Kevin Burn

Matching Crimes Using Burglars' Modus Operandi: A Test of Three Models

’Hard’ forensic evidence (eg DNA) may be the best means of linking crimes, but it is often absent at burglary crime scenes. Modus operandi information is always present to some degree, but little is known of its significance in matching burglaries. This paper evaluates the ability of three algorithms to match a target crime to the actual offender within a database of 966 offences. The first (RCPA) uses only MO information, the second (RPAL) only temporal and geographic data and a third (COMBIN) is a combination of the two. A score of one indicates a perfect match between the target crime and the case selected by the algorithm. The lowest possible rank is 965 showing that 965 cases were selected before the target offence. The RPAL and COMBIN each achieve a perfect match for 24 per cent of the crimes and succeed in matching over half of the crimes at a score of 10 or less. For prolific offenders, using MO information alone is better than temporal and geographic data, although the best performance is achieved when in combination. Behavioural, spatial and temporal information is collected by many Police Services. The value and means of utilising such data in linking crimes is clearly demonstrated.

Adaptive and Nonlinear Fuzzy Force Control Techniques Applied to Robots Operating in Uncertain Environments

For robots to perform many complex tasks there is a need for robust and stable force control. Linear, fixed-gain controllers can only provide adequate performance when they are tuned to specific task requirements, but if the environmental stiffness at the robot/task interface is unknown and varies significantly, performance is degraded. This paper describes the design of two nonlinear, fuzzy force controllers, developed primarily using analytical methods, which overcome the problems of conventional control. Using simulation and an experimental robot, they are shown to perform well over a wide range of stiffness and both a quantitative and qualitative assessment of their performance compared with conventional force control is presented.

Appearance-based localization for mobile robots using digital zoom and visual compass

This paper describes a localization system for mobile robots moving in dynamic indoor environments, which uses probabilistic integration of visual appearance and odometry information. The approach is based on a novel image matching algorithm for appearance-based place recognition that integrates digital zooming, to extend the area of application, and a visual compass. Ambiguous information used for recognizing places is resolved with multiple hypothesis tracking and a selection procedure inspired by Markov localization. This enables the system to deal with perceptual aliasing or absence of reliable sensor data. It has been implemented on a robot operating in an office scenario and the robustness of the approach is demonstrated experimentally.

Measurement of the stiffness of endoscopes—a plea for commonality

Editor,—In a previous issue ( (2000) Gut 46:801–8. [OpenUrl][1][PubMed][2] ), Brooker and colleagues described their experience with an exciting new variable stiffness colonoscope. They made the point that a stiffer colonoscope shaft reduces recurrent looping but makes passage through an angulated sigmoid more difficult and causes more stretching and hence pain when loops do occur. Conversely, the more flexible thinner paediatric instruments are better for negotiating a fixed or narrow sigmoid colon but then tend to allow recurrent loop formation later in the procedure. Their randomised trial using either a standard Olympus CF200HL (13.3 mm shaft diameter) or a prototype (Olympus XCF-SH230L—12.9 mm shaft diameter) variable stiffness colonoscope looked very promising although in one case a paediatric Olympus PCF230I (11.3 mm shaft diameter) was required to get past a fixed sigmoid … [1]: {openurl}?query=rft.jtitle%253DMedical%2B%2526%2Bbiological%2Bengineering%2B%2526%2Bcomputing%26rft.stitle%253DMed%2BBiol%2BEng%2BComput%26rft.aulast%253DWehrmeyer%26rft.auinit1%253DJ.%2BA.%26rft.volume%253D36%26rft.issue%253D4%26rft.spage%253D475%26rft.epage%253D479%26rft.atitle%253DColonoscope%2Bflexural%2Brigidity%2Bmeasurement.%26rft_id%253Dinfo%253Apmid%252F10198532%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [2]: /lookup/external-ref?access_num=10198532&link_type=MED&atom=%2Fgutjnl%2F49%2F1%2F154.1.atom

Environment classification using kohonen self-organizing maps

Abstract: This paper describes a new method for classifying three-dimensional environments in real time using Kohonen self-organizing maps (SOMs). The method has been developed to enable autonomous underwater vehicles (AUVs) to navigate without human intervention in previously unexplored subsea environments, but can be generalized to unmanned aircraft equipped with appropriate sensors flying over unchartered terrains, or spacecraft exploring remote planets, subject to appropriate pre-mission training. The method involves a fuzzy comparison between a SOM created in real time using accumulated sensor data and a class atlas of SOMs derived from previously trained and manually classified environments. This enables mission- and environment-appropriate AUV navigation strategies to be selected in real time. Simulation results using real-world, three-dimensional environment data acquired from digital elevation maps are presented, which demonstrate the potential of the method.

A Self-tuning Fuzzy Robotic Force Controller

Most industrial robots are controlled as position servo-based manipulators. This has made most advanced force control algorithms unpractical and difficult to implement. In this paper a position based fuzzy PID force controller is proposed to regulate contact force of a six degree of freedom industrial robot where the environment contact stiffness varies considerably. Based on a relationship between fuzzy PID and conventional PID control laws and the application of a simple fuzzy self-tuning method, the controller is tuned and satisfying experimental results have been obtained to validate its efficiency.

Visual robot homing using Sarsa(λ), whole image measure, and radial basis function

This paper describes a model for visual homing. It uses Sarsa(lambda) as its learning algorithm, combined with the Jeffery divergence measure (JDM) as a way of terminating the task and augmenting the reward signal. The visual features are taken to be the histograms difference of the current view and the stored views of the goal location, taken for all RGB channels. A radial basis function layer acts on those histograms to provide input for the linear function approximator. An on-policy on-line Sarsa(lambda) method was used to train three linear neural networks one for each action to approximate the action-value function with the aid of eligibility traces. The resultant networks are trained to perform visual robot homing, where they achieved good results in finding a goal location. This work demonstrates that visual homing based on reinforcement learning and radial basis function has a high potential for learning local navigation tasks.

Development of a non-linear force controller using fuzzy logic techniques:

Abstract As the demand for robots to perform complex tasks grows, there is an increasing need to utilize robust and stable force control strategies. Most current schemes can only provide adequate force control with the controller tuned to specific task requirements since, if there is a wide variation in the overall compliance at the robot tool/task interface, the performance is rapidly degraded. This paper describes a method for the design of a fuzzy logic controller to replace a conventional controller in a force control loop. The method combines an analytical approach to controller tuning, with the intuitive properties and self-adjusting gain characteristics associated with fuzzy logic systems. It is demonstrated using a model of a single-axis experimental rig and is shown to perform well over a wide range of stiffnesses. The implementation of the controller on the actual rig is also described. Experimental results compare favourably with those obtained from simulation using an accurate model of the system. Issues relating to the implementation of the controller on multi-axis systems are also addressed.

A software tool for automating the design of robot fuzzy force controllers

This paper describes a software tool to automate a design method for robotic fuzzy force control. The original method was developed to ensure robust and stable force control in situations where environmental stiffness at the robot/task interface is unknown, obviating the use of fixed-gain controllers. It did, however, involve a manual design process requiring significant knowledge of control theory and fuzzy logic. This process has been automated in the form of a Windows-based application, requiring minimal user inputs and incorporating an automatic tuning technique for improved performance in the final controller application. Results obtained from an experimental robot are presented.

A MATLAB toolbox for teaching modern system identification methods for industrial process control

Many complex processes can be successfully controlled using two- or three-term (PI or PID) controllers. However, the task of selecting individual controller gains (tuning) can be time consuming and expensive. An important aspect of tuning is the determination of a mathematical model that captures the dynamics of the process. This is known as system identification, which can be achieved in various ways using either step response or relay feedback tests. This paper describes a Matlab toolbox that is currently being developed to facilitate the teaching of modern approaches to system identification and controller design found in the control research literature. When used in undergraduate teaching, it serves the dual purpose of making the techniques more accessible to students, and of providing challenging project work in the field of control engineering. To demonstrate the toolbox, a recently implemented algorithm is described, together with details of its implementation within the Matlab/Simulink environment. Its effectiveness is then demonstrated by comparing it with a classical approach. Suggestions are also made as to how the software may be used for teaching purposes, and ongoing and future developments are described.