Rail Data Transfer Challenge

01 Vision

Modern inspection trains capture track, overhead line and infrastructure condition with high-accuracy sensors, producing large image/video/sensor datasets. The core challenge is not data capture, but reliable, automated, scalable transfer to land-based systems. With the DB Rail Data Transfer Challenge (with DB InfraGO AG), we invite startups and tech companies to develop end-to-end approaches for automated offloading of very large measurement datasets under realistic depot conditions – optionally with onboard preprocessing.

02 Background
Inspection trains generate up to 10 TB/day, stored locally. Each day the train stands for several hours at a defined stabling location. This downtime shall be used to batch-transfer the data to land-based systems.

03 Goal & Minimum Requirements
Design a system that reliably transfers up to 10 TB per day wirelessly to land within max. 7 hours net transfer time under realistic conditions.

Minimum requirements:

• Up to 10 TB in ≤7h net time

• Distance ≥5 m (to track center)

• Fully automated, no manual interaction on the train

• Data format RCM-DX (input = output, unchanged)

• Robust under realistic stabling conditions (weather, limited access, regular operations)

• Line of sight only if required by the technology

Optional: onboard preprocessing (compression/prioritization), provided complete availability of all source data in the target system is ensured.

We aim to understand:

(1) suitable architectures/technologies for stationary offload

(2) realistic combinations of preprocessing + transmission

(3) long-term options (2030+)

04 Who We’re Looking For
Startups/tech companies experienced with large data volumes, high-performance transmission, and complex system architectures, e.g.:

• High-performance GB–TB data transfer

• Wireless/stationary offload (Wi-Fi, microwave radio, mmWave, hybrid)

• Edge computing/data intelligence (optional)

• Handling intermittent connections (resume/retry)

• Distributed/mobile systems

  Rail experience not required. Partial solutions are welcome if you clearly show how the minimum requirements are met.

05 Evaluation Focus
Not only peak upload speed, but reliable delivery of 10 TB/7h:

• Realistic net throughput models (overhead, restart, error handling)

• Robust architecture & operations concept

• Resilience to interruptions (resume)

• Data integrity, traceability, monitoring
  • Scalability to multiple locations

• Optional: preprocessing supporting complete offload

Future-oriented approaches welcome, not mandatory.

06 Process & Timeline
Application deadline: March 22, 2026 | Feedback by: April 1, 2026.
Submit:

• ·       company/team overview

  ·       system concept

  ·       transmission performance model (“X TB in Y h” net under defined conditions)

  ·       maturity level & expected stable net throughput; architecture sketch

  ·       data handling approach (structuring/prioritization/preprocessing)

  ·       site/infrastructure needs (install/ops/automation/remote ops/security)

  ·       risks & dependencies (frequencies/approvals/integration/safety/EMC)

  ·       rough costs (incl. scaling)
  Phase I (Top 5), online Challenge Days:
  April 27–30, 2026 (onboarding & deep dive Apr 27; concept work + Q&A Apr 28–29; presentations Apr 30 → select Top 3).

Phase II (Top 3), May 2026 (tba):
deepen concepts; mini-prototype/simulation (incl. on-site meeting); focus on logic, feasibility, robustness. Selection Day in Berlin (expected CW 20/21, May 18–21, 2026; date tba) → select 1 team.
100-day Proof of Concept: starts June 8, 2026 (real-world testing with Deutsche Bahn).

07 Funding & Benefits

• Top 5 (Phase I): €1,500 one-time.

• Top 3 (Phase II + Selection Day): +€4,500 one-time.

• Travel reimbursement for challenge events: up to €500/team (with receipts).

• 100-day PoC: €25,000 funding (no equity), DB mindbox startup manager support, access to experts & departments, optional coworking in Berlin.

08 Questions
Technical and organizational questions: dbmindbox@deutschebahn.com
Open consultation hours: Fridays 12:00–12:45
Dial-in link: https://teams.microsoft.com/meet/33918952702339?p=tSNqDCcSzPrfoisVQ0

For further information visit our Challenge Website:
https://www.dbmindbox.com/mindbox/current-calls/db-infrago-rail-data