Provably safe and human-like car-following behaviors: Part 2. A parsimonious multi-phase model with projected braking

Wen-Long Jin · 2025

Read Provably safe and human-like car-following behaviors: Part 2. A parsimonious multi-phase model with projected braking by Wen-Long Jin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-09988.txt.

Introduction

Wen-Long Jin builds in 2025 a car-following model that is provably safe yet drives like a person. Automated cars need both hard safety proof and human-like speed patterns. Plain model extensions with capped braking fail that double test.

Projected braking profiles for leader and follower cars anchor the new design. A nominal model handles calm driving, while a control law takes over when braking looms. Phase plane study proves crash-free runs with bounded speed changes. The lean model keeps the proof set small enough to check fully. Speed-spacing phase maps carry the safety proof in the study.

Catalog record

CREATORWen-Long Jin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-09988

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Provably safe and human-like car-following behaviors: Part 2. A parsimonious multi-phase model with projected braking Wen-Long Jin Abstract Ensuring safe and human-like trajectory planning for automated vehicles amidst real-world uncertainties remains a critical challenge. While existing car-following models often struggle to consistently provide rigorous safety proofs alongside human-like acceleration and deceleration patterns, we introduce a novel multi-phase projection-based car-following model. This model is designed to balance safety and performance by incorporating bounded acceleration and deceleration rates while emulating key human driving principles. Building upon a foundation of fundamental driving principles and a multi-phase dynamical systems analysis (detailed in Part 1 of this study ( Jin 2025 ) ), we first highlight the limitations of extending standard models like Newell’s with simple bounded deceleration. Inspired by human drivers’ anticipatory behavior, we mathematically define and analyze projected braking profiles for both leader and follower vehicles, establishing safety criteria and new phase definitions based on the projected braking lead-vehicle problem. The proposed parsimonious model combines an extended Newell’s model for nominal driving with a new control law for scenarios requiring projected braking. Using speed-spacing phase plane analysis, we provide rigorous mathematical proofs of the model’s adherence to defined safe and human-like driving principles, including collision-free operation, bounded deceleration, and acceptable safe stopping distance, under reasonable initial conditions. Numerical simulations validate the model’s superior performance in achieving both safety and human-like braking profiles for the stationary lead-vehicle problem. Finally, we discuss the model’s implications and future research directions. Keywords : Safe and human-like driving principles; Newell’s simplified car-following model; Lead-vehicle problems; Speed-spacing phase plane analysis; Projected braking; Multi-phase projection-based car-following model. 1 Introduction The planning stage in driving, for both human-driven and automated vehicles, is paramount for ensuring safety amidst a complex and dynamic environment ( Arkin 1998 , Chapter 6) . It bridges perception with action, strategizing trajectories based on myriad factors including safety, comfort, vehicle limits, and traffic laws. While advancements in sensing and AI are significant ( Shalev-Shwartz et al. 2017 ) , the core challenge in planning lies in developing driving models that are not only computationally efficient but also provably safe and exhibit human-like behaviors, attributes often found to be conflicting in existing car-following literature. Part 1 of this study ( Jin 2025 ) introduced a multi-phase dynamical systems analysis framework and applied it to standard car-following models (Newell’s, Intelligent Driver Model, Gipps), revealing their limitations in consistently satisfying a comprehensive set of safe and human-like driving principles, particularly concerning braking dynamics and absolute safety guarantees under bounded deceleration. For instance, naive extensions of simple models like Newell’s to include bounded deceleration can lead to safety violations, the Gipps model is ill-defined when the follower is within the safe jam spacing and inconsistent with the observed fundamental diagram, and the Intelligent Driver model exhibits backward traveling and excessive stopping distance. This motivates the need for novel car-following models that are both parsimonious and can verifiably ensure safety while replicating nuanced human driving patterns. Human drivers inherently employ projection, anticipating potential future actions of lead vehicles, especially during braking, and planning their own maneuvers accordingly. This proactive behavior is crucial for maintaining safety margins. While the concept of safe stopping distance has been explored ( Gipps 1981 ; Shalev-Shwartz et al. 2017 ) , a rigorous, parsimonious car-following model that explicitly incorporates projected braking dynamics into a multi-phase control structure and other human-like behaviors with provable safety guarantees has been lacking. This article (Part 2) proposes such a solution: a parsimonious multi-phase projection-based car-following model, building upon the foundational principles and analytical framework established in Part 1. Our contributions are threefold: 1. We mathematically formalize the concept of projected braking for both leader and follower, defining safety criteria and new driving phases (nominal driving and comfort braking) based on these projections and their relationship to minimum and comfort jam spacings. 2. We develop a novel multi-phase projection-based car-following model that integrates an extended Newell’s model for nominal driving conditions with a new control law specifically designed for scenarios demanding projected braking, ensuring adherence to bounded acceleration and deceleration. 3. We provide rigorous mathematical proofs demonstrating that this model is collision-free and adheres to other critical safe and human-like driving principles under realistic initial conditions, offering a significant improvement over existing models analyzed in Part 1. The proposed model’s parsimony aims for easier calibration and potential for real-world learning, while its multi-phase projection-based nature allows for both human-like nominal driving and robustly safe emergency responses. Numerical simulations validate the theoretical analysis and compare the model’s braking profiles and safety performance against alternatives. We conclude by discussing implications and future extensions for this approach in the pursuit of advanced vehicular automation. The remainder of this article is organized as follows. Section 2 briefly revisits key definitions and the principles of safe and human-like driving behaviors pertinent to this work, as well as the extended Newell’s car-following model, drawing from Part 1. Section 3 establishes the foundation for our projection-based car-following model, analyzing the projected braking lead-vehicle problem, formulating projection-based safety principles, and defining new phases. Section 4 introduces our multi-phase projection-based car-following model, presenting mathematical proofs of its adherence to safe and human-like driving principles. Section 5 presents numerical simulations and analysis, including the model’s performance in the stationary lead-vehicle problem and its braking profile characteristics. Section 6 concludes the study, discussing potential future research directions and the implications of our findings. 2 Brief review of variables, behavioral principles, and extended Newell’s car-following model This section provides a concise overview of essential car-following variables, key behavioral principles for safe and human-like driving, and the extended Newell’s car-following model that will serve as a component of our proposed multi-phase model. For a comprehensive discussion, derivation, and detailed analysis of these concepts and standard car-following models, the reader is referred to Part 1 of this study ( Jin 2025 ) . A list of essential notations used in this article is given in Table 1 . Table 1: Essential Notation and Description for Part 2 Notation Description B(t) Safe stopping distance V(\cdot) Speed-density relation v=V(k) X(t) , X_{L}(t) Location of follower, leader at time t a(t) Acceleration rate of the follower at time t k(t) Density ( k(t)=\frac{1}{z(t)} ) q(t) Flow-rate t Time t^{\prime} Projected time v(t) , v_{L}(t) Speed of follower, leader at time t v_{*}(t) Equilibrium speed of the follower at t z(t) Spacing of the follower ( X_{L}(t)-X(t) ) \Delta t Time-step size \alpha Comfort acceleration bound of the follower \beta Comfort deceleration bound of the follower \beta_{L} Projected deceleration rate of the leader \epsilon Infinitesimal time step ( =\Delta t in discrete version) \kappa Jam density ( =\frac{1}{\zeta} ) \mu Speed limit \tau Minimum time gap (for nominal driving, e.g., Newell’s) \tau^{\prime} Reaction time of follower during braking \zeta Comfort jam spacing \zeta^{\prime} Minimum jam spacing \zeta_{c} Critical spacing ( =\tau\mu+\zeta ) \tilde{B}(t) Projected braking distance of follower \tilde{X}(t^{\prime},t) , \tilde{X}_{L}(t^{\prime},t) Projected location of follower, leader at t^{\prime} during braking (from current time t ) \tilde{X}(t) , \tilde{X}_{L}(t) Projected stopping location of follower, leader (from current time t ) \tilde{z}(t) Projected final spacing when vehicles stop ( \tilde{X}_{L}(t)-\tilde{X}(t) ) \tilde{z}(t^{\prime},t) Projected spacing at future projected time t^{\prime} (based on state

Read the complete text