Search Unity

Augment assembly work planning with digital human simulations

, 1월 19, 2021

Learn how Unity is supporting the MOSIM research project to simulate and analyze complex, realistic human motions for a range of manufacturing use cases. With the ability to simulate assembly worker tasks in minutes rather than weeks, companies can improve production planning, increase worker productivity and safety, and reduce risk and costs.

The MOSIM project is a consortium championed by auto manufacturer Daimler in partnership with Unity and more than 20 other partners. It seeks to provide an open standard for digital human simulations for industrial applications.

Watch the video below for a brief overview of MOSIM and its vision for simulations powered by Unity, and download the open-source project on GitHub.


Download MOSIM on GitHub


The opportunity to improve assembly work processes

Image credit: Daimler

Even in today’s age of automation, assembly of vehicles and other manufactured products still requires significant manual work. Much about the production planning process can be improved, as workers’ tasks are typically not visualized in 3D but described in text. Additionally, validation for these tasks occurs on hardware prototypes. Simulation can make this experience more efficient, but generating a process simulation has historically been time-consuming and required tools for experts.

While it’s possible to simulate specific, individual actions (e.g., affixing a part), simulating a joint sequence of actions – for instance, walking to a part, picking it up, walking to the vehicle, and affixing the part – requires extensive manual effort and physical motion capture of the actual location and setup. Aside from very specific, high-risk scenarios that might justify the complexity and time investment, most companies are deterred from making this effort. 

“Manual assembly for a car or bus requires a lot of different actions, and currently there is no existing tool that can simulate those scenarios comprehensively. The challenge is bringing these individual sequences together, so we wanted to create a framework for digital human simulations that could help us do that,” says Felix Gaisbauer, PhD Researcher at Daimler Buses – EvoBus GmbH and Technical Coordinator for MOSIM.

Image credit: Daimler

As the production of cars, trucks, buses, and more become increasingly more complex and competitive, the need to maximize efficiency is paramount. Reliance on physical testing and optimization not only hinders productivity and inflates costs, but often leaves manufacturers lacking confidence in their operational efficiency.

“While these efforts have improved over the last decade, dynamic simulation of humans is more or less not being done. Right now is really a prime moment for simulation in production,” says Thomas Bär, Manager at Daimler Buses – EvoBus GmbH and Project Leader for MOSIM.


A new open standard for digital human simulations

With MOSIM’s modular, Unity-based human simulations, manufacturers can easily piece together a series of character animations into a comprehensive simulation. They can:

  • Visualize and verify the assembly steps in real-time 3D
  • Simulate hypothetical scenarios to determine the feasibility and optimal sequence of tasks
  • Conduct metric-based assessments (e.g., ergonomics, buildability)
  • Optimize and improve processes before completing physical setups or training teams, which greatly reduces costs and inefficiencies 

Simulations can be created in mere minutes, as compared to the roughly two weeks it would take to create a comparable simulation manually. From identifying ergonomic opportunities during the assembly process to improving worker productivity, safety and training, MOSIM has enormous potential to impact numerous stages of production.

“This opens up a lot of new possibilities. When you can simulate humans moving realistically through assembly sequences, it provides additional opportunities to investigate optimization potential in the factory, such as adjusting shelf positions in order to shorten walk paths. It can even be used to provide immersive training for workers on how to complete their tasks,” says Gaisbauer.

Powering visualization and simulation with Unity

Leveraging Unity’s real-time 3D platform at the core of MOSIM’s framework opens up new possibilities for visualization and simulation, enabling digital avatars and animations that elevate the entire experience. Unity allows for fast prototyping as well as easy deployment to devices like virtual reality headsets.

“I really like the way Unity’s programming works, the large Asset Store, and that most of our project partners have more experience with Unity than other engines. The comprehensive package Unity provides is what makes the difference for us,” says Gaisbauer.  

MOSIM can be accessed via a web-based application as well as in Unity. It is capable of automatically generating worker assembly simulations in 3D using standardized text descriptions, such as “Pick Up & Put Down.”

Get started

MOSIM’s Motion Model Unit (MMU) Generator lets you create new series of human motions

Inspired by the Functional Mock-up Interface (FMI) approach, the industry open-source standard for computer simulations, MOSIM offers Motion Model Interfaces (MMIs) and their implementations called Motion Model Units (MMUs), which range from simple animations like walking to complex tasks like climbing a ladder.

MOSIM includes a library of predefined human motions in its open-source repository on GitHub and encourages the Unity community to contribute to the creation of more MMUs. A Unity-based “MMU Generator” enables the creation of custom MMUs from FBX and other file formats without requiring programming knowledge.

“This is something special. We don’t have this for any other gaming engine – it is unique to Unity,” says Gaisbauer. “It simplifies development and we hope it will attract more animation artists and developers to provide more MMUs.”

Check out the MOSIM repository on GitHub and bring the power of digital human simulation to your application.