Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Safe Merging in Mixed Traffic with Confidence

Heeseung Bang, Aditya Dave, Andreas A. Malikopoulos

TL;DR

This letter designs a controller that ensures effective merging of CAVs with HDVs with safety guarantees and employs conformal prediction to obtain theoretical safety guarantees and uses real-world traffic data to validate the approach.

Abstract

In this letter, we present an approach for learning human driving behavior, without relying on specific model structures or prior distributions, in a mixed-traffic environment where connected and automated vehicles (CAVs) coexist with human-driven vehicles (HDVs). We employ conformal prediction to obtain theoretical safety guarantees and use real-world traffic data to validate our approach. Then, we design a controller that ensures effective merging of CAVs with HDVs with safety guarantees. We provide numerical simulations to illustrate the efficacy of the control approach.

Safe Merging in Mixed Traffic with Confidence

TL;DR

This letter designs a controller that ensures effective merging of CAVs with HDVs with safety guarantees and employs conformal prediction to obtain theoretical safety guarantees and uses real-world traffic data to validate the approach.

Abstract

In this letter, we present an approach for learning human driving behavior, without relying on specific model structures or prior distributions, in a mixed-traffic environment where connected and automated vehicles (CAVs) coexist with human-driven vehicles (HDVs). We employ conformal prediction to obtain theoretical safety guarantees and use real-world traffic data to validate our approach. Then, we design a controller that ensures effective merging of CAVs with HDVs with safety guarantees. We provide numerical simulations to illustrate the efficacy of the control approach.
Paper Structure (10 sections, 2 theorems, 14 equations, 5 figures)

This paper contains 10 sections, 2 theorems, 14 equations, 5 figures.

Table of Contents

  1. Introduction
  2. Modeling Framework
  3. Human Driving Behavior
  4. Learning Human Driving Behavior
  5. Training
  6. Prediction with Confidence
  7. Conformal Prediction for Vehicle Merging
  8. Safe Merging Strategy
  9. Numerical Simulations
  10. Concluding Remarks

Key Result

Proposition 1

Let $\hat{s}_k(t) = \sigma(h_k(t), t)$ be the hidden state of an LSTM from Section sec:learning for any $h_k(t)$ in the dataset $\mathcal{D}_h$, at any $t\in\mathcal{T}$. Then, in the corresponding dataset $\mathcal{D}_{\hat{s}} := \{(\hat{s}_k(t), \bm{\tau}_k: t \in \mathcal{T})\}_{k=1}^K$, the tup

Figures (5)

  • Figure 1: Control zone with one HDV on the highway. The dashed lines represent potential trajectories of the HDV and the CAV. The orange circles represent merging candidates.
  • Figure 2: Decision-making loop for a human driver.
  • Figure 3: Prediction of HDV's arrival time to merging candidates with our trained model. The prediction were made at time $t=0$ (left) and $t=T^\mathrm{m}/3$ (right).
  • Figure 4: Prediction and conformal prediction range for two merging candidates at each time step.
  • Figure 5: Simulation results for two different initial conditions.

Theorems & Definitions (8)

  • Remark 1
  • Remark 2
  • Proposition 1
  • proof
  • Remark 3
  • Remark 4
  • Proposition 2
  • proof