Engineering HDR Seminar Series 18, 2013: Fuzzy Logic Based Mobile Robot Navigation on Inclined Surfaces

Mr Loh Jon Liang, postgraduate from Mechatronics Engineering

Date: 2013-04-15
Time: 12:00 to 13:00
Venue: Engineering meeting room 1, 5-4-22


The primary thrust behind behaviour-based robotics is the idea that any complex task may be decomposed into simpler tasks that can be solved by elementary behaviours. The theory of fuzzy logic provides a framework for designing these behaviours as self-contained fuzzy inference systems, the functions of which are described by their rule bases. In most implementations of fuzzy logic behaviours, the issue of controller complexity is scarcely explored. Our preliminary work has shown that piecewise nonlinear scaling of input membership functions is a promising technique for improving the performance of navigation behaviours with minimal complexity rule bases. Further work in this area concentrated around improving the optimisation process and developing a behaviour arbitration policy based on context-dependent blending (CDB) that is consistent with our proposed framework. Furthermore, this study will explore the possibility of extending this reactive framework to 3-dimensions, particularly for navigation on inclined surfaces in structured environments. This involves developing a real-time RANSAC-based 3D plane extraction method from 3D depth sensor data and integration with the existing behaviour-based reactive navigation architecture.

About the Speaker

Loh Jon Liang graduated with a Bachelor of Mechatronics Engineering (with Honours) from Monash University Clayton in 2009. He is currently pursuing a PhD in the School of Engineering in Monash University. His research focus is on the application of fuzzy logic-based reactive navigation strategies on autonomous mobile robots and real-time 3-dimensional exteroceptive sensing.