Presence Detection: Selection Criteria and Application Boundaries

Presence detection is the quiet first decision in any automated warehouse. Before a conveyor diverts, before a lift raises, before an operator receives a “load confirmed” prompt, something must answer one binary question: is an object actually here? Despite its simplicity, this answer is frequently wrong in practice — not because the sensor is defective, but because it was selected for the wrong context, mounted without regard for its environment, or interpreted through an incomplete understanding of the whole signal chain. This article describes the criteria that should drive sensor selection, the physical and logical boundaries of each sensing principle, and the diagnostic habits that keep a warehouse running after dust, film wrap, and temperature changes have done their worst.

The Operating Context of Presence Detection #

Presence detection lives at the interface between the physical object and the control system. The sensor converts a continuous physical condition — distance, reflectivity, capacitance, sound return — into a single discrete state: present or absent, blocked or clear. The controls team rarely cares how that conversion happens; they care only whether the PLC input turns on at the right moment and stays in the right state long enough for the program to act.

The operating context therefore includes much more than the sensor housing. The bracket, the cable, the connector, the input card, the scan cycle of the PLC, and the downstream actuator all participate in the same detection event. A presence fault observed as a missed divert is often caused by a loose bracket or a marginal power supply, not by the sensing element itself. When troubleshooting, regard the sensor as one component in a detection chain, not as the whole chain.

In a typical warehouse, presence sensors appear at merge points, accumulation lanes, lift gates, shuttle positions, pallet stations, and in front of every automatic door and safety-adjacent zone. At each of these locations the sensor is doing the same fundamental job with different physical constraints. That constraint difference is why no single sensing principle dominates the market.

Sensing Principles and Their Selection Profiles #

Each sensing principle behaves differently with different materials, surface conditions, optical properties, and atmospheric contamination. A short profile is useful before discussing selection criteria.

• Inductive sensors detect only conductive materials, in practice ferrous metals and some non-ferrous metals. They are immune to dust, moisture, and non-metal objects in the path. Their range is short, typically measured in millimetres to a few centimetres, and they are the natural choice for metal pallet detection, carriage position, and mechanical clamping verification.
• Capacitive sensors detect both conductive and non-conductive materials, including wood, cardboard, plastics, and liquids behind a non-metallic wall. They are easily confused by humidity, foam, and changes in material density, so their practical value in a dusty warehouse is limited to applications where nothing else can see the target.
• Photoelectric through-beam sensors place an emitter and a receiver in separate housings. The beam is interrupted by any opaque object, regardless of colour, reflectivity, or surface texture. They offer the longest range and the strongest signal margin, but they require two devices to be aligned and maintained in alignment over their full service life.
• Retroreflective photoelectric sensors house the emitter and receiver together and direct the beam to a reflector. They are easier to install than through-beam versions, but their reliability depends on the reflector staying clean and perpendicular. Polarised retroreflective models are available for detecting shiny or metallic objects that would otherwise confuse a standard unit.
• Diffuse photoelectric sensors rely on light reflected directly from the target. They require no separate reflector and are the least expensive option, but they are the most sensitive to target colour, gloss, angle, and background proximity. A white carton at 500 mm might trigger a diffuse sensor set for a grey carton at 200 mm.
• Ultrasonic sensors emit a sound pulse and measure its return. They detect nearly any material, including transparent film and dark foam, which optical sensors often miss. Their weaknesses are sensitivity to air temperature, turbulent airflow from fans, angled surfaces that deflect the beam, and sound-absorbing materials such as open-cell foam and fabric.

These profiles are not marketing categories; they are physical constraints that determine whether a sensor will work for years or fail weekly.

Component Interactions — The Signal Chain #

Between the sensor element and the final actuator response, several components contribute to the observable behaviour. The power supply must hold voltage under load, especially when the sensor cable runs twenty metres alongside a variable-frequency-drive cable. The connector and cable carry a signal that is small and easily distorted by poor terminaton. The input card applies a debounce or filtering time in many PLCs; a fast sensor output on a slow-filtered input becomes a slow, marginal signal. The PLC program then interprets the input with its own timing constraints, typically requiring the input to be stable for a minimum window before acting.

When a presence signal appears to “flutter” or “miss,” the sensor output may be perfectly healthy at the sensor’s own terminals while the PLC input image shows a different story. Comparing the two signals during an event is one of the fastest diagnostics available. This observation also reveals the most important interaction rule: the symptom is seen at the end of the chain, but the cause may be anywhere along it.

Selection Criteria by Application #

Object Material, Size, and Surface #

The starting point is always the object itself. A metal stillage on a roller conveyor is reliably seen by an inductive sensor, but if the stillage has a thin steel base plate, the sensing distance must be evaluated against the actual metal thickness and mounting gap. A cardboard carton is invisible to inductive sensors and is best served by a retroreflective or diffuse photoelectric unit. Transparent shrink-wrapped pallets defeat standard diffuse units; a polarised retroreflective sensor with a reflective tape target or a through-beam pair is the dependable choice.

Size matters as much as material. A through-beam can detect small boxes accurately because the beam is thin, but a diffuse sensor with a wide sensing cone may fail to detect a narrow carton edge because the reflected signal drops below threshold as the object moves through the beam edge. Always verify the minimum object size against the effective sensing spot at the intended mounting distance.

Environment and Contamination #

Warehouse air carries dust, cardboard fibres, exhaust from forklift engines, and airborne film particles from shrink wrap machinery. In such conditions, diffuse sensors suffer most because contamination on both the lens and the target background reduces or changes the reflected signal. Through-beam sensors tolerate dust well; a layer of dust reduces the beam intensity slightly, but the signal margin is usually large enough to absorb it. Retroreflective units sit between the two, provided the reflector is cleaned on the same schedule as the lens.

Temperature and humidity also shift sensor behaviour. Ultrasonic sensors are particularly sensitive to temperature gradients near loading doors. Capacitive sensors drift with humidity. When an application is near a washdown area or a dock door, choose a sensing principle with inherent immunity rather than compensating with sensitivity adjustments.

Speed, Timing, and Repeatability #

At a typical conveyor merge moving forty cartons per minute, the sensor has tens of milliseconds to respond. Most photoelectric and ultrasonic sensors have response times fast enough for this, but the system as a whole may not. A PLC input card with a long debounce filter, or a program scan that polls the input every 50 ms, can turn a perfectly fast sensor into a source of missed counts. Confirm the total response budget — sensor response time plus PLC filter time plus program scan plus actuator reaction time — and specify the sensor with margin for each stage.

Repeatability matters in lifting and tilting applications where a pallet must stop within a few millimetres of a fixed position. Some sensing principles have wider switching hysteresis than others. If the same pallet type produces a slightly different stop position each cycle, the issue may be the sensor’s hysteresis window rather than the mechanical brake.

Mounting Constraints and Cable Runs #

A through-beam sensor across a wide aisle gives excellent detection certainty, but it also creates a permanent alignment obligation. If the mounting points are inaccessible without a forklift platform, the maintenance team will inevitably slacken on alignment checks. In such positions, a retroreflective or diffuse sensor on a single bracket may be more sustainable