Humans rely on there sensory organs to perceive their surroundings and to decide whether a certain action is required to avoid a potential collision. Most prominent are the visual and auditory cues generated by the eyes and ears, respectively. A significant amount of research has been performed with regard to computer vision and its application to detection capabilities of robots. Vehicles operating at high speeds cannot merely rely on small lightweight digital cameras to detect incoming objects and requires the addition of heavy Light Detection and Ranging (LIDAR) systems. Weight, or reduced lifting capabilities, of UAV’s is a well known design restriction, thereby making computer vision unsuitable for hear and avoid.
The current research explores the notion if it would be possible to use auditory cues to detect nearby aircraft and subsequently to define an appropriate control signal to avoid a collision. This research may be subdivided in a number of research questions that need to be answered. Acoustic sensing, given the prevalent frequencies, ordinarily requires a large acoustic sensor array. State-of-the-art research in the past couple of years however has led to smaller acoustic vector sensors, which inherently define a localization vector to a local source. Filtering of the incoming signal will be necessary to neglect ambient acoustic sources and the noise generated by the airframe increasing the sensing resolution.
Besides acoustic sensing a study is to be performed in advanced control theory, generating the appropriate control commands for agile obstacle avoidance. Robust nonlinear switching controllers will be implemented, exploiting the total control authority of a particular UAV, e.g. fixed-wing aircraft or multicopters.
The performance of the robust controllers discussed above is related to the error of the estimate of the local state of the aircraft. The fast maneuvering regime, in the presence of a broadband vibration spectrum, requires besides an appropriate design the use of nonlinear state filtering techniques.
- Partner and subsidiaries: Microflown Avisa
- Research topics: Advanced control theory, estimation theory, acoustics