Published September 10, 2026 | Version 1.0.1

MoRoOp: A Dataset of Autonomous Mobile Robot Operations in Production Logistics

  • 1. ROR icon KTH Royal Institute of Technology
  • 1. ROR icon KTH Royal Institute of Technology
  • 2. ROR icon Bremer Institut für Produktion und Logistik GmbH

Description

MoRoOp contains 72 hours of operational data from an autonomous mobile robot (AMR). The data was recorded at the KTH IPU Lab as part of the Self-Organizing Production Logistics (SOPL) project. The current release (1.0.0) covers five consecutive days, 2026-07-27 through 2026-07-31, organized into morning and evening shifts.

The dataset is additionally available on Hugginface, see MoRoOp Main

Data organization

The `data/` directory contains one folder for each of the ten shifts. Every shift folder follows this naming pattern:
data/YYYY_MM_DD_{morning|evening}/
YYYY_MM_DD_{morning|evening}_jobs.parquet
YYYY_MM_DD_{morning|evening}_operations.parquet
YYYY_MM_DD_{morning|evening}_dispatch_events.parquet
YYYY_MM_DD_{morning|evening}_robot_state_raw.parquet
YYYY_MM_DD_{morning|evening}_robot_state_cleaned.parquet

`data/kits.parquet` is a shared kit catalogue with the columns `kit_id`, `color`, `size`, and `quantity`.

Tables

  • jobs: Job-level records, including payload, operation definitions, lifecycle state and outcome, termination reason, and release, start, due, finish, and creation timestamps.
  • operations: Operation-level records linked to jobs by `job_id`, including state, outcome, payload, priority, dependencies, retries, termination reason, and timestamps.
  • dispatch_events: OperationDispatcher event logs, tracing the management of operations on the mobile robot. Includes event type, operation ID, job ID, and timestamps.
  • robot_state_raw: Reindexed, postprocessed but unfiltered robot-state and diagnostic telemetry before postprocessing.
  • robot_state_cleaned: Filtered and cleaned robot-state and diagnostic telemetry. Its IDs are retained from the raw data, so gaps identify removed records.

Code examples

Example code for loading and analyzing MoRoOp, together with scripts for reproducing the publication figures, is available in the MoRoOp dataset toolkit.

Technical info (English)

Version 1.0.1. fixed an export bug in which the dependencies column was misspelled, leading to all values in this column to be null.

Files

MoRoOp_V_1_0_1.zip

Files (14.9 MB)

Name Size Download all
md5:818f0b55129e181f3c0fa610b2059c96
14.9 MB Preview Download

Additional details

Funding

The Government's Strategic Research Area initiative (SFO)
Excellence in production research (XPreS) 2009-01024