Martin Colley

Creating an ambient-intelligence environment using embedded agents

The Essex intelligent dormitory, iDorm, uses embedded agents to create an ambient-intelligence environment. In a five-and-a-half-day experiment, a user occupied the iDorm, testing its ability to learn user behavior and adapt to user needs. The embedded agent discreetly controls the iDorm according to user preferences. Our work focuses on developing learning and adaptation techniques for embedded agents. We seek to provide online, lifelong, personalized learning of anticipatory adaptive control to realize the ambient-intelligence vision in ubiquitous-computing environments. We developed the Essex intelligent dormitory, or iDorm, as a test bed for this work and an exemplar of this approach.

A Review of Non-Invasive Techniques to Detect and Predict Localised Muscle Fatigue

Muscle fatigue is an established area of research and various types of muscle fatigue have been investigated in order to fully understand the condition. This paper gives an overview of the various non-invasive techniques available for use in automated fatigue detection, such as mechanomyography, electromyography, near-infrared spectroscopy and ultrasound for both isometric and non-isometric contractions. Various signal analysis methods are compared by illustrating their applicability in real-time settings. This paper will be of interest to researchers who wish to select the most appropriate methodology for research on muscle fatigue detection or prediction, or for the development of devices that can be used in, e.g., sports scenarios to improve performance or prevent injury. To date, research on localised muscle fatigue focuses mainly on the clinical side. There is very little research carried out on the implementation of detecting/predicting fatigue using an autonomous system, although recent research on automating the process of localised muscle fatigue detection/prediction shows promising results.

Evolving spiking neural network controllers for autonomous robots

In this paper we introduce a novel mechanism for controlling autonomous mobile robots that is based on using spiking neural networks (SNNs). The SNNs are inspired by biological neurons that communicate using pulses or spikes. As SNNs have shown to be excellent control systems for biological organisms, they have the potential to produce good control systems for autonomous robots. In this paper we present the use and benefits of SNNs for mobile robot control. We also present an adaptive genetic algorithm (GA) to evolve the weights of the SNNs online using real robots. The adaptive GA using adaptive crossover and mutation converge in a small number of generations to solutions that allow the robots to complete the desired tasks. We have performed many experiments using real mobile robots to test the evolved SNNs in which the SNNs provided a good response.

A hierarchical fuzzy-genetic multi-agent architecture for intelligent buildings online learning, adaptation and control

In this paper, we describe a new application domain for intelligent autonomous systems--intelligent buildings (IB). In doing so we present a novel approach to the implementation of IB agents based on a hierarchical fuzzy genetic multi-embedded-agent architecture comprising a low-level behaviour based reactive layer whose outputs are co-ordinated in a fuzzy way according to deliberative plans. The fuzzy rules related to the room resident comfort are learnt and adapted online using our patented fuzzy-genetic techniques (British patent 99-10539.7). The learnt rule base is updated and adapted via an iterative machine-user dialogue. This learning starts from the best stored rule set in the agent memory (Experience Bank) thereby decreasing the learning time and creating an intelligent agent with memory. We discuss the role of learning in building control systems, and we explain the importance of acquiring information from sensors, rather than relying on pre-programmed models, to determine user needs. We describe how our architecture, consisting of distributed embedded agents, utilises sensory information to learn to perform tasks related to user comfort, energy conservation, and safety. We show how these agents, employing a behaviour-based approach derived from robotics research, are able to continuously learn and adapt to individuals within a building, whilst always providing a fast, safe response to any situation. In addition we show that our system learns similar rules to other offline supervised methods but that our system has the additional capability to rapidly learn and optimise the learnt rule base. Applications of this system include personal support (e.g. increasing independence and quality of life for older people), energy efficiency in commercial buildings or living-area control systems for space vehicles and planetary habitation modules.

Learning and adaptation of an intelligent mobile robot navigator operating in unstructured environment based on a novel online Fuzzy-Genetic system

Abstract In this paper we present our novel Fuzzy–Genetic techniques for the online learning and adaptation of an intelligent robotic navigator system. Such a system could be used by autonomous mobile vehicles navigating in unstructured and changing environments. In this work we focus on the online learning of the obstacle avoidance behaviour, which is an example of a behaviour that receives delayed reinforcement. We show how this behaviour can be co-ordinated with other behaviours that receive immediate reinforcement (such as goal seeking and edge following) learnt during our previous work to generate an intelligent reactive navigator that can deal with unstructured and changing outdoor environments. The system described uses a life long learning paradigm whereby it is able to dynamically adapt to new environments and update its knowledge base.

Programming iSpaces — A Tale of Two Paradigms

‘iSpace, the final frontier’ — this parody of Star Trek encapsulates many of our aspirations for this area as, in the longer term, iSpaces are likely to be the key to mankind’s successful exploration of deep space. In outer space, or hostile planetary habitats, it is inevitable that people will survive in wholly technologically supported artificial environments [1]. Such environments will contain numerous communicating computers embedded into a myriad of devices, sensing, acting, delivering media, processing data, and providing services that enhance the life-style and effectiveness of the occupant and, in outer space, preserving human life. Such environments will also include robots [2]. In today’s iSpaces, while human life will not normally be at stake, the underlying principles and technology are much the same. Today our homes are rapidly being filled with diverse types of products ranging from simple lighting systems to sophisticated entertainment systems, all adding to the functionality and convenience available to the home user. The iSpace approach envisages that, one day soon, most artefacts will contain embedded computers and network connections, opening up the possibility for hundreds of communicating devices, co-operating in communities serving the occupant(s). The seeds of this revolution have already been sown in that pervasive technologies such as the Internet and mobile telephones already boast over 200 and 680 million users, respectively [3].

FPGA implementation of spiking neural networks - an initial step towards building tangible collaborative autonomous agents

This work contains the results of an initial study into the FPGA implementation of a spiking neural network. This work was undertaken as a task in a project that aims to design and develop a new kind of tangible collaborative autonomous agent. The project intends to exploit/investigate methods for engineering emergent collective behaviour in large societies of actual miniature agents that can learn and evolve. Such multi-agent systems could be used to detect and collectively repair faults in a variety of applications where it is difficult for humans to gain access, such as fluidic environments found in critical components of material/industrial systems. The initial achievement of implementation of a spiking neural network on a FPGA hardware platform and results of a robotic wall following task are discussed by comparison with software driven robots and simulations.

A fuzzy incremental synchronous learning technique for embedded-agents learning and control in intelligent inhabited environments

In this paper we introduce a novel learning and adaptation system for embedded-agents embodied and situated in intelligent inhabited environments. The fuzzy incremental synchronous learning (ISL) techniques we describe seek to provide an online, life-long, non-intrusive method for learning personalised behaviour and anticipatory adaptive control for physical environments.

Online Learning and Adaptation of Autonomous Mobile Robots for Sustainable Agriculture

In this paper we will introduce the application of our newly patented double hierarchical Fuzzy-Genetic system (British patent 99-10539.7) to produce an intelligent autonomous outdoor agricultural mobile robot capable of learning and calibrating its controller online in a short time interval and implementing a life long learning strategy. The online and life long learning strategy allow the outdoor robots to increase their experience and adapt their controllers in the face of the changing and dynamic unstructured outdoor agricultural environments. Such characteristics permit prolonged periods of operation within dynamic agricultural environments, which is an essential feature for the realization of a platform vehicle for use in sustainable agriculture and organic farming.

A fuzzy-genetic based embedded-agent approach to learning and control in agricultural autonomous vehicles

This paper describes the design of a fuzzy controlled autonomous robot, incorporating genetic algorithms (GA) based rule learning, for use in an outdoor agricultural environment for path and edge following activities which involve spraying insecticide, distributing fertilisers, ploughing, harvesting, etc. The robot has to navigate under different ground and weather conditions. This paper addresses the development of an online self-learning system based on modified version of the fuzzy classifier system. The proposed technique has resulted in rapid convergence suitable for learning individual behaviours online without the need for simulation. The controller was tested on both an in-door and out-door mobile robot operating with different types of sensors, propulsion and steering. Experiments include operating the vehicle following irregular crop edges under different weather and ground conditions within a tolerance in the order of 2 inches.