Analysis Methods of Bioelectrical and Biomechanical Test Data
Abstract
Biological principles of sensing, information processing and actuation are fundamentally different from those applied in man-made designs which are typically based on specifications chosen to accomplish an a priori identified task. Instead of digital supercomputing found in technical systems, the nervous system of living organisms employs a hybrid signal structure that combines analogue processing with digital signal propagation generated by chemical processes inside and in between individual nerve cells. The resulting bioelectrical signals are finely tuned to the requirements of complex and yet energy-efficient operations. Spatio-temporally distributed signals obtained by hundreds, sometimes thousands of local sensors are selectively integrated to establish filters, the outputs of which are directly used to control specific biomechanical functions, for instance flight control and gaze stabilization in flying insects. The evaluation of design principles such as matched filters for visual state estimation, dynamic range fractioning of visual and inertial sensor systems to expand the tolerable input range, and the parallel extraction of different spectral features requires the availability of adequate test data. Simultaneous measurements of discrete electrical impulses at the neuronal circuit level have to be analyzed in combination with time-continuous behavioural quantities to reveal the functional relationship between sensor signals, actuators and control architectures. We have developed a fly-robotic interface (FRI) as a new analysis tool to generate biomechanical test data which are produced by the system under study itself. The FRI provides sufficient control of relevant boundary conditions essential to test and evaluate parameter-invariant sensing and information processing capabilities which biological system have adapted to operate in a variety of visual environments. The results of the project will inform novel approaches to low SWAP designs for robust GNC on autonomous aerial platforms.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 20, 2023
- Source ID
- FA86552217030
Entities
People
- Holger G. Krapp
Organizations
- Air Force Office of Scientific Research
- Imperial College London
- United States Air Force