Sang-Keon Oh

A system for gait rehabilitation: mobile manipulator approach

A power assisted gait rehabilitation robot system by which a patient who has difficulty in walking can exercise is proposed. Our developed system consists of a robotic manipulator which can adjust the height of the system according to the user and a mobile base with two-driving wheels. It is similar to the architecture of a mobile manipulator. The neural network based control algorithm for a mobile manipulator used to compensate for dynamic interactions, unmodeled dynamics, and disturbances by the user on the system, is derived and experimental results are shown.

Heterogeneous local model networks for time series prediction

The approaches of local modeling have emerged as one of the promising methods of time series prediction. By use of the divide-and-conquer method, local models can exploit state-dependent features to approximate a subset of training data accurately. However, the generalization performance of local model networks is subject to the proper selection of model parameters. In this paper, we present a new method for local model construction for the noisy time series prediction. The proposed method uses the principal component analysis (PCA) and cross-validation technique to construct an optimal input vector for each local model. A heuristic learning rule is also proposed to update the mixture of experts network structure, which determines the confidence level of local prediction model. The proposed method has been tested with noisy Mackey-Glass time series and Sunspot series.

A new distributed evolutionary algorithm for optimization in nonstationary environments

Previous studies of evolutionary algorithms (EA) focus on stationary environments where the evaluation function and constraints of the problem are fixed over time of evolution. This study introduces a new algorithm for optimization in nonstationary environments. In this study, we explore the use of a hierarchically distributed evolution model which enables us to utilize fast local search operators while maintaining the sufficient diversity to adapt the environmental changes. The results are compared with previous techniques with several test functions.

Time series prediction by mixture of linear local models

Local modeling approaches have emerged as one of the promising methods of time series prediction. By divide-and-conquer method, state-dependent local model can approximate a subset of training data accurately. However, the construction of local models need appropriate selection of much larger number of parameters. This paper presents a method to construct a mixture of linear prediction models for the prediction of nonlinear time series. The use of locally linear model reduces the burden on the user to specify parameters using linear optimization method. This method is applied to the modelling of the Mackey-Glass time series.

Robust model reference adaptive control of underactuated robot manipulators

A robust control scheme, overcoming the uncertainty in an underactuated robot manipulator, is proposed based on the sliding mode and MRAC (model reference adaptive control) schemes. By introducing the model reference adaptive technique and robust control algorithm, the dynamic response of each joint of underactuated manipulators can be pre-determined without exact knowledge of the system parameters. To show the effectiveness of the proposed algorithm, simulations for a 2-link underactuated robot with 1 fault joint are done.

Adapting the migration topology of macro-micro evolutionary algorithm by clustering the individuals using self-organizing map

In this paper, we propose a self-adaptive migration rule for macro-micro evolutionary algorithm which was proposed to find several local optima for multi-model optimization problems. The algorithm consists of two evolutionary algorithms which control global species and local individuals respectively. To keep the diversity explicitly, we incorporate a clustering method to divide individuals to several species. Clustering method based on self-organizing map (SOM) can divide individuals to several species and determine the neighboring topology information which defines the migration topology between species. To examine the computational effectiveness of proposed algorithm, we apply the algorithm to standard benchmark problems for numerical optimization.

Balancing the selection pressures and migration schemes in parallel genetic algorithms for planning multiple paths

Parallel genetic algorithms are particularly easy to implement and promise substantial gains in performance. Its basic idea is to keep several sub-populations that are processed by genetic algorithms. Furthermore, a migration mechanism produces a chromosome exchange between sub-population. In this paper, a new selection method based on nonlinear fitness assignment is presented. The use of the proposed ranking selection permits higher local exploitation search, where the diversity of population is maintained by a parallel sub-population structure. Experimental results show the relation between the local-global search balance and probabilities of reaching the desired solutions using test functions and nonstationary route-planning problems.