VORTEX

Infrastructure built for human reflexes is the bottleneck.

With autonomous fleets at critical density, traffic-light infrastructure, designed around human reaction times, becomes the single biggest brake on urban mobility, with social costs of congestion in the hundreds of millions a year. Coordinating over a hundred thousand vehicles is an NP-hard problem that classical solvers cannot handle beyond a few hundred variables; the tolerable latency between trajectory update and actuation rules out any centralised cloud architecture; and a coordination error in a dense intersection is not recoverable. This calls for a paradigm shift, not incremental optimisation.

Quantum-annealing coordination

A Qiskit quantum layer runs annealing on D-Wave hardware to solve macro-clusters of over a thousand vehicles in parallel, shrinking the solution space by orders of magnitude versus classical methods.

Distributed safety-critical edge

Hundreds of Go edge nodes embedded across the road network ingest V2X telemetry at millisecond intervals and apply deterministic local corrections within the latency budget, staying operational in degraded mode even during a network partition.

Zone-partitioned fault isolation

Kafka orchestration with geographically partitioned topics ensures a fault in one zone stays contained, with no cascading failure across the network.

Real-time control room

A WebGL renderer with GPU instancing displays over a hundred thousand trajectories at 60 fps on consumer hardware, processing millions of segments per second to give operators a live view of the whole city.

Sense

Edge nodes gather V2X telemetry from every vehicle at millisecond intervals, building the instantaneous state of the network.

Optimise

The quantum layer formulates the problem as a QUBO and solves macro-clusters via annealing, returning globally coordinated trajectories.

Correct

Edge nodes apply deterministic local corrections within the latency budget, guaranteeing safety even between optimisation cycles.

Actuate

Coordinated trajectories reach the vehicles with no traffic lights, keeping flow smooth and continuous through every intersection.

Standards and functional safety

The edge layer is designed with functional-safety principles (IEC 61508) and verifiable determinism; platform management is aligned to ISO 27001, given its role as critical urban infrastructure.

Data sovereignty

V2X telemetry is processed locally on the edge nodes and does not leave the city perimeter for the safety function; only aggregated, anonymised data feeds quantum optimisation and analytics.

Access and audit

RBAC for control-room operators and an immutable audit trail of every manual override, optimisation cycle and degraded-mode transition, essential for post-incident analysis on critical infrastructure.

Encryption and integrity

AES-256 encryption at rest and TLS 1.3 for V2X telemetry and orchestration channels; V2X messages are authenticated to prevent malicious injection into the coordination loops.

What happens if the quantum layer is unavailable?

The safety-critical edge layer keeps operating in degraded mode with deterministic local corrections: it guarantees intersection safety even without global coordination. The quantum layer optimises flow, but safety does not depend on it.

Does a fault in one zone stop the whole city?

No. Kafka orchestration uses geographically partitioned topics, so a fault stays contained to its zone without cascading. Fault isolation is an architectural principle, not an add-on.

Is quantum really needed or is it marketing?

Coordinating over a hundred thousand vehicles is an NP-hard problem that classical solvers cannot handle beyond a few hundred variables. D-Wave annealing shrinks the solution space by orders of magnitude for macro-clusters; where density is low, a classical solver still suffices.

Does it work without removing existing traffic lights?

VORTEX is built for autonomous fleets at critical density where traffic lights become the bottleneck; it can, however, operate in coexistence during the transition, coordinating V2X-enabled vehicles while traditional infrastructure remains for the rest.

How is it protected from malicious V2X messages?

V2X messages are authenticated and channels encrypted with TLS 1.3, so an actor cannot inject false telemetry into the coordination loops; deterministic edge corrections additionally filter out inconsistent states before actuation.