### Abstract

In this paper we study how a driver could use traffic light information in order to adapt driving speed profile to save fuel. The mission is given by a final destination to reach (through a set of traffic lights) within a specific deadline and the objective is to minimize the fuel consumption. We assume that the speed between each traffic light is constant and we do not take into account the effects of acceleration and gear shifting. Also, we use an existing model for the fuel consumption which depends quadratically on the speed of the vehicle. For simple cases (one traffic light), we derive analytical results using optimization theory and the Karush-Kuhn-Tucker (KKT) necessary conditions for optimality. For more complex and realistic cases, we use Dijkstra’s shortest path algorithm to discretize our decision problem. By “setting nodes” at each distance where there is a traffic light, we can model a realistic situation with an equivalent discrete graph with nonnegative edge costs. Then Dijkstra’s algorithm helps to find the shortest path through this set of nodes. The results indicate that the best strategy is to drive as close to an optimal speed as possible, which is dictated by the vehicle’s characteristics, while avoiding the traffic light blocks. One block look ahead strategies have also shown in simulations to be effective in particular realistic scenarios. It is also shown that, depending on the speed limit, minimal fuel and minimal time strategies may be quite similar.

Original language | English (US) |
---|---|

Pages (from-to) | 672-686 |

Number of pages | 15 |

Journal | WSEAS Transactions on Systems and Control |

Volume | 9 |

Issue number | 1 |

State | Published - Jan 1 2014 |

### Keywords

- Dijkstra’s algorithm
- Fuel minimization
- KKT
- Look ahead strategy
- Optimization
- Traffic light information

### ASJC Scopus subject areas

- Control and Systems Engineering
- Control and Optimization

## Fingerprint Dive into the research topics of 'Fuel minimization of a moving vehicle in suburban traffic'. Together they form a unique fingerprint.

## Cite this

*WSEAS Transactions on Systems and Control*,

*9*(1), 672-686.