Proposal Summary and Objectives

Freight traffic affects the performance of the road network significantly due to their different driving characteristics compared with passenger cars. For example, trucks need extra braking distances and time for deceleration and are slower when starting up. Recently, connected and automated vehicle technology has presented possibilities to control trucks intelligently to improve safety and mobility. Strategies have been developed in previous FMRI projects to formulate multiple trucks’ trajectories considering mixed traffic conditions. The stability problem of vehicle streams has been studied in the third-year project. However, in previous research, deterministic behaviors of human-driven vehicles (HV) in mixed traffic were assumed. In reality, truck drivers may act differently (accelerating, decelerating, reacting) according to their varying driving habits, which will lead to uncertainties in vehicle dynamics and affect overall traffic behavior. To account for this, we propose to consider stochastics behaviors of human driven vehicles while developing control strategies for automated trucks. This proposed project will investigate the method for CAV truck controlling in a truck-only lane with CAV trucks and human-drive trucks. By controlling CAV trucks, the mixed traffic is stabilized and optimized in its mobility given different levels and distribution types of the uncertainty in human-driven vehicles. Two parts make up the modeling part: longitudinal behaviors uncertainty modeling of human driven trucks and the stochastic optimal control development of CAV trucks. The expected outcomes are: First, the proposed scenarios with different types and levels of uncertainty are modeled in simulation. Second, a control method for CAVs to optimize the traffic in terms of safety and stability performances. The expectation is that the CAV trucks as well as mixed traffic flow will operate with collision-free requirements while traveling with reduced travel time, compared to baseline when only a deterministic controller is adopted. Due to the limitation of time (one year), the research will only focus on theoretical modeling. In the implementation stage in the future, local agencies or industry partners may be involved.

Funding Amount:
Status: Active
Duration: Sep 1, 2021 - Jul 30, 2022