Research Activities

1.    Biological Behavior Based Intelligent Plume Tracing System:

Various biological entities are capable of tracing an odor plume in a fluid medium to its source. Autonomous vehicles with such plume tracing capabilities could be beneficial in applications such as mine and unexploded ordinance detection and location. In this study, the insect behavior based plume tracing strategies are investigated, and the performance is improved by soft computing. The results should form a good initial condition for the development of autonomous vehicle tracing strategies.

2.    Intelligent Control and System Modelling:

A behavior-modeling-based approach to design intelligent control system is proposed based on fuzzy-neuro networks. In this study we also propose a hybrid FUZZY P+ID controller and proves its stability. This control system is successfully applied into some industrial system with nonlinear dynamics under uncertainty, such as stoke-fired boilers and mechanical manipulators.

3.    Behavior Control of Mobile Robot Navigation:

A behavior based control strategy for mobile robot navigation in unknown environments is proposed. In order to deal with uncertainty in environments, a complex navigation task is decomposed into each type of behavior with a simple feature. Fuzzy rules are used to quantitatively formulate the defined behaviors, and a neural network is used to fuse information from multiple sensors and to understand environments. Furthermore, a strategy for integration of low-level behavior control and high-level global planning is proposed.

4.    Real-Time Manipulator Motion Planning:

A real-time manipulator motion planning is proposed by fast mapping obstacles from a workspace into a configuration spacet. The proposed approach can be applied in real-time path planning for multiple robot transfer movements in a dynamic environment.

5.    Fuzzy Image Processing and Vision-Guided Vehicle Navigation:

A vision system and other kinds of sensors are integrated on an autonomous vehicle. A vision-guided vehicle navigation approach is proposed by fuzzy inference.  The experiments on an AGV system show that the method is not sensitive to environment changes. In addition, an approach to extraction of line and step edges by fuzzy reasoning is proposed.