Open-Sourcing a Humanoid Robot: Ensuring Integrity and Privacy of Robot Sensor Data

This paper outlines the process of open sourcing key components of a commercial humanoid service robot. Humanoid service robots collect data from their environment, and the integrity of these data streams cannot be ensured for proprietary solutions. The integrity of this data needs to be ensured when deploying robots in sensitive use cases.

We have identified which data streams are most delicate and investigated the implementation of the sensor or feature on the robot using basic data- and network analysis tools. A comparison of sensor data was conducted to prioritise the order of modifications. Some features can be modified to achieve data stream integrity, but other sensors and systems need to be replaced. Other open-source robotics projects, commercial and non-commercial, serve as subjects for comparison.

There are few open-source humanoid service robots. The integrity of humanoid robot sensor data can be ensured and verified. All sensor data is not equally sensitive by nature; some sensors were not considered a priority and needed to be modified. Extensive technical knowledge is required for successful modification as well as validation of the data.

Commercial humanoid service robots can be modified to ensure sensor data integrity with the purpose of deployment in sensitive use cases. The applicability of the implementation is yet to be confirmed.