Neal Snooke

Feature selection Inspired classifier ensemble reduction.

Classifier ensembles constitute one of the main research directions in machine learning and data mining. The use of multiple classifiers generally allows better predictive performance than that achievable with a single model. Several approaches exist in the literature that provide means to construct and aggregate such ensembles. However, these ensemble systems contain redundant members that, if removed, may further increase group diversity and produce better results. Smaller ensembles also relax the memory and storage requirements, reducing systems run-time overhead while improving overall efficiency. This paper extends the ideas developed for feature selection problems to support classifier ensemble reduction, by transforming ensemble predictions into training samples, and treating classifiers as features. Also, the global heuristic harmony search is used to select a reduced subset of such artificial features, while attempting to maximize the feature subset evaluation. The resulting technique is systematically evaluated using high dimensional and large sized benchmark datasets, showing a superior classification performance against both original, unreduced ensembles, and randomly formed subsets.

The Flame system: automating electrical failure mode and effects analysis (FMEA)

It is well known that FMEA is both tedious and time consuming-so much so, that an FMEA analysis on the design of a system is often only completed after a first prototype has been constructed. This situation can lead to time, effort and money being wasted. Automating the FMEA process will improve the speed and consistency with which an FMEA analysis can be performed. The Flame system aims to provide engineers with a knowledge based system (KBS) which is capable of performing automated FMEA. At present, we are concentrating our efforts on electrical design FMEA, however mechanical and software FMEA will be the subjects of future study. The input to the Flame system consists of a physical description of a particular circuit and a description of that circuits functionality. The output from Flame will be a complete (or near complete) FMEA form which can be checked, annotated and signed off by an engineer. The Flame system demonstrates that it is indeed possible to provide engineers with a means of performing automated electrical FMEA. The application considered is automobile systems.

A language for functional interpretation of model based simulation

Functional modeling is in use for the interpretation of the results of model based simulation of engineered systems for design analysis, enabling the automatic generation of a textual design analysis report that expresses the results of the simulation in terms of the systems purpose. We present a novel functional description language that increases the expressiveness of this approach, allowing a system function to be decomposed in terms of subsidiary functions as well as required effects, increasing the range both of systems and design analysis tasks for which the approach can be used.

Combining functional and structural reasoning for safety analysis of electrical designs

Increasing complexity of design in automotive electrical systems has been paralleled by increased demands for analysis of the safety and reliability aspects of those designs. Such demands can place a great burden on the engineers charged with carrying out the analysis. This paper describes how the intended functions of a circuit design can be combined with a qualitative model of the electrical circuit that fulfils the functions, and used to analyse the safety of the design. FLAME, an automated failure mode and effects analysis system based on these techniques, is described in detail. FLAME has been developed over several years, and is capable of composing an FMEA report for many different electrical subsystems. The paper also addresses the issue of how the use of functional and structural reasoning can be extended to sneak circuit analysis and fault tree analysis.

A layered approach to automated electrical safety analysis in automotive environments

Software support for the automotive electrical design process is vital, as many of the safety analysis tasks needing to be carried out, while complex, are repetitive and time consuming. Such support is required throughout the design process, but the available commercial tools are only appropriate at specific points in the design process-providing either an early rough analysis or a late but detailed analysis. This paper describes how the capability and utility of safety analysis software can be improved through separating the types of knowledge used into layers. This allows the maximum amount of information to be reused as the design evolves, and enables software tools to track the consequences of changes to the design so that the repercussions of any design change can be understood. The software capability described has profound implications for the design process. Previously, engineers performed a snapshot design safety analysis at some point in the design process, even if they had an automated design safety analysis tool to assist them. The process and tool arrangement described in this paper enables engineers to continually monitor the status of a design, noting the implications of any changes or refinements to the design.

Hierarchical functional reasoning

This paper discusses the use of hierarchies of function in reasoning about automotive electrical systems. Such hierarchies enable more powerful reasoning for applications such as diagnosis, failure mode and effects analysis, sneak circuit analysis and design verification, while also structuring the domain and thus reducing the complexity at any one level. The context of this discussion is the existing AutoSteve system for performing these tasks. The AutoSteve system works with single level electrical schematics, one schematic for each subsystem in the car, and with a set of functional labels for each subsystem. The functional labels can be used to interpret what is happening in a qualitative simulation of the circuit.

Automated sneak identification

Abstract Sneak circuit analysis is an important design analysis method for detecting unexpected interactions between different parts of an electrical circuit design. It demands a great deal of effort from an engineer, and is also error-prone. A method for the automation of sneak circuit analysis is described. It uses qualitative reasoning and knowledge about which functions are expected to be active in the circuit, in order to identify potential sneak paths in a circuit. The method has been implemented in software and deployed at several automotive engineering design centres. An example run of the software is shown. It is able to import circuits from a standard electrical computer aided design tool, and so demands very little work of the engineer. The time savings through use of the system are estimated to be several orders of magnitude as well as ensuring a thorough exploration of all possible interactions. The possibility of extending this method to cover the circuitry of a complete car at once is addressed, as are the implications of having not just one such tool, but a toolbox of automated design analysis tools.

Automated FMEA based diagnostic symptom generation

The comprehensive on-board diagnosis of faults in many aerospace and other engineered systems requires real time execution using limited computational resources, and must also provide verifiable behaviour. This paper shows how a diagnostic system satisfying these requirements can be automatically generated from the model based simulation used to produce an automated Failure Modes and Effect Analysis (FMEA). The resulting diagnostic system comprises a set of efficiently evaluated symptoms and their associated faults. The symptoms are complete in that they include all necessary observations required to determine applicable system operating states, unlike other work that finesses this problem by having models for each operating state and producing diagnostics for each operating state separately. The symptoms are also efficient because they abstract complex system behaviour based on observations available to the diagnostic system and only preserve sufficient symptom detail to isolate faults given these available observations. This work has been done in the context of diagnosing autonomous aircraft, and is illustrated with examples from that domain. The models used as a basis for automated generation of diagnostics were originally produced to automate the production of a FMEA report, and the paper also considers the relationship between FMEA and diagnostics that provides verification of the failure effects predicted by the simulation and hence validation of the generated symptoms.

Derivation of High Spatial Resolution Albedo from UAV Digital Imagery: Application over the Greenland Ice Sheet

Measurements of albedo are a prerequisite for modelling surface melt across the Earths cryosphere, yet available satellite products are limited in spatial and/or temporal resolution. Here, we present a practical methodology to obtain centimetre resolution albedo products with accuracies of 5% using consumer-grade digital camera and unmanned aerial vehicle (UAV) technologies. Our method comprises a workflow for processing, correcting and calibrating raw digital images using a white reference target, and upward and downward shortwave radiation measurements from broadband silicon pyranometers. We demonstrate the method with a set of UAV sorties over the western, K-sector of the Greenland Ice Sheet. The resulting albedo product, UAV10A1, covers 280 km2, at a resolution of 20 cm per pixel and has a root-mean-square difference of 3.7% compared to MOD10A1 and 4.9% compared to ground-based broadband pyranometer measurements. By continuously measuring downward solar irradiance, the technique overcomes previous limitations due to variable illumination conditions during and between surveys over glaciated terrain. The current miniaturization of multispectral sensors and incorporation of upward facing radiation sensors on UAV packages means that this technique will likely become increasingly attractive in field studies and used in a wide range of applications for high temporal and spatial resolution surface mapping of debris, dust, cryoconite and bioalbedo and for directly constraining surface energy balance models.

Identifying design glitches through automated design analysis

Engineers have developed a number of design techniques in order to detect problems in their designs, such as yellow-boarding, FMECA, FTA and sneak circuit analysis. Concurrent engineering demands that all such design analysis activities are carried out in a timely manner-instantaneously would be ideal. For electrical systems incorporating some electronics, many such techniques can be automated using qualitative simulation to generate appropriate results. There are two main advantages of using qualitative simulation over using numerical simulation in tools such as SPICE or SABER. The first advantage is that the analysis can be performed early in the design lifecycle, identifying problems when it is less costly to fix them. The second advantage is that component models are easy to specify and are highly reusable when compared with models built in SPICE or SABER.