Roboter-Ego-Noise, auch „Eigengeräusche“ genannt, sind Geräusche, die ein Roboter durch Bewegung (z. B. durch Reibung von Motoren und Aktuatoren) oder im Ruhezustand (z. B. durch Lüfter) erzeugt. Bei Robotern mit Mikrofonen kann dieses Ego-Noise die Audiosignalverarbeitung stören. Statt es wie üblich als Störfaktor zu betrachten, wird in dieser Arbeit Ego-Noise als nützliche Informationsquelle untersucht, die Einblicke in den Zustand des Roboters und seiner Umgebung bietet. Es könnte sogar dabei helfen, ähnliche Roboter zu erkennen oder deren Bewegungsmuster zu analysieren. Die Arbeit verfolgt drei Ziele: (1) geeignete Darstellungen von Ego-Noise-Merkmalen zu entwickeln, (2) sensomotorische Abbildungen mit Ego-Noise zu integrieren und (3) den Nutzen von Ego-Noise für mobile Roboter zu bewerten. Dazu werden theoretische Modelle (Vorwärts- und Inversmodelle) verwendet, die durch „motorisches Brabbeln“ – ein kindliches Explorationsverhalten – inspiriert sind. Experimente zeigen, wie Ego-Noise zur Geländeerkennung (z. B. Hangneigungen), Echolokalisierung und Bewegungssimulation eingesetzt werden kann. Zudem wird untersucht, wie Roboter synthetische Audiodaten erzeugen und über Ego-Noise miteinander kommunizieren können. Die Ergebnisse zeigen, dass Ego-Noise die Umweltwahrnehmung und Interaktion von Robotern verbessern und neue Kommunikationsstrategien ermöglichen kann. / Robot ego-noise refers to the sounds robots generate during movement, primarily from motor and actuator friction, or even when stationary, due to components like cooling fans. This noise often interferes with audio signal processing, reducing algorithmic performance. However, this research treats ego-noise as a valuable sensory signal, offering insights into the robot’s state and surroundings. It may even assist in identifying nearby robots and analyzing their motion patterns. This thesis has three main objectives: (1) to develop computationally suitable representations of ego-noise features, (2) to create sensorimotor mappings that utilize these features, and (3) to explore ego-noise as a source of useful information for mobile robots. Using a framework of internal models, including forward (predictor) and inverse (controller) models, the study employs a learning method inspired by infant "motor babbling." It examines how ego-noise can provide information for various tasks, such as detecting terrain changes like slopes, echolocation for wall proximity, and movement imitation based on ego-noise. Additionally, the integration of forward and inverse models enables robots to synthesize audio from motion and share it with other robots. This research highlights how ego-noise can enhance robots’ environmental interaction and, when combined with other sensors, improve perception in complex scenarios.

You do not want your drone to crash. Therefore, safety systems should be put in place to prevent such an event, and obstacle avoidance is a major part of this. Today, the most successful techniques use cameras or light detection and ranging (LIDAR) to find and avoid obstacles; but to improve resiliency, multiple systems should be used. This thesis proposes to use microphones, listening to the drone’s own noise, to estimate the distance to surrounding obstacles. An obstacle echolocation solution for multi-rotor aerial vehicles (MAVs) using ego-noise is developed. The MAV’s noise is captured and auto-correlated to detect echoes at different time delays. This signal is whitened to remove structured measurement noise resulting from the narrow-band components of the MAV’s noise. By recording the MAV’s noise using multiple microphones, a time of arrival (TOA) estimate of the obstacle position is achieved. A beamforming-based solution is used to calculate this estimate. A series of simplified proof-of-concept experiments show that ego-noise echolocation is possible and that the developed solution works in a controlled environment. A prototype implementation of a realistic system is also created. Four signal fusion alternatives are compared, though no best alternative is found for all situations. More work is needed to apply the findings of this work in a robust way, but the principle is shown to work.