The ongoing digitalization of production systems leads to higher temporal and spatial flexibility in production. It furthermore causes an intensified interaction between intelligent technical systems and human beings. Especially the advancing interaction between robots and humans requires sensor systems and measurement methods for automatic and reliable person detection. To guarantee a safe interaction between humans and machines, a reliable detection of the human body has to be assured.
The diversity of sensors and methods requires a well-structured listing concerning fields of application and limits. In the project "Reliable Person Detection in Human-Machine Interaction" an application-matrix was developed, which assigns methods for human detection to fields of application and shows their limits.
To develop the matrix, research in literature and on the market was conducted. From that research, 13 measurement methods and five fields of application have been derived.
In the report and the application-matrix the measurement methods are divided in four groups due to the area they are monitoring: monitoring the whole workplace, monitoring a two dimensional plane, machine-centered monitoring and active personcentered monitoring. For each of the four groups appropriate sensors are mentioned. The principles of all sensors are outlined and methods for person detection presented. Additionally, multi-sensor systems are mentioned in the report, which combine the advantages of different sensor types by data fusion.
The applications are categorized into five fields: industrial robots, automata and handling devices, industrial trucks, vehicles and other transport systems (on rails and in the air). These fields of application are then assigned to the methods of person detection.
Finally, the measurement methods are evaluated due to possible purposes of application (presence detection, positioning, motion coordination between machine and human, reconstruction of body postures), technical parameters and environmental factors. The evaluation is discussed in the report and the results are represented in the application-matrix.
Please download the complete report "Safe person detection for human-machine-interaction" (in German only).
M. Schmauder, K. Höhn, P. Jung, K. Lehmann, S. Paritschkow, P. Westfeld, H. Sardemann:
Sichere Personenerkennung in der Mensch-Maschine-Interaktion.
1. edition. Dortmund: Bundesanstalt für Arbeitsschutz und Arbeitsmedizin 2016. pages 76, Project number: F 2322, paper, PDF file, DOI: 10.21934/baua:bericht20161102
© Federal Institute for Occupational Safety and Health