Trinity College Dublin

Skip to main content.

Top Level TCD Links

Self-organising Motorway Traffic Management

Congestion costs the EU 100 billion Euro per year, while reliable transportation systems are required for just-in-time production and distribution systems. On top of that, lane closures reduce the capacity of a motorway and cause delays and unreliable travel-time estimations.

lego01

The goal of this project is to design an algorithm that allows autonomous vehicles to self-organise to mitigate the effects of a lane closure. Once the design is accomplished, an evaluate of the algorithm through simulation is performed. A further step done in order to investigate the practicalities of such a system is the implementation of the algorithm on robots.

lego 02

Self-organising Motorway

Automated highways tend to increase motorway capacity. One way of automatization is slot-based driving, where vehicles are assigned a position on the road, with a centralised controller co-ordinating traffic. The problem is that existing technologies mainly focus on safety, through systems like adaptive cruise control and lane departure warning.

Self-organisation is a process where a structure or pattern appears in a system without a central authority or external element imposing it through planning (no centralised control or changes to infrastructure)

Novel Self-organising Algorithm

Our algorithm proposes autonomous vehicles categorised by lane number. The vehicles are enabled for vehicle-to-vehicle communication, while being organized in virtual slots. These virtual slots designate a grid around each vehicle, and as a result the vehicle should maintain its position relative to other vehicles.

Simulation

Simulation Software -VISSIM

We have devised a microscopic multi-modal traffic flow simulation with a set scenario. The scenario proposes a three-lane motorway that merges into two lanes after 2km.

simulation 01 Fig 1. 5000 vehicles per hour -Human drivers simulation 02 Fig 2. 5000 vehicles per hour –Self-organising algorithm

Evaluation

Our simulations estimate how human drivers perform vs. vehicles controlled by the self-organizing algorithm when facing increased volumes of traffic.

Result

The road reaches capacity at 5000 vehicles/hour. Travel times are highly reduced when employing the self-organization algorithm for vehicle control.

Implementation on Robots

Lego MindstormsNXT

MindstormsNXT has a programmable robotics kit, with open-source firmware. We added sets of advanced sensors and bluetooth communication

Hierarchical Control System

The control system is implemented as through a self-organising behaviour modelled on nodes, with feedback received from sensors input into nodes. The output from nodes commands actuators.

Conclusion

Based on the findings of this research we can conclude that self organisation can reduce the delays due to traffic congestion and provide more reliable travel time estimations.

People

Niall O'Hara, Dan Marinescu, Suzy Temate, Mélanie Bouroche, Siobhán Clarke, Vinny Cahill

Sponsors


Last updated 12 November 2013 futurecities@tcd.ie (Email).