Geofencing the East Rand: Automating Alerts for High-Risk Logistics Zones

The East Rand is the operational core of Gauteng freight. Kempton Park, Boksburg, Spartan and Isando sit alongside OR Tambo International and form a corridor where bonded warehouses, FMCG depots and cross-border consolidation centres run shoulder to shoulder. The East Rand is consistently named in cargo-crime reports as a high-exposure freight cluster. A circle drawn around a warehouse on a map is no longer an adequate boundary. Fleet operations need polygon-precise geofences that match the real shape of a yard, the real path of an authorised exit and the real perimeter of a known danger zone, with breach alerts that fire when a vehicle deviates rather than when an operator notices on a screen.

Direct answer

A geofence is a virtual boundary, defined by GPS coordinates, that triggers an alert when a tracked vehicle enters or leaves it. On the East Rand, polygon geofences (multi-sided shapes traced around warehouse footprints, cargo lanes and danger zones) outperform simple radius circles because they follow the real geometry of fenced yards, gate access points and high-risk side streets, producing fewer false positives and faster operator response.

Why polygon geofences matter for East Rand warehouse parks

A radius geofence is a circle. A polygon geofence is a hand-traced shape with as many sides as the operator needs. In the open veld a circle is fine. Inside the East Rand industrial belt it is not. A 200-metre radius around a warehouse on Pomona Road can include a competitor's yard, a public service road, a fuel station and two parking bays for unrelated tenants.

The warehouse parks of Kempton Park, Spartan and Boksburg are tight clusters of fenced sites. The relevant boundary is the perimeter fence, not a notional circle. A polygon traced along that fence line, including the gate, the inbound queue and the outbound staging area, gives the dispatcher a clean signal: a vehicle is inside the yard or outside it, on the access road or on the public road. False positives drop because the fence matches the asset.

A second reason matters more on the East Rand than elsewhere. The OR Tambo cargo precinct is bonded. A vehicle leaving a SARS-controlled area without clearance is a customs problem, not just a theft problem. The polygon must follow the bonded perimeter precisely so the breach event is legally meaningful when reviewed by SARS or by the consignee.

Three layer types every East Rand fleet should run

A working geofence stack uses three layer types in combination. Operators who run only one layer (typically authorised stop) miss the events that matter most: route deviation and entry into a known danger zone.

• Authorised stop: the warehouse, depot, fuel station or customer site where the vehicle is expected. Polygon traced to the fence line. Entry, exit and dwell time logged.
• Danger zone: known hijacking hot-spots, abandoned yards used for transhipment, side streets used by syndicates for vehicle swaps. Any entry triggers an alert.
• Route deviation corridor: a thin polygon along the authorised route between origin and destination. The vehicle should remain inside the corridor for the journey. Any exit triggers an alert.

A vehicle in trouble usually breaches all three layers in sequence. A dispatcher seeing the three breach events together has a verified incident, not a guess.

The OR Tambo, Kempton Park, Boksburg and Spartan freight cluster

Cargo around OR Tambo touches multiple stakeholders inside a few kilometres: bonded airline warehouses, consolidators, customs brokers, ground handlers and road carriers. A typical export load can pass through three polygon geofences in 20 minutes.

The geography matters for fence design. Kempton Park sits west of the airport, with warehouse parks along Pomona, Bredell and Bonaero. Spartan is south of the runways and contains automotive and FMCG depots. Boksburg lies further south-east, anchored by East Rand Mall logistics tenants and N12 corridor warehouses. Each cluster has its own access roads, its own gate-throughput patterns and its own exposure profile. A single fleet running all three clusters needs three polygon templates, not one.

The table below summarises typical fence design choices for each cluster.

How the breach alert reaches the dispatcher

A geofence breach is not useful if it sits in a log file. The breach must reach the dispatcher inside the same view that holds the GPS pin, the route history and the driver duress signal. That view is what TG Online provides for fleets running TG Tracking hardware on the East Rand.

When a vehicle crosses a polygon boundary, the device on the CAN bus reports the position to TG Online. The platform compares the position against the active fence layers (authorised, danger, deviation), flags the breach, and surfaces it in the dispatcher's incident view. The dispatcher sees the live map, the breach type, the driver's last reported status and any associated camera or duress event in a single pane. Tools such as TomTom Area Analytics and Wialon's geofencing module are widely used internationally for the same operational pattern, and TG Online's role is to wire those breach signals into a South African control room that already understands the corridor.

The operational value sits in three behaviours. Breach alerts fire when a vehicle deviates without a human pressing a button. Related signals (camera trigger, duress event, route variance) are correlated so the dispatcher reviews one incident, not three. The breach record is timestamped and stored for SAPS, SARS and insurer use later.

A 6-step protocol for designing a new East Rand geofence

Drawing a polygon is the easy part. Designing a fence that produces useful alerts is harder. The following protocol is the working sequence used by TG Tracking implementation teams when commissioning a new East Rand site.

1. Walk the site perimeter with a handheld GPS or phone-based mapping tool. Note the fence line, gate, inbound queue and outbound staging.
2. Trace the polygon along the fence line in TG Online, with vertices at every corner and gate.
3. Add the entry and exit gate as a sub-zone. Gate-loitering events are breaches that radius fences miss entirely.
4. Identify two danger-zone polygons within 5 kilometres of the site, drawn from cargo-crime intelligence or insurer reports.
5. Trace a route deviation corridor between the site and the next planned waypoint, typically 200 metres wide.
6. Test the layer set with a known-good run before commissioning. Adjust vertices until alerts fire only on real breaches.

The cost of a poorly drawn fence is alert fatigue. Dispatchers who see ten false breach alerts a day stop reading them.

Why circle geofences fail in dense industrial parks

Circle geofences are still useful in two settings: open road and rural sites. Inside a Kempton Park or Spartan warehouse cluster they fail in three ways.

• A circle around a warehouse includes neighbouring properties the operator does not control.
• A circle around a gate includes a public service road where unrelated traffic registers as breaches.
• A circle never aligns with a bonded customs perimeter, so a SARS-meaningful breach event cannot be produced.

Polygon fences solve all three. They take more time to draw at commissioning. They take less time to maintain over a year, because the false-positive rate is lower.

TG Online as the operational layer over the fence stack

TG Tracking's role in the East Rand is integrator. The hardware on the vehicle reports position. The platform compares position against polygon, danger and corridor layers. The dispatcher sees the breach in TG Online alongside the rest of the operational picture: polygon fences traced to the fence line with gate and staging sub-zones, danger zone overlays drawn from current cargo-crime intelligence, and route deviation corridors per planned trip with breach alerts surfaced in the same view as the GPS pin.

The point is not the platform alone. The point is that breach signals, GPS, camera triggers and driver duress events arrive in the same operator's hands at the same moment. That is what makes the response useful.

Next step: audit your East Rand fence layer this month

Three questions any operations manager running East Rand routes should answer this month:

• Are warehouse fences traced to the actual perimeter, or are they 200-metre circles?
• Is there a danger-zone overlay near each major site, drawn from current intelligence?
• Does a route deviation produce a dispatcher-visible alert inside the same view as the GPS map?

If any answer is "no" or "not sure", the fence layer is not delivering its operational value. TG Tracking offers a half-day fence audit for East Rand fleets running 20 or more vehicles.

Frequently asked questions

What is the difference between a geofence and a route corridor?

A geofence is a closed shape around a fixed asset such as a warehouse or a danger zone. A route corridor is a long, thin polygon traced along the authorised path between two assets. Both are polygons in TG Online. The geofence answers "is the vehicle where it should be?". The route corridor answers "did the vehicle take the right path to get there?".

How many vertices should a polygon geofence have?

Enough to follow the real fence line and no more. A typical East Rand warehouse fence uses 8 to 16 vertices: one at every corner, one at each gate, one at any kink in the perimeter. Dispatchers should avoid 50-vertex polygons because they slow the breach calculation without improving accuracy.

Can polygon geofences trigger an automatic engine cut-off?

In principle yes, in practice rarely. A breach event can route to a remote immobilisation workflow, but the cut-off itself should never fire without a human operator confirming the vehicle is stationary or in safe traffic. See our companion article on remote engine immobilisation for the procedural detail.

How do bonded customs polygons differ from ordinary fences?

A bonded polygon must follow the SARS-controlled perimeter exactly because a breach is a customs event, not just an operational one. The polygon should be drawn from the bonded yard layout supplied by the operator, not from a satellite image. SARS may request the polygon definition during audit, so the source data must be defensible.

What happens to fence data if the vehicle loses GPS signal?

The device buffers position internally and reports breach events when signal returns. For long signal-loss periods the breach record shows a delayed timestamp; TG Online flags this so dispatchers do not treat the late alert as a current event. Multi-constellation GPS reduces the risk.

How often should a fence layer be reviewed?

Quarterly for active operational sites, more often after any major route or warehouse change. Cargo-crime hot spots move; danger-zone overlays drift out of date inside three months. A fence layer set once and never reviewed produces false confidence rather than useful alerts.

Is geofence data acceptable as evidence in insurance claims?

Yes when the breach record is timestamped, the polygon definition is stored alongside the event and the data chain is auditable. Underwriters increasingly request fence breach logs as part of incident packs. TG Online retains breach data for the operator's contracted retention period.

Can a single vehicle carry multiple active geofences at the same time?

Yes. A typical East Rand load runs inside a route corridor, passes through several authorised stop polygons and skirts at least two danger zones. The platform tracks all active fences simultaneously and reports breach events as they occur.