The wire in the road
In New South Wales, detector loops are set into the road surface near the stop line, generally in the middle of the lane, close behind the line.1 A loop is a coil of wire; the metal in a car changes its inductance, and the controller reads that change as a vehicle.2
Stop on a loop and a demand is placed for your movement. Roll back off it and the controller may no longer know you are there.1 In the drawing, the dashed boxes behind each stop line are the loops. They darken when a car is over them.
Sydney, 1933
Sydney's first traffic lights went into operation at 11am on Friday 13 October 1933, at the corner of Kent and Market Streets. They were already vehicle-actuated, on a two-phase system, and many drivers ignored them at first.3 They are also described as Australia's first.10b
Two phases is what this junction runs too: north-south, then east-west.
The clock
A fixed-time controller gives each movement a set amount of time every cycle, whatever arrives.13 A classic starting point is Webster's formula, C₀ = (1.5L + 5) ÷ (1 − Y), where L is the lost time, usually the sum of the intergreens, and Y is the sum of the critical flow ratios.11
A queue leaves a green light at about one car every two seconds per lane.12 New NSW signals are specified with a nominal 140 s cycle, and in 2018 part of the Sydney CBD was cut from 110 s to 90 s.10b
Detectors deciding
An actuated controller holds a green for a minimum, then keeps it while cars keep arriving. If none arrives within the passage time, the phase gaps out. If another road is waiting and the green reaches its maximum, it maxes out.9
US guidance puts passage times for presence detection at about 1 to 4 s.9 With nobody waiting on the other road, a controller can ignore the maximum and stay green.13 Here: 6 s minimum, 3 s passage, 45 s maximum.
SCATS
The Sydney Coordinated Adaptive Traffic System was developed by the NSW Department of Main Roads and dates from 1975. It now runs at more than 65,000 intersections in 218 cities and 33 countries.4
Arthur Sims was systems manager at the Roads and Traffic Authority, and Peter Lowrie a traffic engineer there, when SCATS was developed.6 Lowrie’s lasting contribution was measuring the empty time between cars on the loop — what SCATS calls non-occupancy — rather than how long cars sit on it.6b
Measuring the gaps
SCATS prefers a detector at the stop line in every lane.7,7b From the empty time between cars it works out a degree of saturation: how much of the green was used by moving traffic rather than empty road.7b,8
Regional computers set cycle length, splits and offsets, in small steps almost cycle by cycle. The local controller can end a phase early or skip one nobody is waiting for.5,7 Transport for NSW’s 2024–25 annual report counts 4,860 traffic signals across the Sydney road network.10
What this model leaves out
One lane each way, no turns, no pedestrians, no buses and no neighbouring junctions. Every driver follows the same car-following rule, the Intelligent Driver Model.14 Amber is 4 s, the low end of the NSW range of 4.0 to 6.4 s,10b then 2 s of all-red.
The SCATS-style controller is a sketch of the published principles, not SCATS. It aims to keep the busier road about 90 per cent used, keeps its cycle between 40 and 120 s (an illustrative model range; published SCATS practice usually sets a lower limit around 30–40 s and an upper limit around 100–150 s),7b and moves it by up to 6 s at a time. Real SCATS also coordinates neighbouring junctions through offsets,5 which one junction cannot show.