Omid Motlagh

Development of a new minimum avoidance system for a behavior-based mobile robot

A new fuzzy logic algorithm is developed for mobile robot navigation in local environments. A Pioneer robot perceives its environment through an array of eight sonar sensors and self positioning-localization sensors. While the fuzzy logic body of the algorithm performs the main tasks of obstacle avoidance and target seeking, an actual-virtual target switching strategy resolves the problem of limit cycles in any type of dead-ends encountered on the way to the target. This is an advantage beyond pure fuzzy logic approach and common virtual target techniques. In this work, multiple traps may have any shape or arrangement from barriers forming simple corners and U-shape dead-ends to loops, maze, snail shape, and other complicated shapes. Robot trajectories are demonstrated by simulation work and compared with results from other related methods to prove the robustness of this method.

An expert fuzzy cognitive map for reactive navigation of mobile robots

A control technique is described for reactive navigation of mobile robots. The problems of large number of rules, and inefficient definition of contributing factors, e.g., robot wheel slippage, are resolved. Causal inference mechanism of the fuzzy cognitive map (FCM) is hired for deriving the required control values from the FCMs motion concepts and their causal interactions. The FCM-based control is proven to be advantageous over rule-based techniques. The developed system is utilized to control a Pioneer platform. The results and comparisons with the related works are given using ActivMedia simulation and a developed FCM simulation tool. An error estimation technique is used to measure the error between the actual and the simulation results.

A genetic algorithm for optimization of integrated scheduling of cranes, vehicles, and storage platforms at automated container terminals

Commonly in container terminals, the containers are stored in yards on top of each other using yard cranes. The split-platform storage/retrieval system (SP-AS/RS) has been invented to store containers more efficiently and to access them more quickly. The integrated scheduling of quay cranes, automated guided vehicles and handling platforms in SP-AS/RS has been formulated and solved using the simulated annealing algorithm in previous literatures. This paper presents a genetic algorithm (GA) to solve this problem more accurately and precisely. The GA includes a new operator to make a random string of tasks observing the precedence relations between the tasks. For evaluating the performance of the GA, 10 small size test cases were solved by using the proposed GA and the results were compared to those from the literature. Results show that the proposed GA is able to find fairly near optimal solutions similar to the existing simulated annealing algorithm. Moreover, it is shown that the proposed GA outperforms the existing algorithm when the number of tasks in the scheduling horizon increases (e.g. 30 to 100).

An FCM modeling for using a priori knowledge: application study in modeling quadruped walking

Fuzzy cognitive map (FCM) is well established as a decision-making mechanism with many applications. This paper presents a new strategy for realistic FCM-based inference named input-sensitive FCM. The problem of lack of influence from initial concepts’ weights or priory knowledge on decision outputs is resolved. The results and comparisons with the existing inference models are included to evaluate the strength of the new strategy. The quadruped walking cycle is simulated as a case study for sanity testing and validation of the developed model in terms of realistic decision outputs.

Virtual force field algorithm for a behaviour-based autonomous robot in unknown environments:

The present paper describes a real-time motion-planning approach which lies in the integration of three techniques: fuzzy logic (FL), virtual force field (VFF), and boundary following (BF). The FL algorithm is used for velocity control based on sonar readings. The path-planning algorithm is based on the VFF and BF methods. The proposed navigation system differs from previous works in terms of using different algorithms for planning robot motion. Other improvements concern functional and computational aspects of the design and integration of the modules. The robot shows robust performance in complex situations and local minimum scenarios. Simulation results show the effectiveness of the developed system in various environments with long walls, U-shaped, maze-like, and other types of clutter.

An agile FCM for real-time modeling of dynamic and real-life systems

Fuzzy cognitive map (FCM) is a well-established model of control and decision making based on neural network and fuzzy logic methodologies. It also serves as a powerful systematic way for analyzing real-life problems where tens of known, partially known, and even unknown factors contribute to complexity of a system. FCM-based inference requires a neural activation function much like other neural network systems. In modeling, in addition to an activation function, FCM involves with weight training to learn about relationships as they exist among contributing factors. Therefore, numerous contributing factors could be analyzed to understand the behaviors of factors within a real-life system and to represent it in form of tangible matrices of weights. This article discusses a new incremental FCM activation function, named cumulative activation, and introduces a new weight training technique using simulated annealing (SA) known as agile FCM. Smooth variation of FCM nodes that is due to cumulative nature of inference results into faster convergence, while a unique minimum cost solution is guaranteed using the SA training module that is entirely expert-independent. A combination of these two techniques suits time-related applications where inclusion of temporal features is necessary. The resulted system is examined through numerical example datasets where the candidate FCM shows sensitivity to dynamic variables over time. A real-life example case is included as well to further support the effectiveness of the developed FCM in modeling of natural and complex systems.

Knowledge-Mining the Australian Smart Grid Smart City Data: A Statistical-Neural Approach to Demand-Response Analysis

Large scale field trials of smart grid technologies provide important insights as they capture the complex interdependencies of all the key variables, including consumer behaviours, which are needed for their effective evaluation. We present the Australian Smart Grid Smart City program and describe its big data using a narrative approach to hasten understanding and further analyses by others. Then we present a novel statistical-neural approach to maximise knowledge extraction from large datasets of diurnal load profiles, and demonstrate its use in evaluating the effectiveness of two cost-reflective product offerings, a Network-type and a Retail-type product bundle. The methods of analyses include Principal Component Analysis and Self-Organising Mapping. The results for the mid-winter electricity consumption profiles of participating households in July 2013 in New South Wales showed consumption behaviour changes with up to 12 % reduction in relative peak demand at 700 households who accepted the offerings compared to the control group. The resultant load factor of the high consuming outliers improved by about 18 % under demand-response compared to the control group. The feature-based classifier also revealed which behavioural components change due to users’ demand-response activities; results compared favourably with third party consumer survey results.

Analysis of household electricity consumption behaviours

Adoption of renewable electricity generation technologies such as photovoltaic (PV) systems is at the early majority stage in most developed countries. Depending on solar capacity, applied feed-in tariff, and other factors, households exhibit different electricity consumption behaviours which can potentially assist in Demand Side Management (DSM) of electricity usage. This article presents three univariate analysis methods to infer deliberative behavioural patterns at households with solar electricity generation capacity. Analysis methods include qualitative Principal Component Analysis (PCA), unsupervised Hebbian-based clustering, and clustering using a semi-supervised Self-Organising Map (SOM). The techniques are individually applied to 300 sample households with rooftop PV panels operating under a Gross Metering (GM) scheme. According to the PCA, the dominant behaviours are often general among most households, and therefore reveal themselves on first and second principal components. However, on the third and fourth components some specific household behaviours related to load-shifting and self-consumption, are observed. The Hebbian model differentiates between at least eight behaviour types, some of which indicate deliberative behaviours by the households. Most effectively, SOM clustering clearly detects a self-consumption behaviour attributed to domestic electricity generation. A control group of 400 households is analysed to ensure uniqueness of the self-consumption behaviour to customers with solar PV installed. The techniques developed herein may be able to be used by electricity utilities to assess the influence that future tariff and technology offerings will have on behavioural aspects of customer electricity consumption.

AN ALTERNATIVE APPROACH TO FCM ACTIVATION FOR MODELING DYNAMIC SYSTEMS

Recurrent neural models such as fuzzy cognitive maps (FCM) are well established in decision modeling through progressive variations of systems’ concepts. However, existing activation functions have shortcomings, such as a lack of sensitivity to weights of initial concepts, which is due to exaggerated focus on the training of networks’ causal links. Therefore, in most cases, decision outputs converge toward lower and higher extremes and do not represent gray scales. Another disadvantage is that current models require sufficient time delay for convergence toward results. This makes FCM unable to handle transient changes in input. A new technique has been examined in this article using a real-life example to improve FCM activation in terms of fast response to dynamic stimuli. A simple expert model of hexapod locomotion is developed without focus on weight training. The systems response to stimuli is evaluated through a complete six-phase stride to validate the effectiveness of the developed activation function.

A novel graph computation technique for multi-dimensional curve fitting

Curve-fitting problems are widely solved using numerical and soft techniques. In particular, artificial neural networks (ANN) are used to approximate arbitrary input–output relationships in the form of tuned edge weights. Moreover, using semantic networks such as fuzzy cognitive map (FCM), single graph nodes could be directly associated with their actual grey scales rather than binary values as in ANN. This article examines a novel methodology for automatic construction of FCMs for function approximation. The main contribution is the introduction of nested-FCM structure for multi-variable curve fitting. There are step-by-step example cases along with the obtained results to serve as a guide to the new methods being introduced. It is shown that nested FCM derives relationship models of multiple variables using any conventional weight training technique with minimal computation effort. Issues about computational cost and accuracy are also discussed along with future direction of the research.