Download Project (360 KB)
Human brains can do amazing things. The motor cortex can induce huge transient responses to perform very gentle and precise movements with the regulation of neuromodulators. The sensory and motor cortices in a human body are shaped by experience. These regulatory mechanisms of the brain enable humans with flexible and robust abilities in adapting to dynamic environments and greatly improve accuracy and fault tolerance which is the bottleneck in the control of complex real-time systems. The ability to identify and replicate these biological control systems could help provide a better understanding to reproduce functional behaviours of humans (like walking running etc.) to yield better results i.e., replace the bits and clocks of digital computation with the spikes and rhythm of human communication.Inspired by the control mechanisms of motor cortex, the study presents evidence on a small scale by focusing on developing a software infrastructure that allows for data collection from a human teacher performing control of a class of non-linear systems (Inverted Pendulum on Cart and Ball and Beam) with uncertain dynamics and external perturbations and ability to learn from the collected data using the Concurrent Learning algorithm to identify the control law of unknown form acquired by a human through direct experience with the system. Owing to high demands of real-time performance, the discrete-time dynamics of the systems are considered . Specifically, numerical results focusing on whether the human subject was able to stabilize the system for a sufficiently longer time or not addressing the efficacy of the data-set is presented . To validate the approach, the identified neuromorphic controller is used to stabilize the non-linear systems in hand.
Primary Advisor's Department
Electrical and Computer Engineering
Stander Symposium, School of Engineering
Institutional Learning Goals
Diversity; Faith; Practical Wisdom
"Identification and Design of Neuromorphic Controller Inspired by Mammalian Neural Control Mechanisms Applying Concurrent Learning Algorithm" (2023). Stander Symposium Projects. 3216.