), a psychological metric quantifying an agent’s level of selfishness versus altruism, to model interactions. Unlike previous approaches that rely on computationally expensive Stackelberg games or assume fixed leader-follower dynamics, this work proposes a socially cooperative optimal control framework based on a simultaneous game and potential game theory. The methodology formulates the interaction between a CAV and an HDV as a simultaneous game where each vehicle minimizes a weighted sum of an egoistic objective (individual goals) and a cooperative objective (shared safety/efficiency goals). The weights are determined by the vehicles' SVO angles. The authors prove that a Nash equilibrium for this game can be obtained by minimizing a single potential objective function. To handle the uncertainty of human behavior, the framework employs a receding horizon control strategy. Crucially, it includes a moving horizon estimation algorithm based on maximum entropy inverse reinforcement learning (IRL) to estimate the HDV’s SVO angle in real-time using observed trajectory data. The CAV then adjusts its own SVO angle to compensate for the HDV’s level of altruism or egoism, ensuring robust coordination. The proposed framework is validated through numerical simulations of a vehicle merging scenario. The HDV’s behavior is simulated using a model learned from the Next Generation Simulation (NGSIM) dataset, allowing for the replication of both egoistic and altruistic driving styles. The results demonstrate that the CAV successfully estimates the HDV’s SVO angle and adapts its control actions accordingly. In scenarios with egoistic HDVs, the CAV adopts a more altruistic stance to facilitate safe merging, while it adjusts differently for altruistic HDVs. The simulations confirm that the potential game approach effectively yields a Nash equilibrium and that the IRL-based estimation accurately captures human driving intentions, enabling safe and efficient coordination without requiring explicit communication of intent from the human driver. The significance of this work lies in its ability to integrate psychological models of human behavior into rigorous optimal control frameworks for autonomous driving. By using SVO and potential games, the authors provide a computationally efficient alternative to Stackelberg equilibria for mixed traffic scenarios. The approach allows CAVs to predict and adapt to human driving styles dynamically, enhancing safety and traffic flow in environments where automated and human-driven vehicles coexist. This contributes to the broader field of autonomous systems by demonstrating how social compatibility can be mathematically formalized and implemented in real-time control algorithms.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• situational awareness