A false call at a signal-controlled junction is rarely just a minor detector fault. It can introduce unnecessary stages, extend cycle time, delay genuine traffic and weaken confidence in the data used to manage the network. Knowing how to reduce false traffic calls starts with treating detection as an operational system, not simply a device installed at the kerbside.
For highways authorities and signal engineers, the objective is straightforward: detect the intended road user reliably, reject irrelevant activity, and provide the controller with a demand that reflects what is actually happening on the approach. Achieving that outcome requires good site design, the right sensing technology and disciplined commissioning.
Why false traffic calls affect network performance
A false traffic call occurs when a detector sends an actuation to the controller although no relevant vehicle, cyclist or pedestrian is present in the intended detection area. The effect depends on the junction strategy. At a lightly trafficked side road, it may call an empty stage and interrupt main-road progression. At a busy urban junction, repeated false demands can add lost time to every cycle and contribute to avoidable queues.
The consequences are wider than delay. Unnecessary green stages increase stop-start driving, fuel use and emissions. They can also complicate signal timing reviews because controller logs appear to show demand where none existed. If the problem affects cyclist detection or pedestrian facilities, poorly targeted calls can make it harder to distinguish genuine active-travel demand from detector noise.
False calls should not be confused with missed calls. A missed call fails to detect a legitimate user, while a false call detects something that should be ignored. Both matter, but their remedies can differ. Increasing sensitivity may solve a missed detection issue while making false actuation more likely. The correct setting is therefore the one that supports the junction’s operational purpose, rather than simply the highest possible detection rate.
Identify what is generating the demand
The first step is to establish the pattern before changing any settings. Controller event logs can show when demands occur, but site observation and detector diagnostics are usually needed to identify why. A short review during different traffic, weather and lighting conditions often reveals whether the issue is constant, intermittent or linked to a particular movement.
Common causes include:
- Detection zones extending into an adjacent live lane, turning pocket, footway or cycle track.
- Vehicles in a neighbouring lane being detected because of poor mounting position, beam geometry or zone design.
- Foliage, signs, street furniture, parked vehicles or moving shadows entering a video detection area.
- Radar reflections from roadside objects, large vehicles or unsuitable alignment.
- Incorrect delay, extension or presence settings causing a brief event to be interpreted as sustained demand.
A detector may be operating exactly as configured while still producing the wrong operational result. This distinction matters. The fault may lie in the detection objective, the area being monitored or the logic used by the controller, rather than the hardware itself.
Observe the approach, not only the detector
Site observations should follow the actual traffic movements that trigger a call. For example, if a right-turn phase is being called unnecessarily, determine whether the trigger is a vehicle in the adjacent ahead lane, a vehicle waiting beyond the stop line, or a passing bus in an opposing lane. Each points to a different correction.
It is also useful to compare peak and off-peak operation. A detector that performs well in free-flow conditions may produce unwanted calls when queues spill back, when larger vehicles occupy nearby space or when low sun affects a camera view. Recording time-stamped video or detector events can provide evidence for configuration changes and later validation.
Start with detector siting and coverage
The most effective way to reduce false traffic calls is often to correct the physical detection layout. Detection zones should match the approach lane, stop-line position and control strategy. A zone that is too broad may capture adjacent movements. One that starts too far upstream can register traffic that will not reach the stop line during the current signal cycle.
Above-ground detection offers a practical advantage here. AI video and radar detectors can usually be adjusted, re-aimed or reconfigured without cutting the carriageway. This allows engineers to refine coverage after observing live operation, avoiding the disruption and maintenance exposure associated with road-embedded inductive loops.
For video detection, camera height, angle and field of view directly affect classification and zone accuracy. The installation must provide a clear view of the relevant lanes while minimising occlusion from signal heads, trees, tall vehicles and roadside equipment. A camera placed for broad network visibility is not automatically positioned for precise lane-by-lane actuation.
For radar, alignment, mounting height and the selected detection field are equally important. Radar is particularly valuable where darkness, glare or difficult weather conditions challenge optical detection, but it must still be configured to reject movements outside the intended approach. Narrower, well-defined coverage may be preferable to a broad field that creates ambiguous targets.
Configure detection logic for the signal strategy
A reliable detector can still create false calls if its outputs are not matched to the controller logic. Engineers should review whether each output is intended to provide pulse, presence, extension, queue or vehicle-class information. These functions should not be treated as interchangeable.
A short pulse may be suitable for registering an approaching vehicle, whereas presence detection is more appropriate where a vehicle must be held at the stop line until it receives service. If a presence output persists after a vehicle has departed, or an extension time is excessive, the controller may continue to act on stale demand. Conversely, overly aggressive filtering can remove genuine calls from cyclists, motorcycles or slower vehicles.
Classification adds useful control where the junction strategy requires it. Video analytics can distinguish relevant road users and help prevent large vehicles, pedestrians or background motion from being interpreted as a standard traffic demand. The practical benefit depends on scene quality and the clarity of the classification task. At complex junctions, separate zones and outputs for separate movements are usually easier to validate than one large, general-purpose area.
Use filtering carefully
Most modern above-ground detectors offer settings that can reduce nuisance activations, such as minimum target size, dwell time, direction of travel and exclusion zones. These controls should be used to remove known unwanted behaviour, not as a substitute for correct siting.
For instance, a minimum dwell time can reject a brief radar return from passing traffic, but setting it too high may fail to recognise a cyclist approaching a red signal. Direction filtering can prevent an opposing movement from calling a phase, provided the detector has a reliable view of travel direction. The best parameters depend on speed environment, lane width, approach geometry and the user groups the junction must serve.
Commission in real operating conditions
Desk-based configuration is only the beginning. Commissioning should include live checks of the whole detection-to-controller chain: detector event, input mapping, demand registration, stage selection and demand cancellation. A correctly detected vehicle is of little value if the controller input is assigned to the wrong phase or remains active beyond the intended period.
Test cases should include representative vehicle types, cyclists where relevant, queued traffic, lane changes and turning movements. In urban areas, observe buses, refuse vehicles and delivery vehicles, as these can create occlusion or radar reflections that are absent during a simple test run. Where the site is affected by low sun, heavy rain or seasonal vegetation, plan a follow-up review rather than assuming a single commissioning visit proves long-term performance.
Performance should be measured against operational outcomes. Useful indicators include the number of empty stages called, frequency of demand during known quiet periods, unnecessary green time, queue discharge and any increase in missed calls after adjustments. This creates an evidence base for deciding whether a change has genuinely improved junction operation.
Maintain detector performance as the site changes
Traffic environments do not stand still. New signage, resurfacing, road layout changes, vegetation growth and altered lane use can all affect detection. A junction that was correctly configured at installation can become prone to false calls months later because the scene has changed.
Routine inspection should therefore include the detector’s physical condition, lens cleanliness for video equipment, mounting stability, line of sight and current zone configuration. Detector health information and event data should be reviewed alongside controller fault reports, particularly where operators identify unexplained stage calls or deterioration in network performance.
For schemes with a high operational impact, periodic revalidation is worthwhile after major changes to signal timings, approach markings or adjacent development access. This is particularly relevant where a junction supports priority measures, active travel facilities or coordinated urban corridors, where a small amount of false demand can affect a wider sequence of signals.
The most dependable approach is to make every call explainable. When the detection area, sensor behaviour and controller logic are aligned, false calls become easier to diagnose and far less likely to compromise safety, capacity or public confidence in the network.