TY - JOUR
T1 - Safe Trajectory Optimization and Efficient-Offline Robust Model Predictive Control for Autonomous Vehicle Lane Change
AU - Nguyen, Hung Duy
AU - Kim, Dongryul
AU - Nguyen, Anh
AU - Han, Kyoungseok
AU - Vu, Minh Nhat
PY - 2024
Y1 - 2024
N2 - Driving autonomous vehicles through diverse roadconditions at various high speeds poses a significant challenge.To address this challenge, we propose a hierarchical controlstrategy consisting of an optimization-based trajectory plannerin the first layer and an efficient-offline robust method employingpath tracking in the second layer. Considering vehicle parametricuncertainties, the proposed hierarchical structure addressesmultiple scenarios with varying road surface conditions andvelocities. In the first layer, using the Pontryagin maximumprinciple (PMP) flexibly with the time-to-collision (TTC) method,the motion planner generates a safe-optimal lane-change trajectorywhen interacting with forward vehicles and adjacentlanevehicles to improve ride comfort while maintaining safedistances. In the second layer, the efficient offline robust modelpredictive control (RMPC) with terminal constraints is appliedto a linear parameter varying (LPV) system. Utilizing linearmatrix inequality (LMI) techniques, the optimization problemaccommodates parametric uncertainties while robustly satisfyinginput and output constraints. To emphasize superior performance,we have considered comparing our proposed approachwith several state-of-the-art methods. Therefore, comparativesimulation results have shown that our safe-optimal trajectorygeneration is better than the Spatio-Temporal Corridors methodregarding path smoothy. The proposed approach (i.e., efficientofflineRMPC) then outperforms the offline MPC method interms of path-tracking performance and parametric uncertaintyhandling while outperforming the online RMPC method in termsof computational complexity reduction. Further, all methods areverified using a co-simulation and testing platform via a highfidelitydynamics testing vehicle control software (i.e., CarSim).
AB - Driving autonomous vehicles through diverse roadconditions at various high speeds poses a significant challenge.To address this challenge, we propose a hierarchical controlstrategy consisting of an optimization-based trajectory plannerin the first layer and an efficient-offline robust method employingpath tracking in the second layer. Considering vehicle parametricuncertainties, the proposed hierarchical structure addressesmultiple scenarios with varying road surface conditions andvelocities. In the first layer, using the Pontryagin maximumprinciple (PMP) flexibly with the time-to-collision (TTC) method,the motion planner generates a safe-optimal lane-change trajectorywhen interacting with forward vehicles and adjacentlanevehicles to improve ride comfort while maintaining safedistances. In the second layer, the efficient offline robust modelpredictive control (RMPC) with terminal constraints is appliedto a linear parameter varying (LPV) system. Utilizing linearmatrix inequality (LMI) techniques, the optimization problemaccommodates parametric uncertainties while robustly satisfyinginput and output constraints. To emphasize superior performance,we have considered comparing our proposed approachwith several state-of-the-art methods. Therefore, comparativesimulation results have shown that our safe-optimal trajectorygeneration is better than the Spatio-Temporal Corridors methodregarding path smoothy. The proposed approach (i.e., efficientofflineRMPC) then outperforms the offline MPC method interms of path-tracking performance and parametric uncertaintyhandling while outperforming the online RMPC method in termsof computational complexity reduction. Further, all methods areverified using a co-simulation and testing platform via a highfidelitydynamics testing vehicle control software (i.e., CarSim).
KW - Trajectory optimization
KW - Pontryagin maximum principle
KW - robust model predictive control
KW - linear matrix inequality
KW - autonomous vehicles
U2 - 10.1109/tiv.2024.3467111
DO - 10.1109/tiv.2024.3467111
M3 - Article
SN - 2379-8858
JO - IEEE Transactions on Intelligent Vehicles
JF - IEEE Transactions on Intelligent Vehicles
ER -