Microsoft Research Seminar

Back in December I gave a seminar at Microsoft Research Cambridge into my research interests around Lots of Humans In the Loop (LHIL) control. The seminar was recorded and can be viewed on Microsoft's site using the link below.

Lots of Humans In the Loop

human control

An experiment was carried out at a proving ground. 30 cars were asked to drive around a figure-of-eight circuit with a traffic light controlled junction in the centre. There were four runs of this experiment. In the first three the junction was controlled by an optimised fixed-time controller. In the fourth a human "controller" switched the lights remotely from a laptop

Read more about the experiment

Download this paper to read more about the experiments and its results.

Get the data

Lots of data were recorded during this experiment from GPS sensors within the vehicles and from track-side bluetooth detectors. If you're interested in the data you can download it

 

Scalextric Track

Scalextric Track

The scalextric with traffic control is an educational tool designed to teach people about some of the technologies behind today's traffic control systems in a fun and interactive way. It was built for an exhibit at the Royal Society Summer science exhibition. But it also pops up at other science outreach events like Science and engineering day at the University of Southampton.

You can watch the videos below which demonstrate how the system works on and early prototype.

 

Want to build your own?

The inductive loop sensors used in this set were made out of a coil of wire with 13 turns. This is connected to an amplifying circuit using LMC662CN Op-Amp. The circuit diagram is below. The traffic lights are three LED's red, yellow and green. The circuit for these is below also.

Scalextric Track

Scalextric Track

The inductive loop sensors and traffic lights are connected up to an Arduino micro-controller. The Arduino is programmed with the junction control logic. You can download the Arduino code here.

Any questions/problems you can always contact me.

For an introduction to my research in Machine Learning methods for signal control please see my lecture to the University's Transport Study Group (UTSG) conference at the Open University in January 2011 this is linked below (in two parts).

For the really fine detail please see my publications

Go!

For my day job I work in the Transportation Research Group (TRG) at the University of Southampton. I am interested in how technology, information and automation can be employed to build a more efficient and sustainable transport system.

My current research is about how we can use Machine Learning methods to build traffic light junction control systems that can learn strategies from Human experts and through experience. Find out more about this here.

I am also interested in fun and engaging ways to inform people about transportation research and the issues we are working on. In 2011 I was involved in building a number of interactive games for an exhibit at the Royal Society Summer science exhibition. Find out about my Scalextric set fitted with traffic control here.

For access to my transport related publications please click here.