%0 Journal Article %T Ecological and safe driving: a model predictive control approach considering spatial and temporal constraints %J Transportation research part D: transport and environment %D 2019 %A Dehkordi, Sepehr G. %A Larue, Grégoire S. %A Cholette, Michael E. %A Rakotonirainy, Andry %A Rakha, Hesham A. %V 67 %N %P 208-222 %X Road transport plays an important role in the economy but generates huge costs in terms of road trauma and greenhouse gas emissions. The vast majority of previous research efforts have focused on driving behaviours that either minimise fuel consumption, or advise on safe driving behaviours. However, a fuel-optimal behaviour is often difficult to implement due to traffic conditions and may lead to unsafe behaviour without proper constraints. While a range of optimisation techniques are currently available, most lack the ability to both consider time and distance dependent constraints necessary for computationally efficient determination of optimal behaviour. In this paper, a methodology to find an ecological and safe (EcoSafe) driving behaviour by formulating and solving an optimal control problem for the minimum fuel driving behaviour while respecting key safety constraints was developed. We first propose a modified distance-based dynamic programming (MDDP) technique that can treat distance and time inequality constraints while optimising. Then we propose a model predictive control (MPC) structure based on MDDP and Pontryagin's maximum principle by considering inter vehicle time and time-to-collision as hard safety constraints. In contrast to existing approaches, the proposed MPC method does not require ad hoc tuning of constraint penalty weights in the cost function to ensure safe following distances. Simulation results of the proposed method show that the proposed algorithm produces lower vehicle fuel consumption compared to existing approaches when safety constraints are limiting. Thus, the results demonstrate the importance of considering safety constraints within an eco-driving system.

Language: en

%G en %I Elsevier Publishing %@ 1361-9209 %U http://dx.doi.org/10.1016/j.trd.2018.11.010