Company Description
Movu Robotics
Innovation, ambitious growth, technology, state-of-the-art robots – that’s Movu Robotics in a nutshell. Due to the in-house design and manufacturing process of our systems, an exciting, hands-on work environment with plenty of challenges and advancement opportunities is guaranteed. Will you help shape the future of intralogistics?
Job Description
Problem statement
In modern automated logistics and warehouse environments, plant control software must operate reliably across a wide variety of pallet types, each with distinct physical and dimensional properties. Currently, no structured digital representation of pallet data exists that can be systematically fed into the plant simulation model, leaving regression testing largely dependent on physical, on-site trials. As a consequence, when new software versions are released, their compatibility with the full range of pallet types—including special or atypical variants—cannot be verified in a virtual environment prior to deployment. This gap forces engineering teams to conduct repeated physical pallet tests, a practice that is both time-intensive and resource-costly. Furthermore, the absence of a centralized, validated pallet database limits the plant model's fidelity and prevents meaningful simulation of edge cases. Closing this gap is essential not only to accelerate the software release cycle, but also to reduce operational risk, decrease physical testing overhead, and progressively build a comprehensive, reusable repository of pallet profiles that enriches future simulation capabilities.
Main issues
- Coverage & Pallet Diversity
The plant software must handle a heterogeneous population of pallet types, including non-standard and special-purpose variants. Ensuring that the simulation model accounts for the full diversity of real-world pallet geometries, weights, and surface properties—without direct physical access to each type during testing—requires a methodical data collection strategy and a robust parametric model that can generalize across variants.
- Data Acquisition & Format Standardization
Pallet test data currently exists in heterogeneous formats across different test setups and measurement instruments. Transforming this raw, inconsistently structured data into a normalized, machine-readable format that can serve as direct input to the plant model presents a significant data engineering challenge. A well-defined schema and automated transformation pipeline are required to ensure consistency, traceability, and scalability.
-Reduction of Redundant Physical Testing
Transitioning from a purely physical to a simulation-first testing approach requires a degree of institutional trust in the virtual model's outputs. Demonstrating that a properly parameterized simulation can replace repetitive physical re-tests—particularly for previously characterized pallet types—demands clear evidence of equivalence and a change management process that aligns engineering, quality, and operations stakeholders.
- Database Design & Long-Term Maintainability
The resulting pallet database must be designed not as a static artifact but as a living, evolving repository that grows as new pallet types are encountered. This implies thoughtful database architecture choices, versioning strategies, and access protocols to ensure that data remains accurate, auditable, and usable by both current and future simulation workloads without significant rework
Solution idea
The proposed work centers on the design and implementation of a structured data framework that captures, normalizes, and integrates physical pallet test data directly into the existing plant simulation model. In a first phase, a data acquisition and transformation pipeline will be developed to extract measurements from pallet tests—covering dimensional, mechanical, and identification attributes—and re-format them according to a defined schema compatible with the plant model's input interface. This pipeline will incorporate validation rules to flag incomplete or anomalous entries before ingestion. In a second phase, the plant model will be extended with a pallet simulation module, parameterized by the ingested data, enabling the controlled execution of software regression tests against a broad and growing set of digital pallet profiles. Special pallet types, which previously required dedicated physical sessions, will be characterized once and stored, eliminating the need for repeated re-testing. Finally, a structured pallet database will be established to serve as the authoritative repository for all characterized pallet types, supporting traceability, version control, and retrieval for future simulation campaigns. The primary tools will include scripted ETL (Extract–Transform–Load) routines in Python for data processing, integration interfaces with the existing plant simulation environment, and a relational or document-oriented database for persistent storage. Validation will be conducted by comparing simulation outcomes with historical physical test results across a representative pallet sample, establishing a quantifiable confidence level for the framework's use in production software release cycles.
Qualifications
Skills
Programming +++
Product development +
Data management ++
Additional Information
What we offer you?
1. A place in an enthusiastic, young organisation with the necessary ambitions
2. The necessary freedom and the opportunity to take initiatives
3. Location Office : Lokeren
4. Satellite office in Berchem ( Antwerp )
Unsolicited representations by third parties (recruitment agencies, headhunters, ...) of CVs via mail and/or telephone for our vacancies are considered as direct applications where no compensation is provided to the third party. Any T&Cs from these third parties will not be accepted unless upon signature of the T&Cs by a person in charge of HR. Candidates remain registered in the system for 12 months and cannot be proposed again during this period.