HighlightsRolling horizon stochastic optimal control strategy for both ACC and CACC.Uncertainties in system dynamics and sensor measurements are considered.Normally-distributed additional disturbances to state and measurement equations.Multi-Objective function with bounded acceleration limit and collision protection.Linearly constrained LQG problem is solved using a separation principle.
AbstractThis paper presents a rolling horizon stochastic optimal control strategy for both Adaptive Cruise Control and Cooperative Adaptive Cruise Control under uncertainty based on the constant time gap policy. Specifically, uncertainties that can arise in vehicle control systems and vehicle sensor measurements are represented as normally-distributed disturbances to state and measurement equations in a state-space formulation. Then, acceleration sequence of a controlled vehicle is determined by optimizing an objective function that captures control efficiency and driving comfort over a predictive horizon, constrained by bounded acceleration/deceleration and collision protection. The optimization problem is formulated as a linearly constrained linear quadratic Gaussian problem and solved using a separation principle, Lagrangian relaxation, and Kalman filter. A sensitivity analysis and a scenario-based analysis via simulations demonstrate that the proposed control strategy can generate smoother vehicle control and perform better than a deterministic feedback controller, particularly under small system disturbances and large measurement disturbances.
Rolling horizon stochastic optimal control strategy for ACC and CACC under uncertainty
2017-07-24
16 pages
Article (Journal)
Electronic Resource
English
Multi-Model Adaptive Control for CACC Applications
IEEE | 2021
|Stability of CACC-manual heterogeneous vehicular flow with partial CACC performance degrading
Taylor & Francis Verlag | 2019
|Configurable Aerospace Command and Control (CACC)
NTIS | 1999
|Truck CACC Implementation and Test to Verify Control Performance
Transportation Research Record | 2019
|