Wei Hong Chin

Incremental on-line learning of human motion using Gaussian adaptive resonance hidden Markov model

In this paper we present an approach for on-line and incremental learning of human motion patterns through continuous observation of motion using novel Topological Gaussian Adaptive Resonance Hidden Markov Model (TGART-HMM). The observed human motion patterns are encoded in a novel modified version of Hidden Markov Model (HMM) called TGART-HMM. The on-line learning process consists of updating the structure of Hidden Markov Model using a topology-learning mechanism based on Gaussian Adaptive Resonance Theory (GART). The model size is adaptable based on the observed motion patterns. The resulting HMM structure is a graph where each node represents an encoded motion pattern. The parameters of TGART-HMM are updated incrementally to incorporate incessant motion patterns. The algorithm is tested on motion captured data to test the efficacy of the system.

Multi-channel Bayesian Adaptive Resonance Associate Memory for on-line topological map building

Graphical abstractDisplay Omitted HighlightsMBARAM enables topological map building with little or no human intervention.Does not require high-level cognitive and prior knowledge to function in a natural environment.Multiple sensory sources are processes simultaneously crucial for real-world robot navigation.Results show capability of generating topological map online used for localization. In this paper, a new network is proposed for automated recognition and classification of the environment information into regions, or nodes. Information is utilized in learning the topological map of an environment. The architecture is based upon a multi-channel Adaptive Resonance Associative Memory (ARAM) that comprises of two layers, input and memory. The input layer is formed using the Multiple Bayesian Adaptive Resonance Theory, which collects sensory data and incrementally clusters the obtained information into a set of nodes. In the memory layer, the clustered information is used as a topological map, where nodes are connected with edges. Nodes in the topological map represent regions of the environment and stores the robot location, while edges connect nodes and stores the robot orientation or direction. The proposed method, a Multi-channel Bayesian Adaptive Resonance Associative Memory (MBARAM) is validated using a number of benchmark datasets. Experimental results indicate that MBARAM is capable of generating topological map online and the map can be used for localization.

Topological Gaussian ARAM for Simultaneous Localization and Mapping (SLAM)

This paper proposes a new neural architecture called Topological Gaussian ARAM (TGARAM) for Simultaneous Localization and Mapping (SLAM). TGARAM is integrating the Gaussian classifier with the incremental topology-learning mechanisms of the Growing Neural Gas (GNG) model for online learning of multidimensional inputs and topological map building. By using the Gaussian classifier, the sensitivity to noise on a number of benchmarks data sets is diminished, and it learns a more efficient internal representation of a mapping. The incremental topology-learning mechanisms of GNG enable TGARAM to connect the generated nodes and build a topology-preserving map. In addition, TGARAM retains multi-channel ARAM network architecture and thus capable to learn multiple mappings simultaneously across multi-modal input patterns, in an online and incremental manner. Multiple sensory sources can be transmitted to TGARAM to build a topological map and improve the estimation of localization, in order to be as generic as possible. The proposed method enables an autonomous agent to perform SLAM in an unknown environment. Finally, we validate the proposed method, through several experiments with several benchmark datasets.

A neuro-based network for on-line topological map building and dynamic path planning

This paper presents a novel combination method for on-line topological map building and dynamic path planning. The proposed method consists of two main components: Bayesian Adaptive Resonance Associative Memory (Bayesian ARAM) and forward-backward propagation path planner. Bayesian ARAM incrementally clusters sensory information and generates topological map. The explored environment is described as a group of neurons (nodes) and edges. Each neuron (nodes) represents a distinct place and it is defined as multi-dimensional Gaussian distribution which does not require any prior knowledge of what a place is supposed to be to make it works in natural environment. The topological map is incrementally generated by Bayesian ARAM. The forward-backward propagation path planner consists of two process: forward propagation determines the possible path while backward propagation with neuron pruning eliminates inefficient neurons and determines the optimum pathway from current location to target location based on the generated map information. The effectiveness of our proposed method is validated by several standardized benchmark datasets.

Biologically Inspired Topological Gaussian ARAM for Robot Navigation

This paper presents a neural network for online topological map construction inspired by the beta oscillations and hippocampal place cell learning. In our proposed method, nodes in the topological map represent place cells (robot location) while edges connect nodes and store robot action (i.e. orientation, direction). Our proposed method (TGARAM) comprises 2 layers: the input layer and the memory layer. The input layer collects sensory information and cluster the obtained information into a set of topological nodes incrementally. In the memory layer, the clustered information is used as a topological map where nodes are associated with actions. Then, topological nodes are clustered together into space regions to represent the environment in the memory layer. The advantages of the proposed method are that 1) it does not require high-level cognitive processes and prior knowledge which is able to work in natural environment, 2) it can process multiple sensory sources simultaneously in continuous space, and 3) it is an incremental and unsupervised learning method. Thus, topological map generated by TGARAM is utilised for path planning to constitutes a basis for robot navigation. Finally, we validate the proposed method through several experiments.

Topological Gaussian ARAM for biologically inspired topological map building

This paper presents a new neural network for online topological map building inspired by beta oscillations and hippocampal place cell learning. The memory layer represents the hippocampus, the input layer represents the entorhinal, and the

Genetic Bayesian ARAM for Simultaneous Localization and Hybrid Map Building

\rho

Episodic memory multimodal learning for robot sensorimotor map building and navigation

ρ is the orientation system. In this model, multiple-scale entorhinal grid cell activations form the input layer feature patterns, which are categorized by hippocampal place cells (nodes) and act as spatial categories in the memory layer. Top-down attentive matching and mismatch-mediated reset (beta oscillations), which are triggered by the orientation system, overcome the stability-plasticity dilemma and prevent the catastrophic forgetting of place cell maps. In our proposed method, nodes in the topological map represent place cells (robot location), while edges connect nodes and store robot action (i.e., orientation, direction). Our method is based upon a multi-channel Adaptive Resonance Associative Memory (ARAM) network architecture to obtain multiple sensory sources for topological map building. It comprises two layers: input and memory. The input layer collects sensory data and incrementally clusters the obtained information into a set of topological nodes. In the memory layer, the clustered information is used as a topological map where nodes are associated with actions. The advantages of the proposed method are: (1) it does not require high-level cognitive processes and prior knowledge to make it work in a natural environment; and (2) it can process multiple sensory sources simultaneously in continuous space, which is crucial for real-world robot navigation. Thus, we combine our Topological Gaussian ARAM method (TGARAM) with incremental principle component analysis to constitute a basis for topological map building. Lastly, the proposed method was validated using several standardized benchmark datasets.