My research seeks to investigate and explore the use of neuromorphic principles and spike-based computation as an adaptive, scalable and low-power control system for a high-speed application. Although control theory is a well-established field, there are many instances where biology trumps sophisticated electronic systems, especially in the contexts of power efficiency and adaptability. In order to demonstrate and explore the potential of a bio-inspired control system, a table tennis robot was chosen to as a real-world problem as it is a well-researched area that continues to pose many challenges to conventional engineering principles.
As a result, the outcome of this research aims to build a biologically-inspired robot capable of playing table tennis against a human opponent. The research will make use of vision and learning systems which have not yet been applied to robotic table tennis. As the vision system has typically been the bottleneck and most processing intensive aspect of robotic table tennis, this research will utilise the ATIS cameras. These cameras are capable of drastically reducing the amount of data required for processing, and will enable rapid object tracking.
My research will also investigate spike-based sensor fusion techniques as a neuromorphic approach to processing data in such a system. This will be implemented in the form of a specialised hit detector, a stereo asynchronous vision system, and a spike-based learning loop. The ultimate goal of this research will be to build a working, robotic table tennis player that can play against a human opponent under standard table tennis rules.
- MSc (Eng) – University of Cape TownBCom (Hons) in Financial Analysis and Portfolio Management – University of Cape Town
- BSc (Eng) Electrical and Computer Engineering – University of Cape Town
For a full listing of my publications, please see my personal publications page.(opens in a new window)