A missed call at a signalised junction can add delay to every following movement. A poorly located counter can distort an entire survey. A guide to traffic sensor procurement therefore needs to start with the operational decision the sensor must support, not with a preferred technology or a like-for-like replacement for an existing inductive loop.
For highways authorities, consultants and contractors, the strongest specification connects detection performance to safer roads, reduced congestion and reliable evidence. It also recognises site realities: limited road closure opportunities, constrained power and communications, vulnerable road users, difficult weather and the need to maintain the asset over many years.
Start with the traffic outcome, not the device
Traffic sensors are often procured as discrete components, yet their value is determined by what happens after detection. At a junction, the requirement may be dependable presence and passage detection to extend a green stage, call a side-road phase or protect a cyclist movement. On a strategic route, the priority may be accurate vehicle classification, speed, occupancy or queue data for network management. For a road safety scheme, it may be early warning of approaching vehicles or evidence of speeds before and after an intervention.
These are different use cases. A sensor that provides highly effective vehicle presence detection may not deliver the classification detail required for a traffic study. Equally, a counter designed for long-term volume and speed data may not be the right interface for a signal controller. Define the decision, the road users to be detected, the performance measures and the action triggered by the data before selecting equipment.
A practical requirement statement should establish whether the system must detect cars, vans, lorries, buses, motorcycles, cyclists or pedestrians. It should state the lanes, approaches and detection zones involved, including turning lanes and shared-use areas. It should also distinguish between a need to detect movement, presence, direction, speed, headway, queue length or classification. This removes ambiguity before suppliers interpret the requirement in different ways.
Guide to traffic sensor procurement: choose the right detection method
Non-intrusive technologies can remove a major source of disruption associated with road-embedded loops. There is no need to cut the carriageway, reinstate surfacing or return to site when an embedded loop fails. That can reduce installation risk, shorten works and avoid unnecessary traffic management, particularly at busy junctions and on recently resurfaced roads.
The right approach still depends on the site. Radar is well suited to vehicle detection, speed measurement and selected bicycle applications, with performance that is not dependent on visible light. It can be a strong choice where reliable above-ground detection and minimal carriageway intervention are priorities. Its detection zone, mounting position and ability to separate road users need to be assessed against the geometry of the approach.
AI-powered video detection provides a different capability. Properly specified, it can identify and track multiple road users across complex areas, supporting vehicle, cyclist and pedestrian detection as well as richer traffic analytics. It is particularly useful where a junction layout requires adaptable virtual detection zones rather than fixed physical loops. However, procurement teams should assess camera position, occlusion from street furniture or high-sided vehicles, lighting conditions, network connectivity and data governance from the outset.
Wireless traffic sensors can be appropriate where rapid deployment or temporary monitoring is required, while fixed above-ground solutions may be preferable for long-term control applications. There is no universal winner. The specification should allow an equivalent technology where it can prove that it meets the operational outcome, installation constraint and required data quality.
Specify performance in measurable terms
A specification that simply requests a “traffic detector” invites assumptions. Procurement documents should set out measurable acceptance criteria that reflect the intended operation.
For traffic signal applications, consider detection reliability by vehicle type, minimum and maximum detection distance, detection-zone configuration, response time, hold behaviour and compatibility with the controller interface. For data collection, define accuracy expectations for volume, speed, direction and classification, along with the reporting interval, timestamp standard and treatment of incomplete records.
Accuracy must also be considered in context. A single-lane rural approach has different challenges from a multi-lane urban corridor with buses, cyclists, parked vehicles and frequent turning movements. Ask suppliers to explain expected performance in the actual environment, not only under favourable test conditions. Where vulnerable road users are part of the requirement, make their detection a stated and testable criterion rather than an implied feature.
Environmental resilience deserves equal attention. The equipment should be suitable for its mounting location and withstand the expected range of weather, vibration, contamination and roadside exposure. Specify any requirements for ingress protection, operating temperature, surge protection and electromagnetic compatibility. Also establish how the sensor continues to operate, degrades or reports faults during poor visibility, low light, heavy rain or communications loss.
Treat installation as part of the technical evaluation
Installation method has a direct operational impact. A solution requiring carriageway excavation can affect programme certainty, road user safety and disruption well beyond the device itself. Above-ground detection can instead be mounted on existing poles, mast arms or purpose-designed street furniture, subject to suitable sightlines and structural assessment.
The supplier should provide a clear site survey process that confirms mounting height, orientation, detection coverage, cable routes, available power and communications. At complicated junctions, a desk assessment alone may be insufficient. Site-specific design is essential where detector fields overlap, visibility is restricted or the layout includes bus lanes, cycle facilities and pedestrian crossings.
Commissioning should not be treated as a handover formality. Require installation drawings, detector-zone plans, configuration records and a defined process for validating detection against live traffic. The team responsible for operating the system should be able to understand what has been configured and why. This is particularly valuable when junction staging, road markings or traffic patterns change later.
Assess integration, ownership and data quality
A traffic sensor is only as useful as its connection to the wider system. Confirm the required interfaces early, whether the output is to a traffic signal controller, a local display, a roadside unit, a cloud platform or an existing traffic management system. Open and documented interfaces reduce the risk of being tied to an opaque integration arrangement.
For analytics deployments, clarify how data is transmitted, stored, retained and exported. Highways teams should be able to retrieve usable data without lengthy manual processing. Confirm whether the platform supports the formats, reporting periods and aggregation needed for asset management, scheme evaluation and operational decision-making.
Video-based systems require particular care. The procurement process should distinguish between video used for detection and video retained for viewing or audit. Define the intended use, access controls, retention arrangements and responsibilities for compliance with applicable data protection requirements. A system that performs valuable anonymous object detection may not need to retain identifiable imagery, but this should be designed rather than assumed.
Data quality also needs ongoing attention. Ask how the system identifies sensor health issues, communication failures, obstructed views and implausible readings. A useful platform should make faults visible quickly and support remote diagnostics where possible. If data informs policy, enforcement-adjacent activity or safety interventions, establish a repeatable validation process rather than relying on one commissioning check.
Evaluate whole-life operational value
The best procurement decision is not necessarily the sensor with the longest feature list. It is the solution that delivers dependable detection at the required locations with manageable installation, clear integration and support that matches the authority’s capability.
During evaluation, request evidence from comparable road environments. Examine how configuration changes are made, whether additional detection zones can be created without physical works and what expertise is available during design, commissioning and fault investigation. These practical factors determine whether the asset continues to perform when the network changes.
A useful evaluation should consider at least five areas:
- detection performance against defined road-user and traffic conditions;
- installation requirements, including traffic management and carriageway intervention;
- integration with existing controllers, communications and data systems;
- maintainability, diagnostics and configuration access; and
- supplier technical competence, site support and evidence of comparable deployments.
Weight these criteria according to the scheme objective. For a resurfacing programme, avoiding future carriageway cuts may carry particular importance. For an active travel corridor, cyclist and pedestrian detection performance may be the deciding factor. For network monitoring, classification accuracy and data accessibility may take precedence.
Build testing into the procurement process
Factory documentation is useful, but live-road validation is where a specification becomes operational reality. For higher-risk or more complex sites, include a demonstration, pilot or clearly defined site acceptance test. The test should cover representative traffic conditions and the road users that matter most to the scheme.
Agree the test method before installation. This might compare sensor outputs with manual observations, surveyed traffic data or a defined reference source over an agreed period. Record the configuration used during testing so that later performance can be traced to a known baseline. If the sensor is intended to support signal operation, verify the actual controller response as well as the detector output.
Acceptance should include training and documentation. Engineers should know how to check status, interpret alarms, adjust authorised settings and escalate faults. A technology partner with practical traffic management experience can make this stage considerably more effective, because commissioning issues are often caused by site geometry, controller logic or operating assumptions rather than by the sensor alone.
The most effective traffic sensor procurements leave room for technical innovation while remaining exact about the outcome. Specify the movement you need to detect, the decision that follows and the evidence required to prove performance. That approach gives project teams a sound basis for selecting non-intrusive technology that keeps traffic moving, improves road safety and avoids turning routine detection into another carriageway maintenance problem.