AINEWS
Moreton Bay to pilot Australia's first movement-based AI traffic signals
Moreton Bay's Australian first pilot replaces pre set phase cycles with sensors that classify road users and allocate green time to the next movement in real time.
Moreton Bay's Australian first pilot replaces pre set phase cycles with sensors that classify road users and allocate green time to the next movement in real time.
The City of Moreton Bay Council is preparing an Australian-first pilot of an AI traffic-control system at the intersection of Moreton Parade and Paper Avenue in Petrie, replacing pre-timed phase cycles with controllers that allocate green time to individual road users in real time (Brisbane Times).
The change moves traffic signals from a baton relay to a per-movement model. Under the conventional setup used across most of Australia, every direction gets a pre-set slice of green in a repeating cycle, sometimes adjusted by induction loops or adaptive systems such as SCATS that tweak the split. The Moreton Bay hardware does not run a cycle at all. Each classified road user, from a car turning left to a pedestrian stepping off the kerb, becomes a discrete movement with its own trajectory and priority (Inside Local Government). The controller decides which movement gets the next slice of green.
A pedestrian finishing a late crossing can hold the walk phase a beat longer. A single truck waiting to turn right can hold opposing left-turners without losing the cross-street through movement. A cyclist approaching on a green-painted lane can get a protected arrow on detection alone. The Petrie pilot will show whether the algorithm delivers those gains, or whether it ends up approximating a smarter version of conventional adaptive control.
This is also the architectural pattern autonomous-vehicle intersection protocols are designed to work with, a model in which signals stop broadcasting a fixed green-amber-red sequence and start negotiating the next movement for each classified road user. The Moreton Bay system uses road-side sensors rather than vehicle-to-infrastructure radios, but the underlying model of traffic is the same.
New controller boxes will replace existing traffic-signal hardware at Petrie, paired with sensors that classify approaching road users and predict traffic flow. Mayor Peter Flannery has framed the trial as a potential gamechanger for motorists and pedestrians (Inside Local Government). The council plans a staged rollout: validate the system at Petrie first, then escalate to a more complex intersection if phase one succeeds (Brisbane Times).
Moreton Bay is not the only Queensland council testing AI signals. Brisbane City Council runs its Smarter Suburban Corridors program, and ABC Brisbane reports Brisbane is separately trialling AI-integrated lights on its own network. Smarter Suburban Corridors targets corridor-level flow across stretches of road rather than per-movement intersection control, and Brisbane's separate trial sits within that corridor frame. The two architectures are siblings, not the same system, and neither council has published a shared performance methodology.
Pilots like Petrie will shape the standard question before any standard does. Brisbane's corridor program and Moreton Bay's movement-based trial run in parallel, with separate data sets, separate vendor relationships and separate council evaluation cycles. If early results diverge, Queensland risks the same fragmentation seen under legacy adaptive deployments: councils running their own logic, with no shared baseline for what AI traffic management means at the road level.
That divergence is where the governance question starts. A Smiths Lawyers analysis of Queensland's broader AI traffic-light push flags which standards body accredits the algorithm, what audit trail exists when a controller overrides a phase cycle, and what happens to the per-road-user classification data the new sensors generate. Austroads, the national road-management standards body, has not published guidance specific to movement-based controllers, and the Queensland Department of Transport and Main Roads has not named the safety case Moreton Bay is being held to.
For now there are no public performance KPIs: no delay-reduction figure, no pedestrian wait-time baseline, no modelled crash-outcome projection. The gamechanger framing remains a council-stated position, not an independent finding. What is logged, how long classification data is retained, who can access it, which vendor is behind the controller, and on what training basis the algorithm operates are also absent from the public material. Whether a movement-based pilot at Petrie becomes a movement-based network, or stops at one intersection, will depend on those answers and on the KPIs the council commits to publishing from phase one. No completion date has been announced.