Real-time predictions
Spot cascading delays minutes before they materialise and protect punctuality KPIs.
We empower operators to understand and prevent cascading delays network-wide and in real-time.
Value
Spot cascading delays minutes before they materialise and protect punctuality KPIs.
Surface the most impactful conflict resolutions in any delay chain regardless of the network complexity.
Share insights with stakeholders across infrastructure, operators, and regulators with confidence.
Use cases
From shunting yards to national networks - Oncron delivers measurable improvements wherever punctuality and efficiency matter.
Local decisions often have network-wide consequences. Minor disruptions, late arrivals, or suboptimal sequencing can quickly propagate into major delays — on the mainline as well as in yards and depots. Oncron detects and visualizes delay propagation before it happens and recommends AI-optimized decisions to contain disruptions early. This proactive approach can reduce total delay minutes by up to 10%, protecting punctuality KPIs and easing signallers' workloads.
Today's delay attribution is fragmented and costly. Oncron analyzes complete delay chains to reveal causal links between events and quantify their impact. Where data access allows, it can go beyond human capability in root cause analysis, creating a consistent, automated foundation for accountability and continuous improvement.
Even a theoretically stable timetable can fail in operation. Oncron analyzes live performance data to reveal where and why real-world conditions diverge from plan, uncovering structural weaknesses early. This gives planners and regulators a fact-based view of timetable feasibility and uncovers risks before annual planning updates.
Operational rules and reactions are often based on experience rather than consistent data. While individual expertise matters, operational excellence depends on making that knowledge available system-wide. Oncron anonymously aggregates decision patterns and outcomes to identify which actions consistently minimize delay propagation — without tracking individuals. By turning experience into institutional intelligence, it enables data-driven conflict resolution and measurable performance across control centers and shifts.
Approach
Our event-based, graph-driven architecture models entire rail networks in real time—simulating hours into the future within seconds.
We know railway data is never perfect. Oncron's algorithms run effectively on station-level or partial live feeds.
Oncron was founded by experts who've spent decades in railway, data science, and system integration. We understand the domain's unique rhythms, safety culture, and decision constraints.
Our Deep Reinforcement Learning approach explores countless operational scenarios to discover strategies even seasoned signallers might overlook. But it doesn't dictate — it recommends.