Showing papers by "Universiti Teknologi Malaysia published in 1970"

Knowledge-Based Trajectory Error Pattern Method Applied to an Active Force Control Scheme

Abstract: The active force control (AFC) method is known as a robust control scheme that dramatically enhances the performance of a robot arm particularly in compensating the disturbance effects. The main task of the AFC method is to estimate the inertia matrix in the feedback loop to provide the correct (motor) torque required to cancel out these disturbances. Several intelligent control schemes have already been introduced to enhance the estimation methods of acquiring the inertia matrix such as those using neural network, iterative learning and fuzzy logic. In this paper, we propose an alternative scheme called Knowledge-Based Trajectory Error Pattern Method (KBTEPM) to suppress the trajectory track error of the AFC scheme. The knowledge is developed from the trajectory track error characteristic based on the previous experimental results of the crude approximation method. It produces a unique, new and desirable error pattern when a trajectory command is forced. An experimental study was performed using simulation work on the AFC scheme with KBTEPM applied to a two-planar manipulator in which a set of rule-based algorithm is derived. A number of previous AFC schemes are also reviewed as benchmark. The simulation results show that the AFC-KBTEPM scheme successfully reduces the trajectory track error significantly even in the presence of the introduced disturbances. Key Words : Active force control, estimated inertia matrix, robot arm, trajectory error pattern, knowledge-based.

Separation of aqueous isopropanol through chitosan/poly(vinyl alcohol) blended membranes by pervaporation

Abstract: Blend membranes consisting of chitosan/poly (vinyl alcohol) (CS/PVA) were prepared from a solution casting method and characterized via the pervaporation separation of isopropanol-water mixtures. The solution of pure chitosan was blended with poly(vinyl alcohol) at different compositions. The miscibility of the blended polymers was determined. Criteria for miscibility was based on the clarity of the blend and the transparency of the membrane obtained. The effects of feed concentration and chitosan content in membrane on the permeation flux and separation factor were also investigated. For the dehydration of 90 wt.% of isopropanol-water mixtures the performance of the blend ratio of 30:70 (CS:PVA) exhibited the highest separation factor. It was shown that under the same operating conditions, the pervaporation separation index (PSI) for homogeneous membranes were higher than that of blended membranes for the entire feed concentrations. Key Words : Blend, Chitosan, Poly (vinyl alcohol), Miscibility, Pervaporation.

Routing Algorithm of Mobile Agents for Query Retrieval Using Genetic Algorithm

Abstract: Mobile agents often have a task to collect data from several predefined sites. This should be done in an efficient way by minimising the elapsed time. Usually these agents only know the list of sites but not the distances between them. This paper proposes a method to minimise a network routing time taken by the mobile agents to collect information from different sites using genetic algorithm (GA). The mobile agents repeat travelling over short routes and avoid longer ones. Mobile agents for query retrieval have used the GA to select the best routes that minimise the query retrieval time. The result shows that the proposed method provides good time minimisation in retrieving the query results by the mobile agents based on different GA parameters.