Available for


The flexible behavior engine.


Please use the following reference if you use FlexBE in your research project and want to cite it. Feel free to contact me if you need help with a particular setup, implementing custom states, or have feature requests.

    Title = {{Human-Robot Collaborative High-Level Control with an Application to Rescue Robotics}},
    Author = {Philipp Schillinger and Stefan Kohlbrecher and Oskar von Stryk},
    Booktitle = {IEEE International Conference on Robotics and Automation},
    Address = {Stockholm, Sweden},
    Month = {May},
    Year = {2016}


This page will give you a brief overview of the scientific background of FlexBE. While the core is based on the state machine framework SMACH [1], the concepts have been significantly extended and adapted to support a built-in mixed-initiative robot-operator interaction. This means, when executing a behavior, both the robot and the operator can initiate requests to ask the other party for help, e.g., by approving a decision or taking a specific transition. This can happen at any time during execution and does not need to be explicitly modeled in any state machine.

Main parts of the internal operator interaction concepts, especially the Autonomy Level for adapting to uncertain situations and the usage of a remote behavior mirror for bandwidth efficiency, have been developed as my Bachelor Thesis [2]. FlexBE as you know it now was the result of my Master Thesis [3]. Main goal was to provide a way for a non-developer operator to make adjustments to behaviors during runtime, primarily desgned for situations when the exact scenario the robot will be facing is not known in advance. Along with this requirement, the user interface has been developed in order to support behavior creation. Motivation behind providing such an extensive user interface was, in addition to the fact that the operator will most likely not be an developer, reducing cognitive load on the operator while the robot is in the field and reduce sources for possible errors by running verification checks and automatically generating syntax-error-free code. Please refer to chapter 3 of my Master Thesis for further details.

[1] J Bohren and Steve Cousins. The SMACH high-level executive [ros news]. Robotics & Automation Magazine, IEEE, 17(4):18–20, 2010.
[2] Philipp Schillinger, Development of an Operator Centric Behavior Control Approach for a Humanoid Robot. BSc thesis, Technische Universität Darmstadt, 2013.
[3] Philipp Schillinger, An Approach for Runtime-Modifiable Behavior Control of Humanoid Rescue Robots. MSc thesis, Technische Universität Darmstadt, 2015.