Session: 02-08-02: Innovative Product and Process Design II

Paper Number: 68910

Start Time: Thursday, 04:40 PM

68910 - Spray-on Capacitive Proximity Sensors in 3d Printed Robotic Links 

Proximity sensors are crucial components for any robotic system when collaborating with humans. The sensor’s ability to detect the presence of human body parts prior to contact or collision allows for collision avoidance and contactless manipulation of the robot. Capacitive sensors are commonly used as proximity sensors as they detect the human body well and can easily cover large surface areas.

The purpose of this research is to integrate capacitive proximity sensors into additively manufactured robots. The actuators in such robots are off-the-shelf parts, while the links between them are individually designed and manufactured. These design freedoms allow us to construct purpose-built robots with individualized geometry and kinematics.

We integrated for the first time proximity sensors into additively manufactured robotic links by subdividing the inner cavity of the links radially, thereby forming separate chambers extending along the length of the link. These robotic links with internal chambers were fabricated by selective laser sintering (SLS) from polyamide (PA). By coating the walls of these chambers with conductive graphite spray, we formed electrodes suitable for capacitive proximity sensing. The number of chambers corresponds to the number of independent sensors. The sensors can therefore detect not only the presence but also the direction of the approach of a foreign object. An off-the-shelf evaluation board inserted in every link measures the capacitance of each electrode and sends the data to the robot controller. This integrated proximity sensor design is perfectly suited to be included into the individualized manufacturing of highly specialized robot geometries and rapid prototyping. Because the required sensor structures are integrated into the geometry of the links, the only additional manufacturing steps to equip a robot with such sensors are the application of the conductive spray and the insertion of the electronics. To simplify and accelerate the sensor design for individualization and rapid prototyping, we developed an automated design algorithm. The algorithm extrudes a pre-defined sensor structure along arbitrary robotic arms and generates a fitted cavity to house the electronics. These spray-on proximity sensors can easily be integrated into different and complex robot link geometries. This includes multidimensionally curved surfaces where we can obtain a maximum electrode area for most given geometries. Furthermore, the internal placement of the sensor electrodes and electronics protects them from environmental influences.

Experimentally, we confirmed that the sensor performance is not influenced by the thickness of the surrounding PA, as this material has very little influence on the electric field due to its low dielectric constant. As expected, a clear relationship was found between the sensor area and its range. The detection range for several approaching objects of different materials was compared. Grounded conductive objects like a human hand achieved the greatest detection range, confirming that capacitive proximity sensors are especially suited in the context of collaborative robots. The proximity detection range for human body parts was measured to be 95 mm. Through simulation, we were able to optimize the chamber shape for minimal parasitic capacitance, thereby reducing the cycle time of the system to 87 ms for a setup with four sensors. The internal spray-on electrodes performed similarly to external electrodes made from sheet metal. The concept was tested on an additively manufactured 4R serial robot by implementing proximity servoing and collision avoidance modes.

The presented method provides a fast and automatic integration of capacitive proximity sensors into individualized robotic links using conductive spray paint. The method is especially well suited for complex, additively manufactured geometries and can be easily and quickly integrated in the overall production process.

Presenting Author: Samuel Detzel Technical University of Munich

Authors:

Samuel Detzel Technical University of Munich
Yannick S. Krieger Technical University of Munich
Robert W. Hoefer Technical University of Munich
Anton Robe Technical University of Munich
Annette C. Sigling Technical University of Munich
Tim C. Lueth Technical University of Munich